EDINBURGH : PUBLISHED BY HIS MAJESTY’S STATIONERY OFFICE

GE ‘STREET, EDINBURGH; throushafy] Bookseller, anbGrown Copytight Reserved.

Sten N | eae

Dates of the several Numbers of this Volume.

Number LXVI, pp. 1-12 for January 1923.

Number LXVII, pp. 13-30 for April 1923.

Number LXVIII, pp. 31-74 for November 1923. Number LXIX-LXX, pp. 75-393 for November 1924.

ae

List of Contents to Vol. XIV, 1923-1924.

—

The Royal Botanic Garden List of Staff at November 1924 Rules and Regulations Historic Notice Regius Keepers Principal Gardeners from 1756 Features of the Garden. With Key-Plan Teaching in the Garden Enumeration of Visitors, 1889-1923 Regional Spread of Moisture in the Wood of Trees. III.

By Professor William Grant Craib, M.A. eee Plates CLXXX-CXC) The enw of Late Frost on the Wood of Acer Pseudoplat- Linn. By James uname B.Sc., ites mice CXCI-CXCHT) ;

aaa amen of Tsuga. By Dorothy G. Downie, B.Sc., Sc. (For.). (With Plate CXCIV) . :

A New Chinese Osmanthus. By Alfred Rehder ;

7 Ustapsbiaed in Sarcococca ruscifolia, Stack: By Matthew

g Orr (with two figures in the text) - :

iv CONTENTS

On the Resin Ducts in the Leaf of Picea brachytyla, Pritzel. By Matthew Young Orr (with one figure in the text)

Note on the Varieties of Geum elatum, Wallich. By W. Edgar Evans, B.Sc. (with Plates CXCV-CXCVI)

New Primulaceae. By Professor William Wright Smith and George Forrest

The Leaf Glands of Dioscorea macroura, Harms. By Matthew Young Orr (with Plates CXCVILCXCVIIT and three figures in the text)

A New Species of Desmogyne. By Mary M. B. Knagg,

Plantae Chinenses Forrestianae :—

Catalogue of the Plants (excluding Rhododendron) collected by George Forrest during his fifth ex- ploration of Yunnan and Eastern Tibet in the years 1921-22. By the Staff of the Meroe Botanic Garden, Edinburgh.

Printed under the authority of His Majesty's Stationery Office by JoHN PeLLow, 3 Lintrathen Street, Dundee.

‘eo Wtx623/8.0.P, 15 375 485 G7

fj *

| (on 8, Bain, No. Lxv1, Jannary 1923.)

Regional Spread of Moisture in the Wood of Trees. III.

BY

WILLIAM GRANT CRAIB, M.A., Regius Professor of Botany in the University of Aberdeen.

With Plates CLXXX-+CXC.

[. Sparta gta ed Trees meh g the Period late Autumn to atly Spring (Contd.). :. IT

Deciduous-Leaved Trees during the Veitene Period. a. Bud opening and hid pe J Phase. : b. Full-leaf conditio III. Evergreen Broad-Leaved ied IV. Coniferae. V. Wind as an Agent in ‘Ries Shee:

PU UO! “Iu wm ww

I. Deciduous-Leaved Trees during the Period late Autumn to Early Spring (Contd.).

AN earlier paper* summarised the results of experiments on the moisture content of the wood of newly felled trees of Acer Pseudoplatanus at various times during the leafless period. Up to that time only two species of true heart-wood trees had been examined, viz. Ulmus montana and Quercus cerris. From the one spetimen of each that had then been examined it was possible to say} merely that the results obtained were comparable with those for Acer and to suggest that in all probability true heart- wood trees would be found to be similar to Acer in their behaviour as regards moisture-spread throughout the entire leafless period.

Subsequently a series of trees of Ulmus was examined in exactly the same way as that of Acer, the months of felling being October, November, January and March. The general results were exactly as predicted. Moisture storage takes place in the heart-wood in the end of the year and this moisture travels out- wards as the season progresses. Slight differences of detail when compared with Acer were observed but these in no way affected the general results and in the present state of our knowledge I

* Notes Roy. Bot. Gard. ae vol. i No. LI. p. 1 (2938) ; for second

2 CRAIB—REGIONAL SPREAD OF MoIstuRE IN Woop oF TREES.

had decided to limit the Ulmus experiments to elucidation of the one broad principle. It is unnecessary to supply the graphs for Ulmus as these are in general but repetition of those for Acer.

It may be interesting to record one small experiment which although not carried out by a method ensuring absolute accuracy, - vet gave rather unexpected results which might be regarded as suggestive of future lines of investigation.

In the case of one of the trees the annual rings were so broad that I decided to split the ring roughly into Spring and Summer wood and to examine these two separately. The result was that the Summer wood contained on an average about 10% more moisture than the Spring wood. This tree was then in the centre-storage condition. Unfortunately no equally broad rings were present in any of the other trees but some were just broad enough to allow of their being split and the result was that the Spring wood showed a decidedly higher moisture-percentage than the Summer wood. Apparently then when the moisture is stored in the centre of the trunk the Summer wood has the higher per- centage but when the moisture is moving outwards the Spring wood has the higher. In recording this small experiment I am well aware of the fact that the methods employed do not give results from which satisfactory conclusions can be drawn. No account has been taken, for example, of the difference in specific gravity between the earlier and the later formed wood of the year. In spite of this I think the results may be worthy of record.*

II. Deciduous-Leaved Trees during the Leaf-bearing Period.

From the condition of the last graph published} to the summer condition with, as I presumed, a large central comparatively dry - _area, the transition would seem easy. It was legitimate to imagine the large percentage of moisture just inside the cambium in March as very rich in sugar and as ready to be drafted to the cambium at its own level or higher up. The exhaustion of this store, rapid owing to the tree’s spring growth activities, would result in the summer condition as presumed. ‘The process is _ however no such simple one. j

This paper is a continuation of that already publishedt and gives the results obtained from the following trees of Acer Pseudoplatanus :—

1. Tree felled entire on the morning of 23rd April 1918. At this time the buds were just open and a few leaves here and there were showing. Growth had not yet begun on the trunk but the _* Notes on decidous Coniferae will be found

oo: Soe Roy. Bot. Gard. Edin., vol. xi. ~ Sos: Pi. out. aus. tN — Bot. Gard, a vol. xi. No, LI. p. 1 (191 18).

Crais—REGional, SPREAD OF MoIs?URE IN Woop or TREES. -

cambium throughout was decidedly swollen and “‘bark’’ was easily separable. Growth had begun on two year old twigs where a row of spring vessels was fully formed.

This tree was not a satisfactory specimen as it showed very slow diameter growth compared with neighbouring trees of the same age and subsequently in chipping up the cuts it was observed that the pith was discoloured all through the bole and well up into the crown and also that in the centre of the lower- most cut a small area of wood was discoloured. ‘The unexpected results obtained might have been due to this diseased condition and it was resolved to examine another tree which appeared quite healthy.

2. ‘Tree felled entire on the morning of 7th May 1918, when practically all the buds were open but only a few leaves here and there were expanded, i.e. this tree was almost in the same con- dition as the April one as regards its buds. The results obtained confirmed those of the April tree.

3- Tree felled entire on the morning of 14th May 1918 when the first leaves were all just fully expanded.

4. Tree felled entire on the morning of 20th August 1918.

These four trees represent two phases—distinct in the tree’s activities and very distinct in the results :—

a. Bud opening and expanding phase as represented by the

April and May trees. b. Full leaf condition as represented by the August tree. These two fall to be treated separately.

a. April and May.

As already indicated the results obtained from these trees were quite unexpected. Buds from the March tree opened three days* after felling. Yet owing to a prolonged cold spell I had to wait from 23rd March to 23rd April before a tree showing the opening bud condition was available. This interval was so long that results which I here refer to from the April tree as representing the bud-opening condition may in reality be the state of affairs in a tree with the buds just ready to burst, i.e. the moisture distribution as published for the March treet may not be the final stage in the closed bud condition. Further experiments alone can settle this point. But however that may be we can still discuss the March and April results as a closed-bud con- dition and the bud-opening condition respectively.

The most striking result obtained from this series is the position in the lowermost cuts of the area of maximum moisture content in the case of the April and the first May trees (Plates CLXXXIIL-CLXXXIII). Here we find the moisture heaped up

otes Roy. Bot. Gard. ge cob OD a a sea BP g (r918). | Notes Rey. Bot. * ssaeont oe ., xi. No. Lt. PA. CLIX (1918) .

4 CRAIB—REGIONAL SPREAD OF MorsturE iN Woop or TREES.

towards the centre—exactly the reverse of the March condition and further in the first May tree we have in the bottom cut a return of the three-wave graph. In short, comparison of the graphs of the bottom cuts for the first May tree (P]. CLXX XIII) and the January tree (P1. CLVII) shows that they are practically identical. Nor does the similarity of the results obtained from these two trees end here. There is a general resemblance throughout all the results as plotted for these two trees.

Again, the second May tree shows results (Pl. CLXXXIV) comparable with those of the March tree (Pl. CLIX) but on a much smaller scale. Here there is once more the maximum moisture percentage almost on the outside, though that maximum is very much smaller than in March. Since then the first May tree resembles so closely the January tree in its moisture distribu- tion and the second May tree is comparable with the March tree, one may legitimately ask whether in the interval April-May (?March-May) the moisture does not go through the same move- ments on a reduced scale and at greatly accelerated speed as in the longer interval October-March.

Again in the second May tree (Pl. CLX XXIV) there is in the uppermost two cuts a decided increase in moisture percentage towards the outside. This increased percentage is correlated with the fact that here growth has now commenced.

The time relationship of bud-opening and the commencement of diameter increase is not the same for all trees. In the trees of Acer Pseudoplatanus examined growth began at the twigs when the buds were still closed, extended back along the branches to the trunk and reached a little way down the trunk only after the leaves were fully expanded. This was not the case in a speci- men of Quercus pedunculata felled in May when only at the top of the tree a few male inflorescences were showing and when not a single leaf was to be seen on the tree. Here a row of new spring vessels was already formed throughout the bole. In Acer _ Pseudoplanatus new wood was formed on the trunk only after leaf-expansion ; in Quercus pedunculata new wood was present throughout the trunk before the buds had burst.

b. Full-Leaf Condition, August (Pl. CLXXXV).

Interpretation here presents practically no difficulty. The transpiration current is represented by the higher percentages in the youngest wood and throughout most of the tree the whole central area has an almost uniform percentage averaging but slightly over 60. Probably in a tree of uniform growth the

graph, excluding the transpiration current, would be a straight ene

eH Bie deciduous Coniferae see under Coniferae.

Crain—REGIONAL SPREAD OF MorstURE IN Woop OF TREES. 5

These results and their bearing on the tree-felling period are referred to under Coniferae.

III. Evergreen Broad-Leaved Trees (Plates CLXXXVI- CLXXXVII).

The experiments proposed on evergreen broad-leaved trees were unfortunately interrupted and of the series of Ilex Aqut- folium selected for examination only two trees were felled, one on 5th December 1918 and the second on 15th February 1919. From

-an examination and comparison of the graphs of the moisture- distribution of these two trees, however, one may conclude :—

1. ‘That in the Holly moisture is stored in the centre. This is shown in the graph of bottom cut of the December tree (Plate CLXXXVI) where moisture is stored in the heart-wood.

2. That later in the season the moisture spreads from the centre as shown by the low moisture-percentage of the heart-wood in February (Plate CLXXXVII).

3. That the movement is much slower than in the case of Acer or Ulmus and also that the storage is apparently on a much smaller scale as evidenced by the lower range of percentages.

IV. Coniferae (Plates CLXXXI, CLXXXVIII-CLXXXIX).

Amongst the Conifers the deciduous and evergreen habit ap- parently makes no difference with regard to moisture-spread, but there is a very great difference between the resinous and non-resinous genera.

f the resinous trees, species of Pinus, Larix, Chamaecypans and Cedrus have been examined and in no case is there storage of moisture in the heart-wood.* The impregnation of the walls with resin renders the passage of moisture from the sap-wood to the heart-wood impossible. Hence we have in these resinous trees throughout the year a condition approximating to the August Acer, i.e. in mature trees a large dry heart-wood area surrounded by the narrow wet zone of sap-wood.

A typical moisture distribution graph for a resinous Conifer is given on Plate CLXXXVIII (Chamaecyparis pisifera, felled and Oct. 1918). The sap-wood gives very high percentages and there is a very sudden drop on entering the heart-wood.

* Ho. M, Oliphant (apud Sir Hugh Stewart in Trans. Roy. Bot. Scot. Arb.

vol. x . 67) that a comparison may be drawn between the eonal + moistare ea s of Picea sitchensis vine Acer Pseudoplatanus. ons cai one one cannot et et e moi distribution in the. pisses to which these tie species belong. Of the neybeal —— of resinous age examined at different wires none with central storag comparable wi hh that of Acer has been found. There } here and “there a slight ‘sen in moisture-percentage towards - at the centre of

hea: n

-wood but so slight and at times so inconstant as to a ed withthe central storage condition of Acer.

6 CRAIB

REGIONAL SPREAD OF MorsturE IN Woop or TREES.

This fact shows quite conclusively that moisture distribution is not the determining factor in the selection of the optimum felling period. For, granted that winter is the period for resin- ous conifer felling, then the season in broad-leaved trees with a similar moisture distribution is that of the full-leaf condition.

As stated in the earlier paper* other factors must be considered before a decision on the optimum-felling season can be given. But in the meantime the results from the Acer tree suggest a method of felling when the tree is in full leaf which might at least be given a trial. ‘The present practice is to cut off all branches almost immediately after the tree is felled. Instead of this let a tree be felled say in August and let the branches be left on until the leaves have flagged and are no longer transpiring. ‘The result would be that the leaves would transpire the bulk of the water in the younger wood and we would have a trunk with, in the case of Acer, an almost uniform 60% of moisture throughout. ;

Non-resinous Conifers have been represented in the experi- ments by Taxus, and in Taxus there are two points to be noted.

‘There is here storage of moisture in the heart-wood. In this respect Taxus behaves as the broad-leaved trees, or I should say apparently so because the experiments were interrupted before the summer condition was examined. ‘The winter condition, however, is on the same lines as that of the broad-leaved trees.

In Taxus there was noted a peculiarity not present in any of the other trees examined. In the cross section of the newly felled tree, three distinct areas can be recognised—central heart- wood, sap-wood on periphery, and separating these two areas a narrow zone of very pale wood (Plate CLXXXI). A very super- ficial examination sufficed to show that this narrow pale zone was much drier than either the heart-wood or the sap-wood and the final drying of the chips served but to emphasize the difference.

The position of this dry pale-coloured zone immediately ex- ternal to the heart-wood suggested that it was some intermediate stage or the primary stage of the conversion of the sap-wood into heart-wood. To test the validity of this conclusion reagents were employed to find out the distribution of tannin in this zone and in the neighbouring heart-wood and sap-wood.

Five tannin tests were used :—Osmic Acid + Hydrochloric Acid; Ammonium molybdate in Ammonium chloride solution ; ae chloride in Ether ; Potassium dichromate ; and Ammonium chloride.

Heart-wood responded readily to the various tests as also did the half of the white zone next to the heart-wood, but the white zone gave no reaction in the part next to the sap-wood. These

* Notes Roy. Bot. Gard. Bdin., vol. xi. No. LI. p. eeg = ca 2

CRAIB—REGIONAL SPREAD OF MOISTURE INWoop OF TREES. 7

tests showing the distribution of the tanrin served to confirm the supposition that this white zone is sap-wood in process of conversion into heart-wood by the extraction of moisture of imbibition followed by impregnation of the walls with tannin.

Further, microscopic sections showed that the tannin is actually in the walls of the tracheids, that there is no trace of tannin in the medullary ray cells and that the medullary ray cells are living in the white band. ‘Thus a means of communica- tion is still open through the medullary rays from sap-wood to heart-wood even supposing the impregnation of the tracheid walls with tannin were to act as resin in preventing diffusion through the walls.

Plate CLXXXIX gives the moisture distribution in Taxus baccata in November, showing the high moisture-percentage of the sap-wood, the sudden dip owing to the low moisture- percentage of the white band and the rise in percentage on entering the heart-wood.

The coloured plate (CLXXXI) is a reproduction of a painting made from the newly felled wood of this tree to show the conspicuous white band separating the heart-wood and sap-wood.

V. Wind as an Agent in Moisture-Spread (Plate CXC).

That wind is a very effective agent in the raising of the sap in trees is a well-known fact but in the accompanying graph the results of a gale are shown to be not only sap-raising ; there is in addition a disturbing of the sap-distribution equilibrium.

The tree examined was a specimen of Populus trichocarpa which was blown down during a severe gale on the early morning of 1st March 1918. The results probably do not give the maximum effects of the gale as, not suspecting any such disturb- ance of the moisture distribution, I treated the tree in an exactly similar way to that adopted for the other trees examined, i.e. by determining the moisture-distribution in strips from north to south and east to west through the centre. Fortunately, however, the gale was from only slightly east of north so that the maximum results shown in the graph are those from north to south

Without attempting a physical explanation of the results I would simply draw attention to the increasing percentage in the youngest wood as we ascend the tree on the side on which the gale was blowing, to the decreasing percentage in the youngest wood on the side away from the gale, and to the great contrast in the percentages of these two sides.

It must be borne in mind that in such a tree as Populus tricho- carpa with its elasticity -we would find a maximum moisture dis-

turbance. In less flexible trees such as Quercus or Platanus the | -

j result of a gale would probably be very much smaller.

8 CRAIB—REGIONAI, SPREAD OF MOISTURE IN Woop OF TREES.

One other point must be noted. The graphs show a com- paratively high moisture-percentage in the centre of the trunk. How far this is a result of the gale I cannot say as the whole series of experiments on moisture-spread was interrupted before another Populus could be examined to elucidate this point.

- The absence of a graph for the bottom cut in the N—S direction * explained by the fact that the tree was not quite sound on the north side at the point selected for the cut.

For the material on which the previous and present papers are based I am indebted to the late Sir I. B. Balfour. Not for the material alone, but also for his encouragement throughout the lengthy and laborious experiments and for his friendly advice and criticism I wish now to express my sincere gratitude. Mr R. L Harrow, F.R.H.S., Head Gardener, Royal Botanic Garden, as before, arranged and saw to the carrying out of the details of the felling operations in a very satisfactory way. For the original painting of the wood of Taxus I am indebted to Miss L. Snelling.

hae We PLATES (CLAAA-CRC) Illustrating Professor Craib’s Paper on Moisture-Spread.

PLATE ee erry es ep ge in bole of Acer Pseudo- plata in Augu

CUXXXI, —transverse and ie tudinal sections of newly felled Taxus baccata showing heart-wood and sap-

CLXXXIT. Moisture distribution throughout the trunk of Acer idoplatanus in April.

CLAS — De do. do. do. in May (1). CLAS TV i— Do. do. do. > in May (2). CLASKV — Do. do. do. in August, CLAXX VI. We ey distribution eroughon the trunk of Ilex quifolium in December CLXXXVIT,— Do. do. do. a in February. CLXXXVIIL.—Moisture distribution thro the ¢ trnnk«vof

Chamaecyparis pisifera in ig ies Ser es ee oo throughout the trunk of Taxus i vember.

CeCe Sie a sae tion throughout the trunk of wind- blown Populus trichocarpa.

¢

explanation of methods adopted in these plates se en published in Notes Roy. Bot. Gard, Z. dn, VOL ah. No rE € previous paper

NOTES, R.B.G., EDIN. PLATE CLXXX

under 60 % 61 — 80% 81 — 100% over 100 %

NOTES, R.B.G., EDIN. PLATE CLXXXI

NoreEs, R.B.G., EDIN,

PLATE CLXXXIL ACER ( ApRric)}

-6 t = 80 60 : ti =4 beeps if 70 ‘gre sar re MMR eg fo slim 60 { = te | ee 80 t ; i as #2) Ett? x eo =< ee i 66

NoreEs, R.B.G., EDIN,

PLATE CLXXXITI,

ACER (May- 1) 9G \% $0 aaa @ - 5.4 Poesy guess pears cgen 80 70 2s ~P Sf 70 60 60 =7 =6 on LZ — 80 ov 70 dt a eet aE eas ocegcat HE 70 Soars “60 690 a 2) nex Pw aews Sasa Raten os 80 ou — 70 70 HUES cat —4 Titiiittierreal itn ti Dy ait 60 60 _ -3 an a 80 ou / —_—_— E 70 as 10 60 “ 60

a oy 80 " Su sceagentss ates ccoey cents cist: ome poe ae eet at erm ee Re 70 60 60 50 50 +]

120 a4 a 110 af a 3 100 Ae ve ‘ig rant i Ht ease ttt ue Hts ae aN 80 60 nae X Set

PLATE CELXXXEV-

ACER {May-z

«

—— 60 70 ci ee | 90 , 90 80 -\— : at ee. -<geaea anes +80 NG kencns S255 2 ates ‘ 6 eee a ce tutes o=- abba 70 70 5 sta gate neaazaas 4 i 80 70 io” Saas CSeSehetas So ced noes 33 Bias =a aa B58 <8 | aetrt + : 70 v4 Es ———-- =A si ree 90 “\ 90 80 Pt at 86 1 3 Pau ihess as ti SESE 2NGes Genes OOK wae ees i Le zs edits acre ence ee eat eget eee EE asi 70 60 50 +9 100 j\ 106 80 bas: } 90 50 7 NE eae: / \ 80 9 gees: pees 10 wae Svanneeses <6 Sch ni geet = oeeee Saas Semen tal eter ny HE BES bid 60 60 vi

a es

Nores, R.B.G., EDIN,

TH 110 tt0 100 | 100 90 | 90 80 3 - 80 70 70 60 60 -6 100 100 90 90 80 = Pia 80 70 7 60 60 =§

a o ———

60 +4

Plate CLXXXYV, ACER (Auc.)

$s, R.B.G., EDIN.

NOTE

PLare CLXXXVI. 5

g2s¢e

= caaeeees casa taats Satttteees srtts sensu =csuv toast Canad canes Soest <—— pa SE GMEEE 8d>= = 08 fe asad on 0 ce er t Seet Hbeeeeeet i Bes Godda tian, i O11 a . Ee he ec ee ce Hee ee 00 pt ES ae Hit: saat at 00 00 See ee StH

Nores, R.B.G., EDIN. . PLATE CLXXXVH. ILEX (Fes.)

fe3se8 \ ese 8 \ fess 8838 8 8.8 ? : - HA 7 4 , ig - i aE sarseqenta ts 8&8 ;

NOTES, R.B.G., EDIN.

250

150

100

§0

6 260 200 150 150 160 100 =H 50 50 —4 m3

PLATE CLXXXVIIL.

CHAMAECY PARIS

260 200 150 100 50

250

200

150

100

PLATE CLXXXEX,. Notes, R.B.G., EDIN, TAXUS

120 120 He 110 100 —— $ 100 90 na 998 80 i 80 70 \ 70 rE 60. 50 60 40 40 A 140 140 y 180 130 8 A 5 a \ +20 126 rte i10 j { ar cums ns aN " 100 100 Hae USE SEES SES | 90 90 : 80 80 70 70 / 60 60 50 50 40 40 5 ie 150 150

Notes, R.B.G. EDEN. PLATE CXC.

POPULUS 170 160 , 160 Las 150 | 140 140 130 130 120 120 \ 120 140 110 110 10¢ 100 100 —— 100 90 90 ioe 90 \A { | 90 80 80 80 \ } 80 70 70 160 ree : 150 140 130 130 120 120 11 eSinnamaras i i - Beast 90 Hd 90 \Ay bee i) 90 80 80 80 Si 80 70 70 70 =9 130 | 150 =8 140 140 140 130 130 130 130 120 120 120 120 110 fd 110 110 110 NV i00 a 100 5 iF E 00 100 W 90 V | 90 90 90 80 80 80 “6 150 150 140 140 130 130

The Effect of Late Frost on the Wood of Acer Pseudoplatanus, Linn.

BY

JAMES TEMPLETON, B.Sc.

Senior Botanist, Department of Agriculture, Cairo.

With Plates CXCI-CXCIII.

In the course of an investigation of the moisture content of the wood of Acer Pseudoplatanus, Linn., a tree was felled in the Royal Botanic Garden, Edinburgh, in March of 1921. Transverse cuts were taken from the bole and crown at measured heights above ground level. A superficial examination revealed the fact that one wood-ring of the lowest cut stood out conspicuously from all the others because of a narrow zone at its inner margin being much lighter in colour than the rest of the wood (Pl. CXCI, Fig. 1a). On the cut surface of the stump, the ring was also very well marked ; but 180 cms. above ground level it was in no way different from the normal. A closer examina- tion of this phenomenon gave results which, though in no way connected with the original line of research, appear to be worthy of record.

The annual ring referred to was apparently laid down in the year 1897. On examining transverse and longitudinal sections of it microscopically it became evident that the cambium, from the commencement of its activity in the spring of that year, and for some time later, had acted in a manner far from normal. In Plate CXCII, Fig. 1 shows a transverse section of two consecutive typical annual rings of Acer: Fig. 2., of the rings laid down in 1896 (a) and 1897 (b). In Fig. 1, it will be observed that the medullary rays run continuously through the adjacent rings; but in Fig. 2 this is not the case, for the cambium, instead of giving rise to tracheides, vasa, and medullary ray cells, has laid down at the commencement of its activity for the year, several layers of parenchymatous cells with rather thick, lignified walls. A transition from these to normal wood elements—formed later in

Notes, R.B.G., Edin., No. LXVI, January 1923.]

10 TEMPLETON— EFFECT oF LATE Frost oN Woop oF ACER.

the year—can be traced, but in no case is there any direct con- tinuation of the medullary rays through the two consecutive rings. Plate CXCII, Fig. 3 and Fig. 4 respectively, show the above rings in longitudinal radial section, and in these the same features are illustrated. It is specially evident from Fig. 4, that an unsuccessful attempt to form vasa was made early in the _ year, the result being the production at intervals of short columns of large parenchymatous cells, the transverse walls of which have not disappeared (a). Later in the year a further stage in the transition to vasa can be seen, for though the columns of cells are still short, the transverse walls have now disappeared (b). The transition from the parenchymatous cells to tracheides is clearly demonstrated in this figure.

This abnormal behaviour ‘of the cambium appears to be the result of frost injury. In this connection it is interesting to

note that, in the Records of the Scottish Meteorological Society

for the year 1897,—the year in which the ring was laid down— frost was recorded in Edinburgh on consecutive days, from March 30th to April 6th, and on the 1st of April the ground was covered with snow.

Hartig* found a somewhat similar condition in the wood of several frosted conifers e.g. Pine, Spruce, etc. According to this author, the formation of an ice layer between the cortex and the young wood, results in the crushing and killing of the latter. The medullary rays extend through the ice zone undamaged, and when the ice melts they become extraordinarily broad, since they can expand without any counter pressure. On the inner side of the phloem, the cambium gives rise to parenchymatous tissue, and only eventually in the wood zone is normal wood formed.

In the cases described by Hartig, normal spring wood of the year had been formed before the frost took effect, while in the tree under discussion, the cambial cells, though probably turgid, had. not undergone division.

Hartig also. states that, in places, liable to frost, it is in he under stem portions of coniferous trees, i.e. up to 1-2 metres in height, that these frost rings may be looked for, and. it was just in that region of the Acer that the phenomenon could be

_ observed.

A transverse, cut, containing ph abnormal wood-ring, was set aside to dry, and on examination some weeks later, exhibited not only the usual fractures along the medullary rays (Pl. CXCI, Fig. 2a), but ale aye defined circular split (PI. CXC Fig. 2b and Fig. rb)..:

-- Tensions are set. up i in CHS, ‘aeabee which, result i in splitting taking place along the lines of least resistance—usually ape

, a

* R. Hartig, Lehrbuch er tentren (1900), p. = see

‘TEMPLETON—EFrFect oF LATE Frost oN Woop oF ACER. II

medullary rays. In this case the zone of parenchyma also con- stituted a line of weakness in the wo

It is noticeable also that in no case is a radial fracture in the outer portion of the wood in line with one in the inner portion, due, no doubt, to the break in the continuity of the medullary rays previously referred to.

Since the medullary rays are the routes, par excellence, along which water and food materials in solution move radially in the wood, the question arises as to whether the passage of these fluids was affected by the interruption of the medullary rays.

It is well known that in Summer, starch is stored in the medullary rays and wood parenchyma, after having been con- veyed to these situations in the form of sugar. If the passage of sugar had been restricted, this would be apparent by an examina- tion of the radial distribution of starch. A narrow transverse strip passing through the centre was therefore removed from each cut, and these were placed for some hours in a solution of iodine. The result is shown in Plate CXCIII.* The wood of strips B-G is normal, and the maximum amount of starch in each strip, judging from the depth of the stain, is found in the outermost wood-rings, with a gradual diminution towards the centre. In strip A there is a sudden, and very well marked de- crease in the amount of starch in the wood within the abnormal wood-ring (a), which indicates, in a very convincing manner, that the inward passage of sugar had been very considerably re- stricted, and the internal economy of the tree consequently affected.

In conclusion, I wish to express my indebtedness to the late Sir Isaac Bayley Balfour, for placing the material for investigation at my disposal, and to Mr R. L. Harrow, Head Gardener, Royal Botanic Garden, for his care in carrying out the details of felling.

black areas at b and ¢ are not indicative of presence of starch, but are discolourations in the wood, due to a knot, and decayed pith, respectiv ely,

NS = eae

12 TEMPLETON—Errect or Late Frost on Woop or ACER.

EXPLANATION OF PLATES (CXCI-CXCIII).

Illustrating Mr fay peers s paper on the Effect of Frost on r Pseudoplatants PLATE CXCE. Fig. 1 Bee, of wood showing bade wood ring (a) and it a ef Wood block showing Shores caused by drying : . 3.—Block split into two portions at the frost ring. OCT the SEE of sections es os wood of Acer Pseudo- latanus. bout

= x a Fig. 1.—Normal wood in tran oe Fig The same, in radial ten itudinal section

3 Fig. 2—Abnormal wood in transverse section. Fig. 4.—The same, in radial longitudinal section. CXCIII. Wood acini hi with iodine, showing starch distribution ous heights above ground level

Onerns an oo. Pisiismpyeae pyc to 65 cms. UR cee een tet 225 cms

— 155 cms C-D 130 cms D-E 1oo cms E-F 85 cms F-G BES ene cae hc sik anes 70 cms. G to top of leader eenrea speecumely ca Ws 1 no

Printed under the pany of His Majesty's ani tas Office by Jo 48 Murraygate, Dund

(to6D) Wt7743/SOP171 375 3/23 Gz.

NoTES R.B.G., Edin. PLATE CXCI.

TY ! p tden-*ceeite ge Pal ya E ae we mann. i : Soe, ty i PS aS -~ hb pote d SS EBB ee ; sare Sass eas MRS 7, | octane wt * they

J , Abt 4 STs He

PLATE CXCIl.

3 <

a aid w

BS

<>

‘Ou Avis oe oe Het Se. i ail Peo e sere bres: tS Dyes 8 Sk ORF

bad Ps ii - TYE * #

as “o C ae Lure Hint ee? } icOm

#

2 ihe Fig. 4.

<]

&

rpg et:

0 Oat, :'8'@a!

we iT |

RG

: 26 se te oe

2 - eae +e

raasee

ee okt eT) ‘tinea on ie a ae

@ oan, 4

: @ e@

Fig, 3.

*

e 7 et a

$ .*

.

s 2 * *

i) " ref ae we > hadnt

at 30?

. “8

; ye ase : se a - a® =. a 4 wh. kee oe ne og ee Oe ae esa Ree ee CF oP Sess,

ee ve

Zu }

{ert (seen

eas at el eer oe”

or

Li

NOTES R.B.G., EDIN. y

ie Ei

NOTES R.B.G., EDIN. PLATE CXCIl],

Vol. XIV. FOR OFFICIAL USE. Number LXVII.

NOTES

FROM THE

‘ EDINBURGH. : APRIL 1923.

CONTENTS.

Chinese Species of Tsuga. By Dorothy G. seers: ges B.Sc.(For.) (with Plate CXCIV).

A New Chinese Osmanthus. By Alfred Rehder.

Polyembryony in Sarcococca ruscifolia, Stapf. By Matthew oung Orr (with two figures in the text). : ;

On the Resin Ducts in the Leaf of Picea brachytyla, Pritzel. By Matthew Young Orr (with one figure in the text).

Note on the Varieties of Geum elatum, Wallich. tasd W. a Evans, B.Sc. (with Plates CKCV-CXCVD). :

. INBURGH re PUBLISHED BY BIS Latent STATIONERY OFFICE.

‘SOLD AT THE GARDEN,

eos STATIONERY OFFICE (Scorr: IsH E oe ae — dete ‘EDINBURGH. =

ROYAL BOTANIC GARDEN,

PAGE

21

24

id to be purchased, either directly or through any Bookseller, from fe

Chinese Species of Tsuga. BY

DOROTHY G. DOWNIE, B.Sc., B.Sc. (For.) | Assistant in Botany Department, Aberdeen University.

With Plate CXCIV.

THE following paper is an attempt to arrange and classify the Chinese specimens of Tsuga in the herbarium of the Royal Botanic Garden, Edinburgh. ‘The material was collected in the south-west provinces of China, chiefly Yunnan, Szechuan and Western Hupeh.

Up to the present, species and varieties known are the following :—

Tsuga chinensis, Pritz. <A bies chinensis, Franchet. Tsuga yunnanensis, Masters. Abies yunnanensis, Franchet. Tsuga Brunoniana, Carr., var. chinensis, Masters. Abies dumosa, Loud., var. chinensis, Franchet. Tsuga Mairei, Lemée et Léveillé.

In an enumeration® of Chinese Tsugas, Masters refers certain collections to T. Sieboldii. I have seen no specimen from Chine which can be referred to this species. In a later work} Masters enumerates under T. yunnanensis collections which in his previous paper he had included under T. Sieboldii. Henry 7156 —one of the specimens so treated by Masters—I have, after com- parison with the type in Paris, referred to T. chinensis. Probably some of the other specimens which, so far, I have had no opportunity of examining, may have been wrongly in- cluded under T. yunnanensis by Masters.

Diagnostic characters. ‘The condition of the leaf margin, generally employed in dividing the genus, was not a constant factor here and a division based on the character of the leaf apex was adopted. ‘The utility of this feature as a diagnostic character was strengthened on finding that all plants with an entire leaf apex bore strictly sessile cones, while those with the leaf apex emarginate had shortly stalked cones. In all cases where the # . Linn. Soc. vol. xxvi, p. 556 (1902).

+ sh Lint Son srok/ cnet bas (eo _ [Notes, R.B.G. Edin., No, LXVII, April 1923.]

14 DownrE—CHINESE SPECIES OF TSUGA.

attachment of cones could be observed the combination of characters held. Of the described species* T. yunnanensis comes under the first group (i.e., leaf apex entire) and T. chinensts under the second. As the cone stalk of the plants of the Chin- ensis group is in all cases very short, the criterion used was the condition of the bud scales at the base of the cone. In stalked forms the scales retain the erect condition of an opening bud above which the stalk may or may not appear, while in the sessile forms the bud scales are flattened out against the base of the cone.

Yunnanensis Group. (i.e. leaf apex entire, cones sessile). Of the four collections thus placed in the Yunnanensis group, three, Ward 1666, Wilson 2098, and 2099, I believe to be the same as T. yunnanensis. The remaining one, Forrest 6747, differs so markedly in the type of cone that I believe it to be a distinct species. The cone scales of T. yunnanensis are finely striate, thin and fibrous, whereas those of Forrest 6747 are smooth, thick and woody in texture.

Chinensis Group. (i.e. leaf apex emarginate, cones stipitate). In this group one specimen, Wilson 2096, stood out very dis- tinctly in the character of the mature cone. The cone scales thick, woody, polished and with truncate apices, open wide at maturity, spreading at anangle little less than go° to theaxis. All other cones in this group and in the previous one, have the scales erecto-patent at maturit

Although the serration on the leaf margins was variable, one or two specimens in this group were found with persistently entire margins. To this sub-division belong Henry 7156, Wilson 2100 and 2097, all of which I believe to be the same as chinensis, and also Forrest 17,169. The latter has a long narrow needle often 25 mm. in length, in contrast to the short broad needle, rarely exceeding 18 mm. of T. chinensis. ‘This together with differences in the cone scale and the colour and indumentum of the twig led me to believe that it was a distinct species.

Leaves of Ward 257 and Maire 203 were generally serrulate on the margin, but there was a marked difference in cone outline and in the size and shape of the cone scale. In Ward 257 the cone is obovate, the cone scales few in number (about 25) measure 14 x 13 mm. while in Maire 203 the oblong-ovate cone is com- posed of a larger number of smaller scales (about 35) 10 x 7 mm. and therefore I believe them to be distinct species.

The young twigs of all the Chinese Tsugas I haveexamined are more or less pubescent, the indumentum occasionally confined to the grooves between the pulvini, more often scattered in the grooves and on the lower part of the pulvini. The leaves, flat

* Excluding 7. Mairez which, ex descr., is probably incorrectly referred to Tsuga.

DownIE—CHINESE SPECIES oF TSUGA. 15

and grooved on the upper surface are either white underneath or concolorous. ‘The one year twigs vary from a distinct yellow to a red-brown colour, the apex of the pulvinus invariably being redder or darker than the rest of the twig.

The anatomy of the leaf did not help very much in the system- atic placing of the species. What differences I have found merely substantiate the formation of the two main groups. A difference in outline is maintained in a general way throughout the groups. Inthe Yunnanensis group there is a deep groove on the upper surface and a slight midrib on the lower surface, while the angles of the leaf are sharp. In the Chinensis group the groove on the upper surface is very slight and there is no midrib on the lower surface while the angles of the leaf are rounded. Excep- tions, however, occur. "The deep groove on the upper surface is absent in some leaf sections of Ward 1666, belonging to the Yunnanensis group (all sections were taken near the middle of the leaf) and sharp angles were observed in serrated species of the Chinensis group, so that variability in outline must be expected. The distribution of the hypodermal cells on the other hand has proved quite a constant factor. In both groups the arrangement on the lower surface is the same, a few cells beneath the resin canal and a few in scattered groups or continuous between the angle of the leaf and the beginning of the stomatic area. The difference is found on the upper surface. In the Yunnanensis group hypodermal cells are practically absent, a few cells sub- tend the deep groove only. In the Chinensis group all species show a large development of hypodermal cells in groups all along the upper surface or forming a continuous layer beneath the epi- dermis.

Key to the Species.

A. Leaf apex entire; cones strictly sessile; leaf margin serrul- ate; branchlets pubescent; hypodermis on upper surface restricted to groove.

I. Cone scale woody in texture, smooth. 1. Tsuga dura. II. Cone scale fibrous in texture, finely longitudinally striate. : : ; ; : ‘ : Ze unnanensis.

B, Leaf apex emarginate; cones shortly stalked; leaf margin serrulate or entire; branchlets pubescent; hypodermis scattered all along upper surface.

TI. Cone scales patent at maturity. : 4.4, paens. II. Cone scales erecto-patent at maturity. a. Leaves generally serrulate on margin. i. Scales up to 14 x 13 mm. about 25 per cone. ‘ . 4. T. Wardu.

4 .

16 DowniE—CHINESE SPECIES OF TSUGA.

ti. Scales up to 10 x 7 mm. about 35 per cone. é . 5+ -T. calcarea, b. heaves alw ave enero on mar i. Leaves up to 18 mm. oe: ‘branchlets yellow. 6.

. chinensis. Leaves up nes 25 mm. long, b bintichlets brown or ‘red-brown. : i 7. T. Forresti.

Enumeration of Species.

1. Tsuga dura, Downie, sp. n. Arbor 30-90-pedalis (ex For- rest) ; ramuli hornotini lutei, indumento sparso pallide brunneo, pulvino subprominente apice fusco-rubro-brunneo subconspicue rimoso alabastro axillari dimidio breviore, annotini teretes, mox cinereo-brunnei et sublaeves ; alabastri squamae setulosae. Folia 6-21 mm. longa, 2 mm. lata, apice obtusa vel rotundata, basi in petiolum angustata, supra sulcata, subtus alba, fasciebus stoma- tiferis duabus e lineis 7-9 constitutis, margine parte dimidia superiore subregulariter denticulata, breviter petiolata. Amenta oO stipite 4 mm. longo suffulta, e ramulis iuvenilibus orta, ag- gregata. Strobili iuventute azureo-purpurei, maturi pallide brunnei, erecto-patentes, obtusi, 20 mm. longi, sessiles, squamis oblongis apice rotundatis vel fere subtruncatis anguste margin- atis minute irregulariter denticulatis rigidis nitidis striis ir- regularibus inconspicuis vel fere deficientibus.

Yunnan. Eastern flank of the Lichiang Range. Lat. 27° 25 N. Tree 30-90 ft. In isolated groups and specimens in mixed forests. Forrest, 6747. Aug. Igro.

2. Tsuga yunnanensis, Masters in Journ. Linn. Soc. vol. XXVI, p. 556 (1902), et vol. xxxvii, p. 421 (1906) pro parte. Abies yunnanensis, Franchet in Journ. de Bot. xiii, p. 258 (1899). Upper Burma, Hpimaw. Ward 1666. Yunnan, Yang in Chan, in woods 2800 m. Delavay ! (type): Lichiang. Delavay 172! Western Szechuan. Wilson 2008, 2 Specimen 172 collected by Delavay in Méniitetia of Lichiang is noted by Franchet as a form of A. yunnanensis with a slightly smaller cone. The material I have examined shows all grada- _ tions in the size of cone from that of Delavay 172 to the type specimen, and I conclude that they are all forms of one species.

3- Tsuga patens, Downie, sp. n. Ramuli novelli luteo- -brunnei, indumento brevi pallido saepissime in sulcis tantum, pulvino rubro-brunneo alabastro axillari dimidio breviore, mox cinerei, vix laeves, pulvino conspicuo apice rimoso inferne am-

DowNIE—CHINESE SPECIES oF TSUGA. 17

pliato. Folia 7-21 mm. longa, 2.5 mm. lata, apice rotundata vel truncata, emarginata, ima basi angustata, breviora integra, longi- ora parte triente superiore serrulata, supra sulcata, subtus saep- issime subalbida, fasciebus stomatiferis duabus e lineis 7-9 constitutis. Strobili iuventute luteo-brunnei, maturi lutei vel castaneo-lutei, ambitu oblongo-ovati vel oblongi, 25-30 mm. longi, breviter sed distincte stipitati; squamae maturae patentes, bractea obovata apice conspicue bilobata 5 mm. longa irregulariter serru- lata dorso puberula, squama seminifera subquadrata margine truncata vel truncato-rotundata 12mm. longa 1omm. lataauguste marginata fere integra parte exserta breviuscula et latiuscula nitida irregulariter distantius longitudinaliter striata, parte ob- tecta dense puberula ; semina ambitu triangularia, 3.5 mm. longa, 2.5 mm. lata, ala 9g mm. longa 4 mm. lata.

Western Hupeh. Wilson, 2096. Type. (Arnold Arb. Ex- ped.) July 1907, and Nov. 1907.

4. Tsuga Wardii, Downie, sp. nu. Arbor 80~r100-pedalis trunco saepe 10-40-ped. supra basem aequaliter bifurcato (ex Ward); ramuli novelli pallide brunnei, indumento brevi sub- denso, pulvino rubro-brunneo subconspicue rimoso alabastro ax- illari dimidio breviore, haud conspicue sulcati, mox supra cinereo- brunnei, infra fusco-brunnei, pulvino per annos 3-4 persistente. Folia omnia immatura, 10-20 mm. longa, 1.5 mm. lata, apice rotundata vel subtruncata, emarginata, basi in petiolum brevem angustata, margine praesertim apicem versus serrata, supra sul- cata, subtus alba, fasciebus stomatiferis duabus e lineis 5-7 con- stitutis. Inflorescentia ( distincte stipitata. Strobili iuventute rubro-purpurei, maturi erecto-patentes, ambitu oblongo-obovati, 21 mm. longi, stipite brevi squamis persistentibus obtecto suf- fulti, bracteis apice bifidis; squamae seminiferae obovatae vel rotundato-obovatae, apice truncatae vel subtruncato-rotundatae, 14 mm. longae, 13 mm. latae, anguste et haud conspicue margin- atae, fere integrae, parte exserta maiuscula luteo-brunnea nitida longitudinaliter irregulariter corrugata, parte obtecta dense pu- berula sulcata.

N.W. Yunnan and E. Tibet. Chung valley, 9—10,000 ft. also in mixed forests of Doker-la, Mekong-Salween Divide. Lat. 28°. Ward, 257, May 1913.

5. Tsuga calcarea, Downie, sp. n. Ramuli iuventute - pal- lide rubro-brunnei, indumento brevi pallido subdenso, pulvino apice rubro-brunneo, annotini haud conspicue sulcati, mox fusco- grisei vel griseo-brunnei. Folia 6-20 mm. longa, 2 mm. lata, apice truncata vel subtruncata, emarginata, apud basem angus- tata, breviora integra, longiora saepissime serrulata, pagina

18 DowNIE—CHINESE SPECIES OF TSUGA.

utraque concoloria vel inferiore subalbida superiore sulcata, fasciebus stomatiferis duabus e lineis 7-9 constitutis, breviter petiolata. Strobili maturi erecto-patentes, pallide brunnei, am- bitu oblongo-ovati, 21 mm. longi; squamae seminiferae oblongae, apice rotundatae vel subtruncatae, angustissime marginatae, haud recurvae, fere integrae, parte exserta brevi lata pallide brunnea nitida irregulariter longitudinaliter sulcata.

Yunnan. Mi-monts arides, calcaires 4 Tché-hay, altitude 2600 m. Grand et superbe résineux a aubier rose, fruits trés petits, dressés-rameaux tendant A la verticale (rare). Maire, 203. Octobre.

6. Tsuga chinensis, Pritz. in Engler’s Bot. Jahrb. vol. xxix, p- 217 (1900); Masters in Journ. Linn. Soc. vol. xxvi, p. 556 (1902) et vol. xxxvii, p. 421 (1906).

T. Sieboldi, Masters, |.c. vol. xxvi, p. 556, saltem pro parte, non Carr.

T. yunnanensis, Masters, l.c. vol. xxxvii, p. 421, saltem pro parte, non Carr.

A bies chinensis, Franchet in Journ. de Bot. xiii, p. 259 (1899).

Szechuan, near Tchen keou tin, altitude 2500 m. Farges 808!

(type).

Szechuan. Henry 7156.

Western Szechuan. Wilson 2097, 2100. 7. Tsuga Forrestii, Downie, sp. n. Arbor 30-80-pedalis (ex Forrest); ramuli primo brunnei vel rubro-brunnei, indumento brevi pallido, pulvino rubro-brunneo nitido inferne gradatim pal- lescente ramulis annotinis subprominente infra apicem margin- atum ampliato, mox supra griseo-brunnei, infra rubro-brunnei, cortice transverse fisso pulvino subconspicuo. Folia 10-25 mm. longa, 2 mm. lata, apice rotundata, emarginata, ima basi in petiolum angustata, integra, concoloria, supra sulcata, subtus fasciebus stomatiferis duabus e lineis 6-8 compositis instructa, breviter petiolata. Strobili lutei vel pallide luteo-brunnei, am- bitu ovati vel oblongo-ovati, 25-30 mm. longi, breviter stipitati; bracteae ovatae, apice breviter bilobatae, 4 mm. longae, irregul- ariter denticulatae, breviter petiolatae, dorso puberulae, inferne margine ciliatae ; squamae seminiferae truncato-rotundatae, dorso convexae, basi auriculatae, 12 mm. longae, 13 mm. latae, longi- Sadinalitas subregulariter striatae, marginatae, integrae, parte exserta nitida glabra luteo-brunnea, parte inclusa dense puberula. Semina 3 mm. longa, ala 10 mm. longa.

Yunnan, Lichiang range. Lat. 27° 30/N. ‘Tree of 30-80 ft. in pure and mixed forests 10-13,000 ft. Forrest, 17,169. Nov. 1918.

DownIE—CHINESE SPECIES OF TSUGA. 19

NotE.—Very similar to this species is Maire 692, 1914. Grand et superbe résineux (rare). Fl. jeunes. Coteaux arides a Siao-ou-long. Alti. 2550 m., Juin.

Species Exclusae vel Dubiae.

Tsuga Brunoniana, Masters in Journ. Linn. Soc. vol. XXXVl, P. 421 (1906), an Carr. ?

Tsuga Brunoniana, Carr., var. chinensis, Masters in Journ. Linn. Soc. vol. xxvi, p. 556 (1902). Abies dumosa, Loud., var. chinensis, Franchet in Journ. de Bot. xiii, p. 258 (1899).

Tsuga Sieboldii, Masters in Journ. Linn. Soc. vol. XXXVli, p. 421 (1906), an Carr. ?

Tsuga Mairei, Lemée et Léveillé in Le Monde des Plantes, vol. XVi, p. 20 (1914). Ex descr. this species is probably incorrectly referred to Tsuga. There is no specimen in the Léveillé herb- arium in Edinburgh.

In conclusion I wish to thank Professor W. G. Craib at whose suggestion the study was undertaken and to whom I am indebted for helpful criticism throughout the work and in the preparation of the manuscript. Acknowledgment is also made to the officials of the herbarium of the Paris Museum for permission to examine type specimens and to Professor W. Wright Smith for the pre- paration of the plate. I should also like to express my indebted- ness to the late Professor Sir I. Bayley Balfour for the use of the Edinburgh herbarium material.

EXPLANATION OF PLATE CXCIV. Illustrating Miss Downie’s Paper on Chinese Species of Tsuga.

Cone of Tsuga patens, Downie—Wilson, 2096. Cone of Tsuga Forrestii, Downie—Forrest, 17,169.

SIO ww DH % SB o g — “a oe R ‘ = a = y a 8 a t i] a S)

A New Chinese Osmanthus. BY

ALFRED REHDER, Arnold Arboretum, Harvard.

Osmanthus Forrestii, sp. nov.

Frutex 3-6-metralis ramulis junioribus pallide griseo-flavidis glabris ; gemmae perulis late ovatis acuminulatis glabris margine ciliato excepto. Folia persistentia chartacea, ovato-lanceolata,

vel oblongo-lanceolata, basi late cuneata vel rotundata,

acuminata, integra vel spinoso-dentata, dentibus utrinque 18-20 erecto-patentibus satis angustis basi 1-2 mm. latis, glabra, supra nigro-punctulata, utrinque nervis 13-18 supra ut costa leviter, subtus magis elevatis, reticulo nervulorum subtus praecipue in foliis dentatis conspicue elevato, supra minus vel interdum saltem in foliis integris vix elevato; petioli validi, 5-7 mm. longi, glabri. Flores cremeo-albidi, fragrantes, fas- ciculati, e gemmis axillaribus; pedicelli graciles, circiter I cm. longi ; lg brevissimus irregulariter vel interdum indistincte 4-lobatus lobis latioribus quam longis eroso-denticulatis vel integ- ris; corolla circiter 5 mm. longa, fere ad basin 4-partita lobis el- lipticis; stamina paullo supra basin inserta corolla paullo breviora; antheris ovato-ellipticis basi cordatis apice obtusis, filamentis brevissimis. Fructus non visus.

Cuina : Northwestern Yunnan; mountains west of Yungspels, Lat. 26° 30’ N., Long. 100° 50’ E., altitude 8—gooo ft., amongst scrub on open rocky hillsides and in open mixed forests. G. Forrest, No. 21,050 (spinous-leaved shrub of 9-20 ft.; flowers sweetly fragrant, creamy white).

This new species seems most closely related to O. armatus, Diels and O. serrulatus, Rehder; the former is easily dis- tinguished by the minute pubescence of the winter-buds, young branchlets and petioles, by the comparatively narrower leaves not reticulate beneath and with fewer and broader spiny teeth, and by the less deeply divided corolla; O. serrulatus differs chiefly in its cuneate leaves usually broadest above the middle, not reticulate beneath and with smaller and more numerous teeth pointing for- ward, in the slenderer and longer petioles and pedicels, in the ‘Linger calyx and in the narrower anthers on longer he ati

Wvotes, R.BG., Edin. No. LXVII. April.]

Polyembryony in Sarcococca ruscifolia, Stapf. BY

. MATTHEW YOUNG ORR, Assistant in Laboratory, Royal Botanic Garden, Edinburgh.

With two figures in the text.

THE number of Angiosperms in which Polyembryony has ap- peared as an occasional or habitual feature of their reproductive processes is comparatively small, and present-day botanical literature is singularly barren of further additions to the list. A new case of polyembryony is therefore of interest, more par- ticularly if it be found in a Family which has not previously furnished any example of the phenomonon. Such a case has — recently come to light in the embryogeny of Sarcococca, and, so far as is known to the writer, this is the first mention of a de- parture from the normal mode of embryo-formation in a member of the Buxaceae.

While examining fruits of Sarcococca ruscifolia, which had been brought to the Botanic Garden Laboratory from an estate in East Lothian, it was discovered that fully one half of the ripe seeds contained more than one embryo, there being usually, in each seed, two fully-formed embryos with radicle, plumule and cotyle- dons, and a varying number of minute, immature embryos in various stages of development.

Fic. 1.—Two embryos from a seed of Sarcococca ruscifolia. X 10.

The only apparent distinction between the two mature embryos taken from any one of these seeds was a difference in size, and it

(Notes, R.B.G., Edin., No. LXVII, April 1923.]

22 ORR—POLYEMBRYONY IN SARCOCOCCA RUSCIFOLIA, STAPF.

is a matter for conjecture whether both of these embryos might not have developed, pari passu, into seedlings had the seed been allowed to germinate. Two such embryos are shown in fig. I.

A further supply of seed was obtained from plants of the same species growing in the Royal Botanic Garden, and these were also found to be polyembryonate; in one case, as many as seven embryos, showing different degrees of development, being isolated by dissection from a single seed.

The problem of the origin of these several embryos was not one which could be solved satisfactorily by an investigation of the material then available; but, by removing the endocarp, and em- bedding the seed in paraffin, it was possible to obtain a series of sections in which the relative positions of the embryos, and their relationship to the tissues within the seed, could be demonstrated. The anatomical features disclosed by these preparations can best be illustrated by reference to fig. 2, which is a camera lucida drawing of a longitudinal section through the micropylar region of a ripe seed.

S Bare tanec

=

Fic. 2.—Longitudinal section through the micropylar region of a seed of Sarcococca ruscifolia, a,a, mature, and b,b, immatu bryos. x about Loo.

In the upper half of the figure, there appear, in section, the radicle apices of two embryos similar to those represented in fig. 1, while, in the lower, micropylar area, intermixed with the dis- organised tissue, there can be seen a number of distinct cell-

masses which are the immature embryos referred to above.

ORR—POLYEMBRYONY IN SARCOCOCCA RUSCIFOLIA, STAPF. 23

From the figure, it will be seen that the position of these immature embryos in the disintegrating tissue at the apical end of the seed is somewhat suggestive of their genesis in the nucellus of the ovule at an earlier period of growth. This derivation of embryos from sporophytic tissue outside the embryo-sac is by no means an uncommon feature of certain forms of polyembryony, and has been observed in both Monocotyledons and Dicotyledons by various investigators.

In the Euphorbiaceae, for example, it has been established by the researches of Strasburger* on Coelebogyne ilicifolia (Al- chornea), and by the later work of Hegelmaier+ on Euphorbia dulcis, that embryos produced by these two polyembryonic species have their origin in the nucellus. Bearing in mind the close re- lationship existing between the Euphorbiaceae and Buxaceae, it might be reasonably inferred that the undifferentiated embryos in the seed of Sarcococca have arisen from the same ovular tissue.

Whether the fully-developed embryos shown in fig. 1 originated in like manner, or in some other way, is a debatable point at this stage. Absolute proof regarding the morphological origin of both the mature and immature embryos in the seed of Sarcococca ruscifolia can only be attained by a study (which is now being prosecuted) of the whole process of embryo-building throughout the entire period of maturation. All else is pure speculation.

* EF. Strasburger in Zeitschr. Naturwiss. Jena, xii ia P. 659. + F. Hegelmaier in Ber, Deut. Bot. Gesell., xxi (1903),

On the Resin Ducts in the Leaf of Picea brachytyla, Pritzel.

BY

MATTHEW YOUNG ORR, Assistant in Laboratory, Royal Botanic Garden, Edinburgh.

With one figure in the text.

Picea brachytyla, Pritzel*, one of the flat-leaved Spruces of Western China, has now been in cultivation for a number of years ; but, as all who have studied Conifers will admit, it is by no means an easy matter to identify specimens of this Picea, with any degree of certainty, by its vegetative characters alone. The difficulty of separating this species from Picea complanata is a case in point, and it is recorded that Franchet himself referred a specimen of the latter to his “‘A bies brachytyla.”’

This difficulty of accurately placing a non-fruiting specimen is not confined to Picea brachytyla alone, but, if the unusual anatomical feature which has come under the observation of the writer, and which forms the subject of this note, should prove to be constant in the species, this problem of identification, in the case of Picea brachytyla at least, ought to be the more easily solved.

In the leaves of Piceas, belonging to the Omorica section, there are, as is well known, two laterally placed, marginal resin ducts, which extend from the base of the leaf throughout its whole length, terminating abruptly immediately behind the apex. These resin passages are in contact with, or may be partially enclosed by, the hypoderm of the under side of the leaf.

The more minute details of their morphology need not be dis- cussed here, since the chief purpose of this paper is to call atten- tion to the diagnostic feature referred to above, which has apparently been overlooked in the past, since no reference to it can be found in the literature appertaining to this species.

Some little time ago, while examining microscopic preparations of the leaves of Picea brachytyla, grown in the Royal Botanic Garden from seed collected in Western Hupeh by the Veitch Expedition in 1900, it was noticed that in transverse sections, cut from near the base of the leaf, there were four resin ducts visible,

* See Plantae Wilsonianae, vol. ii (x914), p. 33.

_ R.B,G,, Edin., No, LXVII, April 1923.]

ORR—LEAF OF PICEA BRACHYTYLA, PRITZEL. 25

in place of the usual two, and that these were in pairs, occupying the normal position belo the lower epidermis (Fig. 1a). The outer duct of each pair was slightly smaller than the inner one, and both were surrounded by moderately thick-walled cells, con- tinuous with the hypoderm. These “‘twin’’ resin ducts gave to the leaf section a very striking appearance, forming a marked contrast to the normal arrangement found in all other flat-leaved Spruces.

Ns apace

2o9

] a2 ws Se BOSS 1,00

Fic, 1,—(a) Diagram of a Fonrreni age see of a leaf of Prcea brachytyla, showing the pai resin ut 3 (b) Transverse a of a resin aoer: showing the origin of the secondary duct. s, the septum.. x a

In order to test the constancy of this feature, and thereby form an estimate of its possible value as a specific indicator, similar preparations were made from the leaves of a herbarium specimen of a native plant of brachytyla, collected by Wilson in Western Hupeh, and bearing the number 1896. Here, as in the cultivated plant, a similar duplication of the resin ducts was observed. An examination of the leaves of a fruiting specimen grown in Ireland furnished additional evidence of the universality of the phen- . omenon, while indicating further that it is not confined entirely to the leaves of non-fruiting branches.

Unfortunately, it was not possible to extend the investigation to include the type and co-types of the species, as these were not available ; but in the leaves of all plants that were examined, the doubling of the resin passages had taken place, forming an unmis- takable feature of the transverse section.

The morphological origin of these secondary resin ducts is interesting and has been worked out in detail by means of serial sections. In those sections cut across the leaf immediately above its constricted base, only two resin passages were visible; but at a point about 1 or 2 mm. distant from the base, the outer, secondary ducts take their origin from the primary ducts by a process of division. ‘The first indication of this division is the production of a partial septum, formed by the elements of the epithelial layer, which have extended inwards at two points. By

the subsequent growth in length of the cells composing the two

26 ORR—LEAF OF PICEA BRACHYTYLA, PRITZEL.

portions of this incomplete septum, the projecting parts become united, and the division of the resin passage longitudinally is ac- complished. ‘This stage is represented in Fig. rb.

The paired resin ducts, separated only by the septum, remain in close contact with each other, running parallel through the leaf for about one third of its length; but beyond this point, the secondary ducts diverge in a lateral direction, away from the parent ducts, and in successive sections cut through this region of the leaf, it is possible to trace their course outwards to the margin, where they ultimately die out, about midway between the base and apex of the leaf.

The two main resin passages are continued almost to the apex, and transverse sections of the upper half of the leaf present the normal arrangement seen in the leaves of other flat-leaved Piceas, except that in one case a small lateral duct was observed near the apex, midway between the main duct and the leaf margin, but it only extended for a very short distance.

Why the leaves of Picea brachytyla should possess this peculiar feature of paired resin ducts it is not easy to say; but it is a structural phenomenon which has every appearance of being con- stant in the species, and, so far as is known, it is not found in any other species of Picea.

While it is recognised that in many plants anatomical differ- ences have a questionable taxonomic value, yet in the genus A bies, the position of the resin canals in the leaf is often made a deciding factor in the identification of a species, and it is suggested, on this analogy, that the unusual arrangement of the resin ducts in the leaf of Picea brachytyla might also be used with advantage as an aid to diagnosis in doubtful cases.

Note on the Varieties of Geum elatum, Wallich. BY

W. EDGAR EVANS, B.&c., Assistant in Herbarium, Royal Botanic Garden, Edinburgh.

With Plates Nos. CXCV and CXCVI.

Wui.e working out certain Chinese specimens, which had been regarded as forms of the well-known Himalayan Geum elatum, Wall., I had occasion to examine somewhat critically a consider- able number of authentic examples of that species. In the hope of helping to elucidate its varieties, the results obtained are here set down.

The sub-genus Sieversia of the genus Geum, distinguished by the permanent attachment of the whole style to the achene, is represented in the Himalaya by two alpine species only, which may be recognised as follows :—

Radical leaves having well-developed lateral leaflets; the terminal leaflet or lobe not greatly exceeding the upper lateral ones, not cordate. : ‘ : . G. elatum, Wall.

Radical leaves having only very much reduced lateral leaflets ; the terminal leaflet or lobe very greatly exceeding the lateral ones, deeply cordate. : . G. sikkimense, Prain.*

Of these, the latter seems to be both rare and local, having apparently been collected only on three occasions, always in Sik- kim ; the former, on the other hand, has a wide range throughout alpine Himalaya, extending from Kashmir to Sikkim at least ; while Dr L. Diels in his ‘‘Flora von Central China’’} states that the form of it known as var. humile, Royle is distributed through China from Kansu and Shansi to Yunnan, and Franchet records that it was collected by Delavay in the latter province.

The Chinese distribution, just referred to, seems to be based, largely at least, on misidentification and will in consequence re- quire careful revision.§ I have, it is true, seen only three Chinese gatherings of reputed Geum elatum, all from Yunnan

* Journ. As. Soc. Beng., Ixxiii (1914), P. 200, PL 7. + Engler’s Jahtbuch., xxix (1900), p. 404.

+ Plantae Delavayanae, p. ‘

§ See Cardot in Notul. System. Paris, iii (1916), p. 225. (Notes, R,B.G., Edin., No, LXVII, April 1923]

28 Evans—GeEuM ELATUM, WALLICH.

and only one of them vouched for by Dr Diels*, but since all these specimens have ascending ovules and basally deciduous styles and are thus referable to the genus Coluria not Geum, the area, if any, occupied in China by the species now under consideration must, for the present, be regarded as doubtful.

In the account of the varieties of Geum elatum, Wall. which follows, it will be seen that the type number Wall. Cat. 711 covers two of these. From the Catalogue itself it would appear that two distinct gatherings, one from Kumaon and one from Sirmore, were distributed under this number and it is therefore likely that all the specimens from one of these localities were what I here name var. typicum, those from the other being the full-statured form with almost glabrous achenes for which I propose the varietal name leiocarpum. Unfortunately, it-is quite impossible to say from which locality any given specimen originated, but it is worthy of note that the whole of the type sheet from the Kew Herbarium (Plate No. CXCV), kindly lent me for examination, is the former, while the whole of that in the Edinburgh Herbarium is the latter. Owing to this uncertainty it has not been possible to include either in the distributional notes, though the specimens in question will form the types of the respective varieties should further investigation prove their validity.

With regard to the third variety (var. humile, Royle) I am of the opinion that it may prove merely a stunted state of var. leiocar- pum. ‘The plant from Gossain Than, distributed as Wall. Cat. no. 712, may be regarded as the best available example of Geum elatum, Wall. var. humile, Royle as understood by Sir J. D. Hooker.+ The only description of this, which does more than refer to its reduced size and generally solitary flowers, is that of Lehmann.} From this we learn that the few Nepalese specimens in question were in fairly advanced condition, the petals being much withered or fallen and that, in the portion received by him, the inflorescence did not exceed the leaves and the achenes were glabrous. An examination of the single representative speci- men in the Kew Herbarium (Plate No. CXCVI) shows that it agrees with Lehmann’s account, except that the inflorescences are almost twice the length of the leaves and the achenes are tipped with a very few, short, bristly hairs. From this it seems clear that var. humile, Royle has achenes which show just the same characters and degree of variation as do those of var. leiocarpum and in addition that its stature is subject to at least some variation towards the taller-growing forms of the species. In this connec- tion it should be borne in mind that almost equally reduced forms with the — sabe ee of var. typicum also occur, as an

. Edinb. vii (1912), p. 108.

y. Bot. G: + Flore a British india. ii, P- 343- ; Revis. Potentill., p. 47, t. 17.

Evans—GEUM. ELATUM, WALLICH. 29

example of which I may cite a Sikkim specimen in the Edinburgh Herbarium (W. W. Smith No. 3451). For the reasons above detailed, then, I am inclined to regard var. humile, Royle as merely a reduced state of the tall-growing var. letocarpum; I have, however, retained it pending fuller investigation.

I wish here to express my thanks to the Director of the Royal Botanic Gardens, Kew, for the loan of the type specimens figured and to Professor Wright Smith for help in various ways.

Enumeration of the Varieties of Geum elatum, Wall. with Synonymy, etc. Varietates tres sequentes fortasse ita sunt distinguendae :—

1. Var. « typicum W. E. Evans (Plate No. CXCV) ; foliorum pinnis lateralibus haud approximatis, caulibus floriferis elongatis vulgo 2-6-floris, carpellis dense hispidis recognita.

Geum elatum, Wall. Cat. no. 711 p. p.; Hook. f. Flor. Brit. Ind., ii, p. 343 (excl. var. humile).

Sieversia elata, Royle IIl., p. 207 (excl. var. humilis), t. 39.

ALPINE Himalaya; generally distributed. Kashmir, J. F. Duthie s.n.!; Chamba, J. H. Lace no. 1233! Bashahr, J. H. Lace no. 314!; Sikkim, W. W. Smith no. 4289!

2. Var. @ leiocarpum, W. E. Evans; quae cum varietate a foli- orum atque inflorescentiarum forma quadrat, carpellis glaberrimis vel apicem versus parcissime hispidis differt.

Geum elatum, Wall. Cat. no. 711 p. p.

ALPINE HIMALAYA; apparently much less frequent than var. a. Ihave seen only two specimens, one distributed under Wall. Cat. no. 711 and another collected by J. R. Reid, Esqre., C.I-E. ; both are in the Edinburgh Herbarium and have almost or quite glabrous achenes. Unfortunately, in each case, the place of origin is uncertain.

_Var. y humile, Royle (Plate No. CXCVI); foliorum pinnis approximatis, catlibus floriferis 1~-(rarissime 2-) floris folia haud vel paulo excedentibus, carpellis glabris vel subglabris ab var. typico divergit.

Geum elatum var. humile, Hook. f. Flor. Brit. Ind. 11, p. 343-

Geum adnatum, Wall. Cat. no. 712.

Sieversia elata, var. humilis, Royle Il., p. 207.

Sieversia humilis, Royle MSS, apud Royle l.c.

Potentilla adnata, Wall. MSS. cae Wall. Cat.; Lehm. Pugill. ix, p. 9; Revis. Potentill., p. 47, t. 17.

ALPINE HIMALAYA oNeesl, Wallich no. ae Sikkim, W. W. Smith no. 3174!

30 Evans—GEuM ELATUM, WALLICH.

Key to the Varieties of Geum elatum, Wall.

Lateral lobes of the radical leaves generally distant, the reduced alternate ones ae numerous; flowers 2-6; plant tall- growing (Plate No. CXCV

Achenes densely hispid. var. a typicum. Achenes glabrous or very sparsely iepld above var. 8 Poeeron

Lateral lobes of the fadical fedves more 5 clese-cat: the reduced al- ternate ones fewer; flowers ere eplitacy: ; plant low- growing. (Plate No. CXCVI). : var. y humile.

EXPLANATION OF PLATES. Illustrating Mr Evans’ Paper on the Varieties of Geum elatum, Wall. (Photographs by R. M. Adam, Assistant in Studio, Royal Botanic Garden, Edinburgh.)

Pirate No. CXCV. “e elatum, Wall. — typicum, W.E. Evans, Wall. Cat ores "Herb. Whole plant x y Showitar form of ‘Teaves and fiihoresbente CXCVI.—G. a Wall. var. humile, Royle. eg 0 natum in Herb. Kew Who in ali Sout Psy eeieg pe tte form of leaves fe gt marten ry

(x0o8D Wt 1x105/S.0.P.177 375 8/23 J.P G7

NOTES, R.B.G., EDIN. PLATE CXCIV

Chinese Species of Tsuga—Cones,

NOTES, R.B.G., EDIN. PLATE CXCV.

Geum elatum, Wall, var, typicum, W. E. Evans,

NOTES, R.B.G., EDIN. PLATE CXCVI.

f

Geum elatum, Wall, var. humile, Royle,

New Primulaceae BY Proressor WILLIAM WRIGHT SMITH AND GEORGE FORREST.

‘Tux following new species and varieties are described along with notes on a few previously published :

Primula Agleniana, Balf. f. et Forrest, var. alba, Forrest (Nivalis),

p- 32. aromatica, W. W. Sm. et Forrest (Malvacea), p. 32. atrotubata, W. W. Sm. et Forrest (Malvacea), p. 33. aurantiaca, W. W. rie $ Forrest eraser p- 34. brachystoma, W. W. Sm. (Candelabra), p. 35 caldaria, W. W. Sm. et : Hottest (Farinosae), p. 35- Pager W. W. Sm. et Forrest, var. nana, W. W. Sm. et

rest (Farinosae), p. 36.

caitianttie: Franch. var. pe Farrer (Nivalis-Calliantha),

Pp. 39. Coryana, Balf. £. et Forrest (Nivalis), p. 37- pea Watt, var. Pantlingii (King) (Obtusifolia?),

Dickieana, Watt, var. chlorops, Forrest (Obtusifolia ?),

p- 3 aumicola W. W. Sm. et Forrest (Obconica), p. 40. a, W. W. Sm. et Forrest (Malacoides), p. 40. praca ti W. W. Sm. et Forrest (Geranioides), p. 41. cone W. W. Sm. et Ward (Petiolaris), p. 42. V. W. Sm. et Perseit Wibererdt Pp. 43-

malvacea, Franch. var. alba, Forrest (Malvacea), p

peer dapine Franch. var, intermedia, Forrest ( F seesaeete )y

P- 4°. melanops, W. W. Sm. (Nivalis), p microstachys, Balf. f. et Forrest Pity acea), p. & monanth . W. Sm. et Forrest (Macrocarpa), p oxygraphidifolia, W. W. Sm. et Ward (Aorieulate, P. 50. Pauliana, W. W. Sm. et Forrest (Geranioides), p pseudosikkimensis, Forrest (Sikkimensis), p. 52.

Notes, R.B.G., Edin., No. LXVIII, November 1923.]

Ivo

\2tt

32 Smitru—Nrw PRiIMULACKAR.

Primula tapeina, Balf. f. et Forrest (Suffruticosa), p. ee Valentiniana, Hand.-Mzt. (Amethystina?), p aK ma haan bs , W. W. Sm. et Ward ( (Muscarioides), Pp. 54. ingtonensis, Forrest (Obconica), p. 55. iisioorien Sots Forrest, p. 55. ‘ s, Hand.-Mzt., p. 56.

Primula Agleniana, Balf. f. et Forrest var. alba, Forrest.

A typo floribus albis variat; in specie nobili fortasse men- tione digna. :

“S.E. Tibet :—Salween-Kiuchiang divide, Tsarong. Lat. 28° 24’ N. Long. 98° 24’ E. Alt. 15,000 ft. Plant of 12-16 inches. Flowers fragrant, white, faintly flushed rose exterior. Moist alpine pasture on the margins of cane brakes. July 1921.”’ G. Forrest. No. 20081.

It is possibly worth recording that the typical P. Agleniana was found in July 1905 on the rie ea, Page ikist aan sae at an altitude of 11~-13,000 ft., Lat. 27° 28’ N., under Forrest No. 479, an incomplete apchiien which remained for long uniden- tified. See Notes R.B.G., Edin. VI. (1912) 3

Primula aromatica, W. W. Sm. et Forrest.

Species valde distincta cujus affinitas cum sectione Malvacea esse videtur; foliis in vivo carnosulis aromaticis, in sicco deli- catissimis, scapo floribusque gracillimis ab affinibus facile cognoscitur.

Planta 7-12 cm. alta calcicola caespitosa glebam parvam formans radicibus permultis gracilibus. Folia pauca _petiolo 3-5-5 em. longo flexuoso pilis longis albidis densius ornato prae- dita; lamina plerumque 2.5~3.5 cm. longa atque fere aeque lata, late ovata, ad quartam partem 7-9-lobata, lobis rotundatis et nunc trllobiilatis: apice rotundata, basi Breviter cordata vel subtruncata, textura in sicco tenuissima, sparsim pilosula vel utrinque fere glabra, margine pilis placate pitti minutis ciliolata. Scapus ad 12 cm. altus folia plerumque bene superans pilis ut petioli indutus inflorescentiam 1~5-floram spiciformem gerens; intervalla inter flores 5-10 mm.; bracteae 5-7 mm. lon- gae lineares vel e basi petioliformi lanceolatae vel anguste oblan- ceolatae, ad basim pedicelli positae vel nunc sub calyce subad- natae pilis glanduliferis minutis ciliolatae; pedicelli brevissimi

Vix I mm. superantes. Calyx circ. 6 mm. longus foliaceus fere

ad imum in lobos lanceolatos acutos fissus glanduloso-pilosulus.

- Corollae tenuissimae roseae circ. 1.5 em. dougee tubus_ cylin-

Y

SmMitHu—NEW PRIMULACEAE.

r

. we

- >

dricus 1 cm. longus flavidus pilis minutis conspersus, lobi anguste obovati emarginati glabri. Stamina in flore longistylo ad quadrantem tubi corollini inferiorem inserta antheris vix I mm. superantibus, in flore brevistylo paulo infra os posita. Ovarium subglobosum; stylus glaber in flore longistylo ad os pertinens, in flore brevistylo ad quadrantem tubi attingens; stigma capitatum. Fructus globosus apice valvis 5 dehiscens seminibus numerosis angulatis circ. 0.5 mm. longis.

“West China :—Goo-du Shan, Yunnan. Lat. 28° 16’ N. Long. 100° 12’ E. Alt. 9-10,000 ft. Plant of 3-5 inches. Flowers rose, tube green. Foliage fragrant. Moist situations on limy cliffs and slopes. July 1921.’ G. Forrest. No. 20568.

_A most remarkable Primula for tenuity of foliage and flower —in the dried state at least-—for both have the fragility asso- ciated with dried flowers of Impatiens. It is placed with some doubt in the series Malvacea, but the alternative seems to be a new section.

\) Primula atrotubata, W. W. Sm. et Forrest.

Species szechuanica ex affinitate P. malvaceae, Franch. a qua habitu minore, floribus saturate roseis, corollae tubo atropur- pureo lineato, oculo kermesino inter alia divergit.

Planta 10-20 em. alta. Folia petiolo laminam subaequante vel saepe multo breviore pilis albidis patentibus longis bene induto munita; lamina 3-7 cm. longa, 2.55 em. lata, subovalis vel late ovata, apice rotundata, basi breviter cordata, margine grosse crenata atque ciliolata, textura in sicco tenuis, supra pilis conspersa, subtus ad costam nervosque dense pubescens, ceterum hie illic glabrescens. Scapus ad 20 cm. altus dense pubescens umbellas 2-3 superpositas in modum P. malvaceae gerens, bracteae 6-8 mm. longae lineari-lanceolatae foliaceae pubescentes ut pedicelli S13 mm. longi. Calyx 1 cm. saepe superans post anthesim accrescens, late cupularis pubescens atque glandulosus ‘in lobos late ovatos acutos ad duas partes fissus. Corollae saturate roseae tubus anguste cylindricus calycem aequans vel nunc paulo superans infra medium paululo constrictus atropur- Pureus lineatus extra sparsim pubescens, annulus conspicuus, lobi 6-7 mm. longi obovati alte bilobulati. Stamina in flore longistylo ad tubi corollini partem constrictam inserta. Ovarium globosum parvum; stylus in flore longistylo vix ad os pertinens ; stigma capitatum. Fructus haud maturus calycis tubo bene inclusus

“West China :— Mountains east of Yung-ning, S.W. Szechuan. Lat. 27° 50’ N. Long. 100° 56’ E. Alt. 10-11,009

4. “Plant of 4-8 inches. Flowers deep rose with crimson eye. Dry stony meadows. July 1921.”

34 Smirn—Nrw PRIMvLACEAL.

“Muli, S.W. Szechwan. Alt. 6-7ooo ft. Flowers bright rose-pink, striped outside. In the bed of the Litang river under shrubs and in thick growth, well shaded. Calcareous soil. A species in which the calyx is larger than the flower! A very late flowering species. Though more advanced on 16.7.21 it was clearly not yet in full bloom. When properly out, it is a beautiful plant, the rich rose-pink and dark eye show ing up against the foliage in striking manner. July rgar.’’ F. K Ward. No. 4351.

P. atrotubata is closely akin to P. malvacea, Franch. but we are unable to match it with any of the numerous collections ot that species from Yunnan. When in flower the dark-coloured corolla-tube with longitudinal lines and the crimson eye serve for discrimination.

Ua Primula aurantiaca, W. W. Sm. et Forrest.

Species e grege Candelabra, habitu P. burmanicae, Ball. cet Ward; a speciebus ejus_ sectionis flavidifloris calycis forma lobisque facile distinguitur ; scapus fructiferus saepe decumbens apice gemmas foliosas profert.

Planta ad 25 cm. alta. Folia sat numerosa rosulata petiolata, majora 12-16 cm. longa, 2.5-3.5 cm. lata, oblanceolata apice plus minusve rotundata margine eroso-denticulata, deorsum in petiolum lamina duplo-breviorem alatum vaginantem sensim attenuata, in sicco tenuiter papyracea, glabra, utrinque nervis in sicco conspicuis, costa media lata; superficies infra pallidior obscure pulverulenta. Scapi 1-2 satis robusti ad 25 cm. alti glabri umbellas plerumque 6~12-floras 2-6-superpositas inter- vallis 2-3 cm. longis gerentes; bracteae g-11 mm. longae lineares ad basim pulverulentae; pedicelli 6-8 mm. longi sub fructu erecti; scapi fructiferi nunc decumbentes gemmas bene foliosas radicantes proferentes: Flores saturate rubro-auran- tiaci. Calyx anguste campanulatus circ. 7 mm. longus extra’ subatratus intus viridis in lobos 5 anguste oblongos 4 mm. longos apice acutos vel nunc obtuse 2~-3-lobulatos divisus. Corollae tubus cire. 12 mm. longus angustus annulatus, lobi 4 mm. longi anguste obovati vel suboblongi emarginati. Antherae in flore brevistylo ad os corollae insertae. Ovarium globosum ; stylus brevis calyce brevior, longus ad os corollae pertinens ; stigma magnum capitatum. Capsula globosa ad 5 mm. diametro e tubo calycis ampliati paulo exserta seminibus 1 mm. diametro - angulatis ah papillatis.

“West China :—Chien Some Eekong divide, Yunnan. Lat. 26° a N. Long. 99° 40’ E. Alt. 11,000 ft. Plant of 7-10 inches. Flowers deep ruddy-orange. Moist alpine pasture

Smitu—Nerw PRIMULACKAL. 35

by streams. July 1922.’’ G. Forrest. Nos. 21497, 22318 (Sept. 1922 in fruit).

The following numbers are conspecific, representing a form with greener calyx and paler corolla but otherwise agreeing with the typical plant even in the presence of occasional rooting buds at apex of the scape :—

Same locality. Alt. 11-12,000 ft. Plant of 8-12 inches. July 1922. G. Forrest. Nos. 21501, 22287 (Sept. 1922 in fruit).

\{//Primula brachystoma, WW. W. Sm.

Species ex affinitate P. serratifoliae, Franch. a qua foliorum forma atque denticulatione, calycis lobis brevibus, corollae tubo aeque et anguste cylindrico, lobis corollinis brevibus subquad- ratis inter alia divergit.

Planta ad 20 cm. alta efarinosa epilosa. Folia petiolo vix discreto alato brevi suffulta; lamina 6-12 cm. longa, 1.5-2 cm. lata, oblanceolata vel sublanceolata, apice angustata atque breviter apiculata, basi sensim in petiolum decurrens, margine denticulis brevissimis induratis praedita, in sicco tenuiter Papyracea, glabra. Scapus ad 20 cm. longus gracilis glaber umbellam simplicem 2-6-floram gerens; bracteae ad 6 mm. longae basi latiusculae lineari-lanceolatae acutae glabrae; pedi- celli 5-10 mm. longi; flores plus minusve cernui. Calyx circ. 5 mm. longus anguste campanulatus glaber submembranaceus pallido-viridis lobis vix 1 mm. superantibus triangularibus indurato-apiculatis. Corollae flavae annulo valde distincts tubus aeque cylindricus circ. 8 mm. longus glaber, lobi circ. 4 mm. longi subquadrati breviter emarginati. Stamina ad medium tubum inserta. Stylus (? bene evolutus) vix calycem acquans ; stigma magnum capitatum; ovarium globosum. F ructus deest.

“Upper Burma :—Shing Hong region, Burmo-Chinese fron- tier at gooo ft. A few plants only as yet seen; in the wet precipitous grassy margin of a cascade in the lower alpine tegion but probably seeded down from a higher station. The flower is going to be golden yellow. June 1920.” R. Farrer. No. 1635. ye

The affinity of the species seems to be with P. serratifolia, Franch. Although showing no sign of superposed umbels in the present state, for the specimens are of young plants, it ts referred to the Candelabra series.

-\699 Primula caldaria, W. W. Sm. et Forrest.

Species ex affinitate P. Knuthianae, Pax, specie shensicisi a qua habitu nunc triplo elatiore, bracteis basi gibbosis, calycis

36 SmitH—NEW PRIMULACEAE.

brevioris lobis angustioribus, corollae albae tubo e calyce magis exserto divergit.

Planta 1 5-50 em. alta gracilis albo-farinosa glabra. Folia varia nunc maxima ad 25 cm. longa 3.5 cm. lata; lamina a petiolo vix discreta ad ro cm. longa oblanceolata apice rotundata basi in petiolum aequilongum vel nunc longiorem alatum ad inser- tionem dilatatum sensim angustata, margine obtuse denticulata, in sicco membranacea, subtus dense albo-farinosa; nunc in speciminibus parvis folia multo minora saepe circ. 3.5 cm. longa,

mm. lata apice magis acutata, petiolo laminam haud aequante. Scapus ad 50 cm. altus flexuosus gracilis glaber sub inflores- centiam albo-farinosus umbellam 20-30-floram gerens; bracteae 6-8 mm. longae a basi lata lanceolatae acuminatae basi gibbosae ; pedicelli 8-16 mm. longi erecti graciles nunc albo-farinosi. Calyx ad 5 mm. longus campanulatus albo-farinosus in lobos anguste lanceolatos acutos fere ad medium fissus. Corollae albae tubus ad 7 mm. longus cylindricus supra dilatatus annul- atus, lobi circ. 3.5 mm. longi et lati obcordati emarginati. Stamina in flore brevistylo ad basim partis corollinae dilatatae inserta antheris 2 mm. longis ad corollae os pertinentibus. Stylus in flore brevistylo tantum 2 mm. longus calycem haud aequans, stigmate capitato. Capsula globosa 3 mm. diametro apice valvis dehiscens seminibus numerosis angulatis

“West China :—Yang-dza Shan, __Mckong-Salween divide, Yunnan. Lat. 28° 18’ N. Long. 98° 43’ E. Alt. 8-gooo ft. Plant of 6-20 inches. Flowers theirs fragrant. Amongst bush herbage in the are of hot-water springs. May 1921.” G. Forrest. No 20

var. nana, W. W. oh et Forrest.

Varietas pusilla alpina tantum 2-5 cm. alta.

‘“‘Vang-dza Shan. Alt. 11-12,000 ft. Plant of %~2 inches. Flowers fragrant white or white flushed lilac. On boulders and dry stony pasture. April-May 1g21.’’ G. Forrest. No. 20140.

The above species appears to be the first member of the Farinosae (in the restricted sense) to occur in Yunnan. Its nearest ally is P. Knuthiana, Pax. ‘The variation in size in the various specimens collected is remarkable—from dwarfs of an inch in height to plants of 20 inches. The bracts on the inflorescence do not all spring from the base but are arranged in a spiral series—and naturally the pedicels likewise.

eo Primula calliantha, Franch. var. nuda, Farrer mss. : Foliis bracteisque calyceque omnino efarinosis a typo recedit.

‘Burma :—On the Moku-ji, Chimili and other passes. Alt. — 11-13,000 ft. July 14th 1920.” R. Farrer. No, 1721,

Smitu—NEw PRIMULACEAE. 37

Mr Farrer has given the following note :—‘‘ The whole series on these alps represents not P. calliantha at all but the almost exactly similar but efarinose P. nuda which last year I only once saw and only in seed [No. 1052}. Here it abounds vastly on the heights sheeting the upper lawns in colour like the Arthritics of Europe. It varies like them too in tones of vinous amethyst or lavender, and, though the lobes are usually entire, specimens oceur in which they are erose-dentate.’

This at first sight looks very distinct but is not worthy of more than varietal rank. find in other specimens of P. calliantha a tendency to loss of farina. In the Farrerian specimen the dense base of fleshy scale leaves shows within copious farina, although the ordinary leaves, scape, bracts and calyx are quite efarinose.

ot Primula Coryana, Balf. f. et Forrest mss. descr. W. W. Sm.

Species nobilis sectionis Nivalis atque ex affinitate P. nivalis, Pallas sed ab ea specie ejusque varietatibus corollae forma facile recognoscitur ; quoad habitum atque pulchritudinem cum sf ee Balf. f. et Forrest comparanda.

Planta 35-45 cm. alta farinosa epilosa foliis sub anthesi bene ae alabastri squamis plurimis 3-5 cm. longis 2 cm. latis in sicco tenuiter membranaceis cincta. Folia magna 20-36 cm. longa, 4-8 cm. lata, lanceolata vel oblanceolata, apice acutata nunc apiculata, hake in petiolum lamina breviorem ad 10 cm. longum 1-2 cm. latum pergamentaceum sensim angustata margine crebre atque subregulariter denticulata, dentibus hydathodo apiculatis, utrinque glabra, supra efarinosa vel fere nuda, infra farina albida vel pallido-sulphurea diu persistente undique induta, costa media lata rubida marginibus stramineis (in superficie inferiore conspicua) excepta. Scapus robustus supra dense farinosus umbellas 1-3 superpositas approximatas plerumque 4-o9-floras gerens; cigs ad 1.5 cm. longae lanceolato-acuminatae farina conspersae; pedicelli longiores 2.5-3.5 cm. longi subreflexi dense farinosi. Flores majores a 4 cm. longi et lati in sicco, roseo-lavendulacei vel roseo-purpurel. Calyx ad 2 em. longus tubuloso-campanulatus in lobos oblongos vel lanceolato-oblongos acutos ad.12 mm. longos ultra medium fissus extra sparsius, intra densius farinosus. Corollae magnae tubus infundibuliformis ad 2.3 cm. longus, parte inferiore tubulosa circ. 13 mm. longa, parte superiore late ampliata,

glaber, extra farina conspersus, lobi ad I. 7 cm. longi, I.5 cm,

lati obovati emarginati nunc margine erosi, ad os latissimum farina copiosa impletum vix annulatus. Stamina in flore _ longistylo ad medium tubi inferioris inserta, in nore brevistylo

38 SMITH—NEW PRIMULACEAE.

prope os posita; antherae 2.5 mm. longae. Ovarium subglo- bosum apice incrassatum atque rubido-tinctum; stylus (longus) circ. 8 mm. vix ad os pertinens (brevis) circ. 3 mm.; stigma magnum capitatum. Capsula cylindrica ad 2.7 cm. longa, ad 5 mm. lata, apice valvis brevissimis dehiscens, seminibus nigridis circ. 1 mm. diametro minute papillatis.

““West China :—Mountains around Mu-li; S.W. Szechuan. Tat. 28° 12’ N. Long. 100? 50’ E. Alt. 12,000 ft. Plant of 14-18 inches. Under surface of foliage peculiarly aromatic when rubbed. Flowers fragrant, fleshy, soft rese-lavender ; throat filled with light sulphur-coloured farina. Half-shady ager situations in conifer and mixed forest. June 1921.”’

. Forrest. No. 20448.

1Modirthins east of Yung-ning, S.W. Szechuan. Lat. 27° 15 N. Long. 1019 EK. Alt. 11-12,000 ft. Plant of 12-16 inches. Flowers fragrant, fleshy, soft rosy-purple; throat and tube yellowish-grey. Farina on under-surface of foliage aromatic. Margins of shady forests and thickets and on open meadows. June 1922.’ G. Forrest. No. 21383.

“Mountains N.E. of Muli, S.W. Szechuan. Lat. 28° 24’ N. Long. 1019 6’ E. Alt. 12,000 ft. Plant of 12-18 inches. Flowers fragrant, soft rosy-purple; throat and tube yellowish- grey. Farina on under-surface of foliage peculiarly aromatic. Margins of forests and > ae scrub in side valleys. June 1922.’’ G. Forrest. No. 21413.

‘Litang divide. Plant ée 15-18 inches. Flowers very large, the petals more or less reflexed, nodding and pushed forward as in P, vinciflora. The colour varies from red through purple to violet, large central portion of the corolla silver-white with meal, which also powders the calyx, pedicels and upper portion of the inflorescence-axis. Under surface of leaves brilliant silver. In damp loam under Rhododendrons at the summit of the range. Seen in several places—not uncommon. Generally growing socially. Fragrant. The finest Primula I have ever seen, a truly Ge an Nivalid. June 1921.’’ F. Kingdon Ward. No. 4181

The followit ing qanibers are in fruit :—

“‘Mu-li mountains. Aug. 1918.’’ G. Forrest. No. 16846.

‘*Mu-li mountains. Sept. ro18.’’ G. Forrest. Nos. 16984, 17007.

**Mountains around Mu-li. Sept. rg21.’’ G. Forrest. No.

39-

‘Mountains around Mu-li. Sept. 1922.’’ G. Forrest. Nos. 22501, 22502. A truly noble plant of the Nivalid section and very distinct . _ from the other members of the series. The specific name is

Smitu—Nrw PRIMULACEAE. 39

in honour of Reginald Cory, Esq., of Duffryn, Cardiff, to whom the Royal Botanic Garden is indebted for much material collected by George Forrest in Western China. ‘The name was attached by Sir Isaac Bayley Balfour to the original fruiting specimens, but no description has been found.

Primula Dickieana, \Watt var. Pantlingii (King).

‘“‘S.E. Tibet :—Salween-Kiuchiang divide, Tsarong. Lat. 28° 20’ N. Long. 98° 27’ E. Alt. 14-14,500 ft. Plant of 2-3 inches. Flowers purplish-blue, throat and tube yellow. Moist alpine pasture. July 1921.’’ G. Forrest. No. 19886.

““S.E. Tibet :—Salween- -Kiuchiang divide, N.W. of Si-chi-to,

Province of Tsarong. Lat. 28° 35’ N. Long. 98° 28’ E. Alt.

14~14,500 ft. Plant of 2-4 inches. Flowers deep lavender-blue,

eve yellow, fragrant. On open moist alpine meadows. May

1922.’’ G. Forrest. No. 21633.

_ In fruit. July 1921. G. Forrest. No. 19975; in fruit. Sept. 1921. G. Forrest. No 20271; in fruit. Oct. 1921. G. Forrest. No. 20856; in fruit. Oct. 1922. G. Forrest. No. 22814.

We are unable to separate these specimens collected in S.E. Tibet from the Himalayan P. Dickicana, Watt or its ally P. Pantlingii which latter seems to us to be only a 1~2-flowered variety of P. Dickicana. Hitherto P. Dickweana was known only from Sikkim where it is a comparatively rare species. Its foliage is peculiarly gland-dotted and the S.E. Tibet specimens show the same character. We would add here that P. Valen- tiniana, Hand.-Mzt. shows the same sunken glands and resembles P. Dickieana in foliage characters. In P. Valentiniana, however the flowers are pendent and the calyx entirely different. It has been referréd by its author to the group of the Amethystinae. Its affinity is certainly with P. Dickieana, whatever the correct section may be.

wes var. chlorops, W. W. Sm. et Forr eisdem locis crescens sed fbas minoribus plerumque eens saturate purpureo-coeruleis, oculo viridi nec flavido praeditis divergere videtur.

“SB. Tibet :—Salween-Kiuchiang divide, Tsarong, west of Chamatong. Lat. 28° 18! N. Long. 98° 27’ E. Alt. 14,000 ft. Plant of 2-4 inches. Flowers deep purple-blue, eye green. On moist stony meadows. June 1922.’’ G. Forrest. No. 21674.

A variety more removed from P. Dickieana than the var.

Pantlingii,

40 Smitu---NEW PRIMULACEAE.

NA Primula dumicola, W. W. Sm. et Forrest.

Species ex affinitate P. obconicac, Hance atque P. sino-Listeri, Balf. f. a quibus foliorum forma atque aspectu, umbellis saepe superpositis divergit; quoad inflorescentiam P. malacoidem, Franch. suggerit sed est efarinosa atque indumento pilorum longorum albidorum differt.

Planta 10-30 cm. alta efarinosa. Folia ad ro cm. longa ; lamina late ovalis vel suboblonga, ad 5 cm. longa, ad 4 cm. lata, apice ambitu rotundata, basi plerumque breviter cordata nunc sub- truncata, margine circ. 11-lobulata, lobulis inciso-dentatis, supra pilisarticulatis albidis bene conspersa, intra praesertim ad nervos marginemque pilis eisdem praedita; petiolus laminam subae- quans longiuscule albido-araneoso-pilosus. Scapi plures (2-5) ad 30 cm. alti sat robusti paulo flexuosi nunc basi ascendentes undique albido-pilosi umbellam plurifloram vulgo 6—10-floram saepe duas superpositas gerentes; bracteae lineares vel sublin- eares 4-7 mm. longae albido-pilosae, pedicelli 8-13 mm. longi stricti pilis albidis conspersi. Calyx 4-5 mm. longus in fructu (vix maturo) ad 6-7 mm. ampliatus campanulatus, viridis pilis - longis albidis bene conspersus in lobos erectos triangulares acutos ad tertiam vel quartam partem fissus. Corollae roseae

. oculo albido-flavo tubus fere-7 mm. longus cylindricus glaber intra vix annulatus, lobi 3 mm. longi obcuneati apice emarginati. Stamina in flore brevistylo vix supra medium inserta atque stylum paululo superantia. Ovarium globosum; stylus brevis calyce brevior. Capsula haud matura globosa calycis tubo inclusa.

““S.E. Tibet :—Near Hsi-ku, on the descent from the Londre La to the Salween valley, Salween-Mekong divide, Yunnan. Liat, 45° 74’ NT Lene: 96° 95 Eo. AN Yr Goo ff: Pian of 4-12 inches. Flowers fragrant, rose, with eye yellowish white. On the margins of thickets by streams. May 1922.” G. Forrest. No. 21657.

Ths fa Primula effusa, WW. W. Sm. et Forrest.

Species valde afinis P. malacoidei, Franch. et P. Forbesii, Franch. ; a priore foliis carnosulis subreniformibus margine sub- regulariter dentatis, tubo corollae calycem duplo superante inter alia divergit; ab alter foliis longe petiolatis cordato-reni- formibus margine valde diverso praeditis praeter alia signa distinguitur.

Planta ad 50 cm. alta, haud annua, radice sat robusta praedita, gemmas proferens. Folia numerosa cum petiolo ad 15 cm.

longa; lamina majorum 4.5 cm. longa et lata, cordato-reniformis -apice plus minusve rotundata, basi plerumque alte cordata, -paululo sinuato-lobata, margine argute atque subregulariter

Smitu——New PRIMULACEAE. AI

dentata, utrinque sparsim hispida cito subglabrescens ; petiolus ad 12 cm. longus laminam longe superans, pilis longiusculis bene conspersus. Scapi plures, nune 8, ad 50 em. alti flexuosi graciles glabri efarinosi umbellas superpositas 2-3 intervallis 5-8 em. longis plurifloras (5-16-) gerentes; bracteae 3-4 mm. longae virides lineares glabrae; pedicelli ad 5 cm. longi inaequales substricti glabri vel subglabri. Calyx ad 4 mm. longus anguste campanulatus viridis sparsim albo-farinosus ad medium in lobos anguste lanceolatos acutos fissus, in fructu multo latior atque ad 5 mm. elongatus. Corollae lavendulaceo- roseae tubus S mm. longus anguste cylindricus extra flavidus vix annulatus, lobi ad 6 mm. longi obcuneati alte emarginati. Stamina filamentis brevissimis in flore longistylo infra medium tubi corollini inserta, in flore brevistylo circ. ad medium posita. Ovarium subglobosum; stylus longus inclusus vix ultra medium tubi corollini pertinens, brevis calycis tubo paulo brevior. Capsula globosa calycis tubo paululo exserta ab apice valvis 5 presisecis dehiscens, seminibus 0.5 mm. diametro brunneis . sciformibus nune angulatis.

“West China :—Yangtze-Yung-ning divide; Yunnan. Lat. 27° 40’ N. Alt. 7-Sooo ft. Plant of 1o-20 inches. Foliage fleshy. Flowers lavender-rose, a Shady moist situa- tions in thickets. June 1914.’’ G. Forrest. No. 12446.

Vest China Ficaie the pen to Semraet on the Yangtze, ¥unnan.:» Lat. 279.35! N.,, Long: 100° 301, K.-Alt. 11,000, ft. Plant of 6-12 inches. Flowers rose fragrant. Foliage fleshy, leathery. Shady moist banks on slate formation. Note—this is not the ad P. malacoides—it was first collected in June 1913. May 1922.’ . Forrest. No. 21236

“West co :—Fengkow, ‘Gecrean: By irrigation streams amongst grass and ferns in the cultivated terraces below Fengkow and just above the Yangtze. ‘The rootstock appears to be perennial. Alt. 7ooo ft. Flowers pinkish-mauve with white eye and orange tube, darkly stri eaves almost glabrous or with a few spattered bristly hskee under surface powdered with white meal. Meal also on calyx, pedicels, etc. April 15th, 1922.’’ Kingdon Ward. No. 5044.

In Mr Forrest’s opinion too distinct a plant to be reckoned as a variety of either P. malacoides or of P. Forbesii. In its fresh condition the leaves are succulent like the leaves of some of the Diptera Saxifrages but dry to a very thin consisténcy. It is found only on moist slate formations.

Primula eucyclia, WW. W. Sm. et Forrest. Species inter congeneres chinenses facile distincta atque cum P. vaginata, Watt specie — tantum comparanda; ab

2 SmitH—Nrw PRIMULACEAE. 4

illa foliis textura tenuioribus magis regulariter orbiculatis, floribus 1~3-natis plerumque binis, corollae tubo atque limbo fere duplo majoribus divergit; petiolis basi manifeste vaginant- ibus species hae duae conspicuae.

lanta perennis ad 8 cm. alta rhizomate elongato anguste cylindrico basibus foliorum delapsorum bene obsito. Folia saepe ad 12, alia petiolo ad 4 cm. longo basi vaginante praedita, alia (minora) petiolo vix 2 mm. longo munita sed vagina lata. ad r cm. longa et 4 mm. lata margine ciliolata instructa; lamina vulgo 2.5 cm. diametro, ambitu orbicularis, basi cordata, ad medium subseptemloba, lobis ad trientem incisis, lobulis acutiusculis, supra minute puberula, sub lente quasi sieivicluias infra pallidior glabra vel subglabra, margine minute ciliolata. Scapus ad 8 cm. altus gracilis folia stperans minutissime puberulus umbellam plerumque bifloram gerens; bracteae 2-4 lanceolatae 3-4 mm. longae acutae minute ciliolatae; pedicelli graciles 7-9 mm. longi minute puberuli. Calyx late campanu- latus circ. 5 mm. longus ad medium fissus glaber vel subglaber ; lobi ovati obtusiusculi margine minutissime ciliolati. Corollae pallide roseae tubus cylindricus fere 1 cm. longus ore annulatus ; limbus 1.6-1.8 cm. latus in lobos obovatos 4~5-dentatos dentibus truncatis divisus. Stamina in flore longistylo calycem paululo superantia. .Ovarium subglobosum; stylus in flore brevistylo calycem aequans, in flore longistylo ad summum corollae tubum attingens; stigma parvum capitatum. Fructus e calyce haud aucto vix exsertus subglobosus apice valvis 2-5 dehiscens; semina pauca 1 mm. diametro paulo complanata minute areolata.

“S.E. ‘Tibet :—Province of ‘Tsarong, Salwin-Kiuchiang divide, N.W. of Si-chi-to. Lat. 28° 35’ N. Long. 98° 28’ E. Alt. 14,000 ft. Plant of %-2% inches; perennial. Flowers pale shell-pink. On rocks and stony alpine meadows. Mav 1922.’ G. Forrest. No. 21632.

‘tLats 528° saotiaNe hong. ae 27! E. On open alpine meadows. In fruit. July roar.’’ . Forrest. No. 19868.

at) 280124" N. Long. 98° 24’ ae Alt. 14-15,000 ft. In fruit. Sept. 1921.’ (;. Forrest. No. 20270.

This species is the Chinese representative of the Himalayan P. vaginata, Watt. As in that species the leaves are markedly dilated at the base into a broad vagina. P. eucyclia is a slenderer plant than the Himalayan with fewer but much larger flowers.

ws Primula gentianoides, \V. W. Sm. et Ward.

Species anomala cujus affinitas nobis dubia est; scapis cuique spenenag, 1-4 ebracteatis unifloris, floribus magnis praecocibus,

SMITH—NEW PRIMULACEAE. 43

corollae lobis valde elongatis integris, staminibus stylogue non vel vix calycis tubum superantibus inter alia signa bene conspicua.

Planta sub ee. ad 6 cm. alta epilosa efarinosa basi squamis carnosulis 1-1.5 ecm. longis obsita radicibus crassiusculis praedita. Folia sub anthesi haud bene evoluta petiolo alato laminam subaequante praedita; lamina cire. 1.5 cm. longa, vix t em. lata, late spathulata, apice rotundata, basi in petiolum sensim angustata, integra vel subintegra, carnosula atroviridis glabra. Seapi 1-4, oc a ee flores aequantes sub calyce sensim ampliati uniflori axillares glabri. Flores in sicco 3 cm. longi erecti glabri_ pallido-coeruleo-violacei. Calyx 8-9 mm. longus late tubulosus viridis textura tenuis glaber in lobos ovatos erectos apice rotundatos fere ad medium fissus. Corollae textura gracillimae tubus 11-12 mm. longus late cylindricus supra paululo dilatatus extra glaber, lobi 17-18 mm. longi oblanceolati medio § mm. lati apice rotundati integri; os exannulatus sed capillis longiusculis delicatis munitus ut tubus intus. Stamina in flore longistylo tantum 2 mm. a basi tubi corollini remota, in brevistylo calycis tubum paululo superantia, antheris circ. 1.5 mm. longis. Stylus longus circ. 5 mm., brevis vix 3 mm. superans; stigma capitatum; ovarium globosum placenta ut videtur pulvinato-discoidea, seminibus immaturis multis. Fructus deest.

_ ‘West China :—Yungning, N.W. Yunnan. Alt. 12-13,000 ft. €Flowers pale bluish violet with no meal. Leaves glabrous with no meal. In open grassy meadows along the summit of the forested range. 25th April 1922.” F. K. Ward. No. 5088.

A very peculiar species. The plant is dwarf but the flowers are Jarge and precocious. “There may be one to four scapes to each plantlet and the scapes are single-flowered and ebracteate. The habit in some ways recalls P. sonchifolia, Franch., but the structure of the flower is very different. The one-flowered ebracteate scapes lead mechanically to Omphalogramma but the calyx and corolla do not accord with that genus; there is, more- over, the want of any indumentum; the inside of the tube is lined with very fine but longish hairs but not forming a tuft in the throat as in the Bella section. It may be placed provisionally in the Petiolaris section. The placenta is broad and convex with the seeds lying upon it. This may point to it being an aberrant Petiolaris (cf. Balf. f. in Notes, R.B.G., Edin. IX. (1916) p. 175 sub P. hylophila.)

og Primula ingens, W. W. Sm. et Forrest.

: Ex affinitate P. russeolae, Balf. f. et Forrest et P. sinonivalis, _ Balf. f. et Forrest a quibus habitu multo elatiore, seapo pedi-

“

4d Smitu-—-New PrmoulAackak.

cellisque dense luteo-farinosis, floribus pernumerosis divergit ; omnes pro subspeciebus P. nivalis, Pall. fortasse rectius aestimari debent.

lanta florifera ad 50 cm. alta vel ultra, in fructu ad go em. attingens, farinosa epilosa radicibus multis crassis; foliorsin rosula sub anthesi bene evoluta, alabastri squamus plurimis aq 8 em. longis, 3 em. latis rubidis in sicco tenuiter membranaceis cincta. Folia magna ad 33 cm. longa, ad 5 em. lata, lanceolata vel oblongo-lanceolata apice acutata vel obtusa, deorsum in petiolum late alatum attenuata margine integra vel nunc obscure denticulata, supra nuda, subtus sparsim luteo-farinosa cito subnuda; petiolus 8-9 em. longus, ad 2 cm. latus, in sicco tenuiter membranaceus rubidus squamis basalibus subsimilis. Scapus robustissimus apicem versus dense luteo-farinosus umbellam multifloram vel duas superpositas gerens; bracteae 6-20 mm. longae lanceolatae acuminatae sparsim farinosae margine ciliolatae; pedicelli cire. 1.5 cm. longi, in fructu 4-8 em. longi, reflexi vel superiores subreflexi (in fructu erecti), dense luteo-farinosi; anthopodium magnum. Calyx ad 12 mm. longus in babecdasapaanhabe ad medium in lobos lanceolatos vel suboblongos acutatos fissus extra fere nudus atro-viridis vel atro-brunneus intus dense farinosus venis distinctis longi- tudinaliter pererratus. Corollae purpureo-coeruleae tubus ad 1.5 cm. longus infundibuliformis glaber annulo distincto praeditus, lobi obovati vel subrotundati 11.5 cm. longi integri margine pilis minutis glanduli-capitatis muniti. Stamind# in flore Jongistylo ad partem tubi corollini quartam inferiorem inserta antheris circ. 2.5 mm. longis. Ovarium subglobosum in dimidio superiore incrassatum atque rubro-tinctum; stylus (longus) 7-8 mm. longus haud exsertus; stigma capitatum. Capsula cylindrica 1.5-2 em. longa e calyce venoso exserta straminea apice valvis dehiscens seminibus circ. 1.5 mm. longis brunneis spongiosis.

“West China :—Bei-ma Shan, Yunnan. Lat. 28° 18/ N. Long. 99° 10’ E. Alt. 14-15,000 ft. Plant of 12-20 inches. Flowers fragrant soft purple-blue with grey eye. Open stony

meadows by streams. June 1921.’’ G. Forrest. No. 108 ; 9 9849.

“‘Bei-ma Shan. Plant of 2%-3 ft. In fruit. Flowers shades of lavender and purple-blues, occasionally pale rose; fragrant. On moist pasture by streams. This is the most robust and one of the most vine Nivalid Primulas I know. Sept. 1921.” G. Forrest. No. 2

** Na-shu-to a Mekong-Vangte divide, Yunnan. Lat. 28° 30/. Long. 99242! Bo AN. 14 \¢00: Plant of 2-2% ft. In fruit. On open seeatioss by streams. Oct. - UP AS. For-

rest. No, 20617.

SmMitH-——New PRIMULACEAE. 45

The above plant appears to come nearest P. sino-nivalis and P. russeola. In habit it has a strong resemblance to the white- flowered P. chionantha, Balf. f. et Forrest. With these it must be placed as one of the many members of the aggregate P. nivalis, Pall. It should prove a magnificent species in cultivation if as amenable as P. chionantha.

\ 4: _ reo \5° Primula lacerata, WW. W. Sin.

Species burmanica affinis P. Veilchianac, Petitm. atque P. pirolaefoliae, Lévl. (quae fortasse sunt eadem); glandulae peculiares in bracteis et calycibus et corollis plantarum supra citatarum in P. lacerata omnino desunt; foliorum textura atque corollae laciniis insuper ab eis removitur.

Planta 15-18 cm. alta scapis 3-9 folia longe superantibus. Folia petiolo 2-6 em. longo minute puberulo praedita ; lamina

2-4.5 cm. longa et lata, orbicularis vel suborbicularis, apice rotun-

data, basi breviter cordata nunc vix cordata, margine obscure crenato-dentata, textura chartacea, subtus subglauca utrinque glabra. Scapi axillares ad 18 cm. alti minutissime puberuli 3-7-flori; bracteae 3-4 mm. longae anguste lanceolatae sub- glabrae (nonnunquam singula ex involucro longe petiolata atque laminam 1~2 cm. diametro gerens) ; pedicelli circ. 1 cm. longi, in fructu ad 2.5 cm. elongati, dense furfuraceo-puberuli; flores erecti fragrantes lavendulaceo-coerulei oculo aurantiaco. Calyx circ. 7 mm. longus anguste campanulatus minutissime puberulus vel subglaber in lobos ovatos subacutos vel obtusiusculos circ. 3 mm. longos fere ad medium fissus. Corollae tubus calycem paululo superans cylindricus, annulus distinctus convolutus, lobi cire. 8 mm. longi anguste obovati bilobi lobulis ipsis pectinato-incisis. Stamina in flore longistylo multo supra medium tubi corollini inserta, antheris circ. 1 mm. longis ad annulum pertinentibus, in flore brevistylo antherae ex ore exsertae. Stylus in flore longistylo ex ore saltem 3 mm, exsertus, in flore brevistylo calycem paululo superans. Capsula fere matura globosa seminibus multis angulatis; dehiscentiam non observavi.

“N.W. Burma :—Between Shingrup Chet and Akhyang. Alt. 7,500 ft. On ledges of hard shale bluffs, on moss on dead trunks and on pathside banks of the uppermost rain-forest or lowest alpine. Flowers almost over by April 15th, in shades of lavender-blue with an orange eye, very fringy and sweetly scented like P. farinosa. Range seems limited, no sign of it was seen on the east side of the pass and it was only found

hin a very small belt of the western. Seen again on one Tittle wet granitic sligcenty at Sooo ft. on the wooded crest

Ab SmitH— NEW PRIMULACEAE.

(Kum Ja Bum) between Kenglungpu and Salangpa, in full Hower and character. Its beauty is remarkable and its flower-form so very unstable that it should develop widely in cultivation. The eve can be a solid ring of orange or a mere many-rayed corona. with a penumbra of pallor. Note too the far exserted style. Nowhere seen in the woods of the Upper Akhyang. April 1920.”’ R. Farrer. No: 1512.

A very beautiful species not yet in cultivation. It has no afhnity with any Himalayan or Burmese species known to me but comes very near to the Szechwan P. Veitchiana, Petitm. as well as the eastern Yunnan plant described by Léveillé as P. pirolaefolia. Of this last I have ample material; of P. Veitch- iana only a scrap. But both are characterised by peculiar internal glandular areas in the bracts and calyx-lobes while the corollas of both are sprinkled with dark glandular dots. Microscopic examination confirms the glandular nature of these marks which are visible from without. PP. lacerata is devoid of these but its close relationship with the Chinese plants is undoubted. P. Vettchiana and P. pirolaefolia appear to me to

the same species but unfortunately my material of the former is very limited. ‘These plants appear to be connected with the section Carolinella in which they are placed provisionally. They agree with one another in the laciniation of the corolla- lobes—not simply emarginate as in the typical members of that section.

et es leucops, \W. W. Sm. et Ward.

Ex affinitate P. aluntzuensis, Balf. f. et Forrest et P. sino- plantagineae, Balf. f{.; planta nana sectionis Nivalis; folits angustissimis, floribus pro planta magnis, corollae oculo magno albo bene notata.

Planta dense caespitosa ad 6 cm. alta, in fructu nune vix acerescens, nunc paulo altior, pallido-luteo-farinosa. Folia 2-3 em. longa, 3-5 mm. lata, petiolo vix discreto alato praedita, sub anthesin vix omnino evoluta; lamina lineari-lanceolata vel anguste lanceolata, acuta vel subacuminata, marginibus revolutis obscure crenato-denticulata, subtus farinosa. Scapus foltis duplo longior umbellam 2-5-floram gerens; bracteae 4-5 mm longae pedicellos plus minusve aequantes basi latae ipnootatae acutae vel acuminatae virides vel nunc purpurascentes leviter farinosae. Calyx circ. 8 mm. longus cylindricus atropurpureus

tubus circ. 1 cm. longus supra calycem ampliatus, lobi subro- undati 1 cm. —— Te sina Stamina in flore brevistylo

SMITH—NEW PRIMULACEAE. 47

supra medium tubum inserta antheris 1.5 mm. longis ab ore 2 mm. remotis. Stylus in flore brevistylo vix calycis tubum superans. Capsula cylindrica 1.5 cm. longa.

“West China :—Gu-ja-la, Muli district, S.W. Szechwan. Alt. 15-16,000 ft. Flowers various shades of mauve, lilac and purple but always with white eye. Fragrant. On open alpine slopes. May 1922.”’ F.K. Ward. No. 5133.

The aggregate P. nivalis, Pall. and its allied species appear to be prolific in West China in the development of various subspecies and varieties. At one extreme there are tall plants such as P. ingens, attaining 3 ft. in the fruiting stage and at the other dwarf plants such as P. atuntzuensis, P. brevicula, P. minor, P. petraea, and P. rigida. P. leucops finds a place with these last.

\y Primula lichiangensis, Forrest var. hapala, Balf. f. et Forrest.

tial es

Varietas foliis subtus plus minusve dense albido-araneoso- tomentosis a planta typica divergens.

“West China :—Mountains east of Yung-ning, S.W. Szechuan. Lat. 27° 48’ N. Long. 1019 E. Alt. 12-13,000 ft. Plant of 8-12 inches. Flowers fragrant, dark purple-rose, eye orange. Qn ledges of dry limestone cliffs. June 1922.” G. Forrest. No. 21227.

“Eastern flank of the Bei-ma Shan, Yunnan. Lat. 289 12’ N. Alt. 12-13,000 ft. On rocks and humus-covered boulders.

June 1917.’’ G. Forrest. No. 13848.

3! Primula Littoniana, Forrest var. robusta, Forrest.

A typo habitu robustiore, foliis majoribus, spica multo longiore ad 25 cm. longa apicem versus multo attenuata, floribus pallide lavendulaceis divergit.

‘‘West China :—Mountains S.E. of Mu-li, $.W. Szechuan. Lat. 279 30’ N. Long. 1019 E. Alt. 11,000 ft. Plant of 2-2% ft. Flowers pale purplish-blue. Open moist meadows on the margins of conifer forests. Aug. 1922.” G. Forrest. No. 22170.

In fruit. Sept. 1922. G. Forrest. Nos. 22490, 23246.

This variety has been found so far only on the mountains around Mu-li. All the plants of P. Littoniana in the area are the same. In other areas where typical P. Littoniana occurs, this giant form has not been seen. It has been in cultivation, but, although its spike may attain sometimes 15 inches in length, its paler-coloured flowers do not make it so.attractive a plant as the typical form, S

48 SmitH—Ne&W PRIMULACEAE.

1 Primula malvacea, Franch. var. alba, Forrest.

A typo floribus albis recedit.

“West China :—Sungkwei Range, Yunnan. Lat. 26° 18’ N. Long. 100° 12’ EK. Alt. 10-11,000 ft. Plant of 9-14 inches. Flowers fragrant, pure white with yellowish eye and tube. On open alpine meadows. July 1922.’’ G. Forrest. No. 21557.

Primula malvacea, Franch. var. intermedia, W. W. Sm. et Forrest.

Varietas nana P. malvaceae, Franch. ; quoad habitum foliaque ad P. blattariformem, Franch. spectans sed inflorescentia P. malvaceae praedita; folia plerumque breviter petiolata ; flores in umbellis superpositis sed, si nunc racemose dispositi, pedicellis elongatis muniti.

“West Chinn :—-Mountains of the Chungtien plateau, Yun- nan. Lat. 27° 30! N. Alt. 11,000 ft. Plant of 3-6 inches. Flowers very pale 1 rose. Open stony pasture. July 1914.” G. Forrest... No. 12

“Muli sicwbbctaitte, valley of the Litang, S.W. Szechuan. daats28° 12 N. Alt. 10,000 ft. Plant of 4-9 inches. Flowers rose with eye yellow. Stony pasture and on ledges of cliffs. July 1918.”’ G. Forrest. Nos. 16276, 16495, 16795.

“Muli. Alt. 9-10,000 ft. Flowers large for the size of the

‘plant, delicate mauve with white eye. On shady limestone

cliffs or in the woods below the cliffs. A charming little plant of refined colour not, however, fragrant. July 1921.’ F. K. Ward. No. 4521.

‘Mountains east of Yung-ning, S.W. Szechuan. Lat. 27° 48’ N. Yong. 1019 KE. Alt. 9-10,000 ft. Plant of 4-10 inches. Flowers rose. Open stony slopes and on cliffs and rocks in dry

side valleys. June 1922.’’ G. Forrest. No. 21425.

Primula melanops, WW. W. Sm. et Ward.

Ex afiinitate P. sinopurpureae, Balf. f. et P. sinoplantagineae, Balf. f. atque cum eis pro microforma P. nivalis, Pall. habenda. Ad P. sinopurpuream potius approximat sed minor, foliis subtus albido-farinosis, floribus saturate atro-purpureis oculo nigro ornatis, capsula magna cylindrica ad 2.5 em. longa. Planta florifera ro-15 cm. alta, in fructu nune ad 35 cm. elata. “West China :—Litang divide, S.W. Szechwan? Alt. 15,000

ft. Flowers deep Tyrian purple, black in the centre, fragrant.

yy?

On screes by glacier lake under the lee of sre. June 1921. rs » Ward, No. 4080

SmitH—NEW PRIMULACEAE. 49

“‘Gu-ja-la, S.W. Szechwan. Alt. 14-15,000 ft. Flowers deep Tyrian purple, the eye darker almost black. On open alpine slopes facing the sun. June 1922.”” F. K. Ward. No. 5132.

Another member of the aggregate P. nivalis, Pall.

\$l0Primula microstachys, Balf. f. et Forrest.

Planta nana affinis P. blattariformi, Franch. cujus pro varietate vel microforma rectius aestimanda; omnibus partibus multo minoribus a planta typica recedit.

Herba 7-20 cm. alta. Folia petiolo 5-10 mm. longo dense pubescente munita; lamina ad 3 cm. longa, ad 2 cm. lata, late oblonga vel ovalis, apice rotundata, basi late cuneata vel subrotundata, ut ea P. blattariformis induta. Scapus gracilis dense pubescens flores 3-12 racemose dispositos gerens ; bracteae lineares ad 5 mm. longae; pedicelli 1-3 mm. longi. Calyx circ. 5 mm. longus. Corollae roseae oculo flavo tubus cire. I em. longus anguste cylindricus gracillimus, lobi 5-7 mm. longi cuneato-obovati alte emarginati. Stamina in flore longistylo ad medium tubi corollini inserta. Stylus fere ad os annulatum pertinens.

“West China :—Mekong-Yangtze divide, Yunnan. Lat. 27° 4o' N. Alt. g-10,000 ft. Plant of 3-8 inches. Flowers rose, eye yellow. Open stony pasture. Aug. ro14.’’ G. Forrest. No. 13137.

A very different looking plant from P. blattariformis, Franch. at first sight, of dwarf habit and with very slender flowers and narrow corolla tube. But it will spveccei prove to be an extreme variety of that species.

\ud? Primula monantha, WW. W. Sm. et Forrest.

Species affinis P. gemmiferae, Batal. atque P. chrysopae, Balf. f. et Forrest a quibus habitu multo minore, foliis fere orbicu- latis, floribus solitariis recedit.

Planta nana 3-6 cm. alta radicibus paucis gracilibus. Folia 5-6 rosulam parvam formantia, petiolo alato latiusculo laminam plus minusve aequante praedita ; lamina 7-9 mm. longa, 6-7 mm. lata, vel nunc minor, suborbicularis vel late spathulata, apice rotundata, basi subabrupte in petiolum contracta, margine dentibus + 20 glanduli-capitatis instructa, textura carnosula, utrinque levissime albo-farinosa epilosa. Scapus ad cm. longus flexuosus leviter albo-farinosus florem solitarium suber- ectum gerens; bracteae 2, major circ. 4 mm. longus, minor 2_ mm., anguste lanceolatae acutae basi obscure gibbosae albo- marinosac ut pedicelli circ. 5 mm. longi. Calyx circ. 5 mm.

50 SmitH—-NEw PRIMULACEAE.

longus anguste campanulatus extra leviter farinosus intus densius in lobos ovato-lanceolatos acutos ad medium fissus. Corollae glabrae purpureo-roseae tubus circ. 9 mm. longus cylindricus supra sensim ampliatus, lobi obovati circ. 9 mm. longi emarginati. Stamina in flore jAuesety's ad medium tubum inserta. Ovarium globosum; stylus in flore longistvlo ad os pertinens; stigma capitatum ructus deest.

est China :—Me kong- Yangtze divide, Yunnan. Lat. 27° 36’ N. Long. 99° 10’ E. Alt. 14,500 ft. Plant of 1%-2% inches. aoe soft purplish-rose. On moist stony pasture. July 1921.’ . Forrest. No. 19678.

We have ae doubt of the affinity of this dwarf species with P. gemmifera, Batal. and P. chrysopa, Balf. f. et Forrest. It is distinguished easily by the rounded pluridentate small leaves and the solitary flowers. Only one collection was made and in the area the plant appeared to be very rare.

yw? Primula oxygraphidifolia, WW. W. Sm. et Ward.

Species nana valde distincta cujus affinitas est cum P. tibetica, Watt, atque P. fasciculata, Balf. f. et Ward, a quibus foliis diversis, floribus multo majoribus solitariis (ut in P. fasciculata), calycis lobis obtusis praeter alia signa removitur

Planta 3-5 cm. alta caespitosa efarinosa epilosa rhizomate parvo vaginis emarcidis foliorum vetustorum induto. Folia petiolo 8-18 mm. longo alato praedita; lamina variabilis nunc late ovata circ. 1 cm. longa et 7 mm. lata, nunc fere orbicularis vel ovato-lanceolata vel lanceolata, apice rotundata vel obtusa, basi rotundata vel in foliis angustioribus in petiolum angustata, margine integra vel dentibus 1-3 utrinque munita, textura in sicco carnosula, supra atroviridis subtus olivacea; folia anni prioris saepe persistentia et aspectu albo-spongiosa. Scapus. ad 5 em. longus axillaris ad medium bracteis 1-2 membranaceis circ. 2 mm. longis appressis lanceolatis acutis munitus. Calyx 6-7 mm. longus anguste campanulatus glaber efarinosus in lobos ag cire. 3 mm. longos et 1.5 mm. latos apice rotundatos

c undulato-erosos fissus; lobi atque juga viridia intervallis pallido-viridibus submembranaceis. Corollae lavendulaceo- coeruleae glabrae tubus 11-12 mm. longus cylindricus supra paulo dilatatus, annulus valde distinctus, lobi 8 mm. longi

obcordati alte emarginati. Stamina in flore longistylo calycis

tubum medium haud superantia, in flore brevistylo paulo supra

_ medium tubum corollinum inserta. Ovarium globosum; stylus _ longus vix ad os pertinens, brevis vix 1 mm. superans; stigma =. capitatum. Fructus (ex reliquiis anni prioris) haud exsertus ne valve gage

ya?

‘

SMITH---NEW PRIMULACEAE. 51

“Muli Range. Flowers mauve or purplish, with pale yellow eve. No meal. On limestone precipices in nooks and crannies. Not common. Plants nearly always occur singly, with one large flower closely pressed against the rock. The previous year’s leaves clothe the stem. 5th June 1921.’ F. K. Ward. No.

2 -““Muli Range. Alt. 15,000 ft. Flowers solitary and without meal. Grows at high alitades on the limestone cliffs in mossy Savers. t5th June 1922.” F. K. Ward. No. 5222.

““Mountains around Mu-li, S.W. Szechuan. Lat. 28° 12’ N. Long. 100° 50’ E. Alt. 12,000 ft. Plant of 1-2 inches. Flowers lavender-blue, eve grey-yellow. Open stony moist pasture by streams. July 192r.”’ G. Forrest. No. 20473.

‘‘Mountains N.E. of Mu-li. Lat. 28° 22’ NY Long. 101° 6’ E. Alt. 12-13,000 ft. Plant of 1%-2 inches. Flowers fragrant, deep lavender, tube white. Open stony moist alpine meadows by streams. June 1922.” G. Forrest. No. 21332.

A peculiar species akin to P. tibetica, Watt. The scapes are one-flowered. The leaves of the previous year are sometimes persistent, drying to a white spongy appearance,

Primula Pauliana, W. W. Sm. et Forrest.

Species in sectione Geranioides inflorescentiis elongatis laxe racemosis secundifloris atque floribus pallide flavis valde conspicua.

Planta satis robusta 25-40 cm. alta. Folia 3-9 petiolo ad 18 em. longo pilis articulatis sparsim instructo munita; lamina foliorum majorum ad 9 em. longa, ad 8 cm. lata, orbiculari-ovata, basi cordata, textura in sicco tenuissima, ad partem quintam subseptemloba, margine grossi-dentata atque ciliolata, supra glabra vel hic illic pilis articulatis sparsim praedita, infra pallidior ad nervos rubidos basi radiantes pilis similibus bene conspersa. Scapus ad 4o cm. longus, gracilis flexuosus supra medium floriferus, pilis articulatis paucis atque pube minuta indutus; inflorescentia elongata laxe racemosa 20~30-flora secundiflora ; bracteae 3-5 mm. longae plerumque lanceolatae vel Sublineares integrae, nunc oblongae apice 3~4-lobulatae, pubes- centes, nonnunquam pedicello suffultae; pedicelli 6-10 mm. longi ut scapus induti. Calyx 6 mm. longus minute puberulus ad medium in lobos triangulares acuminatos indurato-apiculatos fissus. Corollae pallido-flavae tubus aurantiacus circ. 11 mm. ongus anguste cylindricus extra minutissime pubescens obscure

annulatus; limbus circ. 16 mm. diametro lobis obovatis alte ‘Mmarginatis. Stamina in flore longistylo in partem inferiorem

corollini disposita, tubum calycis vix superantia, in flore

52 SmitH—NkEW PRIMULACEAE.

brevistylo ad os corollinum inserta. Ovarium globosum ; stylus subexsertus in flore longistylo, in flore brevistylo ovarium aequans atque vix tubum calycis superans, glaber; stigma capitatum. Fructus circ. 9 mm. longus e calyce haud aucto exsertus, oblongus, apice dentibus parvis dehiscens.

“West China :—S.W. Szechwan, in the mountains S.E. of Yung-ning. Lat. 27° 30’ N. Long. 100° 50’ HE. Alt, II-12,000 ft. Plant of 9-16 inches. Flowers pale yellow, eye and tube orange. Margins of thickets and forests and on open meadows. Aug. 1922.’’ G. Forrest. No. 22073. Also Sept. 1922. No. 22452 in fruit.

A striking species of the Geranioides series and apparently the only yellow-flowered member of that series. “The very elon- gate lax racemose inflorescence is also an easy mark for recognition. ‘The specific name is in honour of the Very Rev. David Paul, D.D., LL.D., an enthusiastic student of the European Primulaceae.

yd Primula pseudosikkimensis, Forrest.

Microforma Primulae sikkimensis, Hook.; planta calcicola habitu humiliore, foliis textura minus avenits: floribus intensius atque laete luteis a planta typica divergit. Plantae ambae in provincia yunnanensi crescunt; P. pseudosikkimensts montibus calcareis Lichiang dictis favet.

‘West China :—Lichiang Range, Yunnan. Lat. 27° 20’ N. Alt. 11-12,000 ft. Plant of 12-18 inches. Flowers bright canary-yellow, fragrant. In crevices and ledges of limestone cliffs. July 1906.’’ G. Forrest. No. 2642.

“‘Kastern flank of the Lichiang Range. Alt. 12-15,000 ft. Plant of 6-9 inches. Flowers soft canary-yellow, fragrant. Alpine meadows. June 1gr1o.’’ G. Forrest. No. 5978. July 1913: No. 10329.

The plant from the Lichiang Range does not quite match the Himalayan species. Where dwarf plants occur in the Himalaya with increase of altitude, there is, at the same time, a decrease in size of flower. The Lichiang plant is of stout habit with large brightly-coloured flowers.

ee Primula tapeina, Balf. f. et Forrest.

Species affinis P. Dubernardianae, Forrest a qua habitu humiliore valde florifero, foliis supra farinosis infra albo- farinosis utringue multo minus glandulosis, calyce minore

_ textura crassiore magis farinoso quam i. corolla alba

divergit.

_ Planta nana suffruticosa ad 2.5 cm. ascendens densissime . eee. glebam 30-60 cm. diametro formans, habitu P.

SMItH-—NEW PRIMULACEAE. 53

PDubernardianae, rhizomate valido lignoso foliis marcidis dense obsito. Folia petiolata 3.5-6 cm. longa ; lamina 6-8 mm, lata, anguste lanceolata vel oblanceolata in petiolum anguste alatum glanduloso-puberulum laminam subaequantem angustata, supra fere plana leviter albo-farinosa vel nunc breviter glanduloso- pubescens, subtus dense albo-farinosa, apice obtusa, margine subintegra vel paulo repanda. Scapus ad 2.5 cm. altus, saepe multo minor, leviter farinosus vel breviter glanduloso-pubescens umbellam 2-6-floram gerens; bracteae 6-8 mm. longae lanceo- latae vel suboblongae leviter farinosae atque pubescentes ; pedicelli ad 1.5 cm. longi subglabri vel nune pubescentes apicem versus farinosi. Calyx campanulatus cire. 6 mm. longus carnosulus albo-farinosus epilosus vel nunc breviter pubescens in lobos ovato-triangulares obtusos ad medium fissus. Corollae albae fauce luteo tubus 1 em. paulo superans e calyce longe exsertus extra primo dense farinosus ut lobi obcordati 7 mm. longi alte emarginati. Stamina in flore longistylo ad quartam partem tubi iter ice inserta, in flore brevistylo supra medium, antheris 2 mm. longis ab ore 2 mm. remotis. Ovarium globosum ; stylus in flore longistylo circ. 6 mm. longus, in brevistylo vix 1.5 mm. aequans; stigma parvum capitatum. F ructus deest.

“S.E. ‘Tibet :—Salween-Kiuchiang divide, Tsarong. Lat. 28° go! N. Cushion plant of 6-12 inches diameter. Flewers white with orange eye. On dry shady rocks and clifls. June 1919. ” - G. Forrest. No, .28367,

‘‘Salween-Kiuchiang divide, sarong. Lat. 28° yo’ N. Long. 98° 15’ BE. Cushion plant of 1 inch. One to two feet in diameter. Flowers white occasionally tinged rose, eye yellow. Dry situations on limestone cliffs. July 1919.’’ G. Forrest. No. 18896.

P. tapeina presents a very distinct appearance from P. Dubernardiana, Forrest in its very compact habit and freedom of flowers. It differs moreover in the character and distribution of the meal, in shorter more fleshy calyx and in the white corolla long exserted from the calyx tu

Primula Valentiniana, Hand.-Mzt. in Sonderabd. aus Anz.

Akad. Wien, No. 26-27 (1922), Pp. 5.

A beautiful dwarf species of the section Amethystinae collected by Dr. Handel-Mazzetti and published in 1922. It was alsa collected by Farrer and by Forrest.

urma :—Chawchi Pass. Alt. 12-13,000 ft. <A little beauty, sheeting the flat fine lawns of the high alps in broad expanses of colour. Flowers intense pure crimson, paling to

54 SmitH—NrEw PRIMULACEAE

their rim and darkening bloomily up to the intense maroon

blackness of the calyx. In number they are usually two, but can be as many (in one unique case) as 5, borne in a genuine head, though sometimes they almost look superimposed. July 14th 1920.”’ erarrer. Mov-1723.

““S.E. Tibet :—Salwin-Kiu Chiang divide. Lat. 28° 12/ N. Long. 98° 24’ E. Alt. 15-15,500 ft. Plant. of 2-3 inches. Flowers fleshy, deep wine-crimson, identical in shade to the darker forms of P. secundiflora. On open, peaty alpine meadows. July 1921.’’ G. Forrest. No. 20000. ‘The same in —

g2t. “‘Salwin-Kiu Chiang divide, west of Chamatong, sarong. Lat. 28° 18’ N. Long. 98° 27’ E. Alt. 13-14,000 ft. Plant’ of 1%~3 inches. Flowers solitary, occasionally in pairs, pen- dulous, fragrant, deep purple-crimson. Moist alpine meadows and moorland. June 1922.” G. Forrest. No. 21792. The same in fruit. Oct. 1922. G. Forrest. No. 22931. he leaves show remarkable resemblance to those of P. Dickieana, Watt both in shape and in the peculiar glandular markings on the under-surface.

vx) Primula violacea, W. W. Sm. et Ward.

Species ex affinitate P. apocliiae, Balf. f. et Forrest atque P. cyananthae, Balf. f. et Forrest a quibus foliis utrinque minute scabridulo-pubescentibus inter alia divergit.

Planta 20-25 cm. alta foliis floribusque coaetaneis, squamis basalibus sub anthesin destituta. Folia 5-7 cm. longa, 1-1.7 em. lata, petiolo laminam subaequante; lamina oblanceolata vel lanceolata, apice obtusa vel subrotundata, basi in petiolum alatum sensim angustata, margine nunc fere integra, nunc obscure denticulata, nunc irregulariter dentata, per totam superficiem minute scabridulo-pubescens in sicco tenuiter papyracea. Scapus ad 25 cm. altus gracilis epilosus sub inflorescentia albo-farinosus spicam florum sessilium pen- dulorum numero plerumque 6-10 parvam gerens; bracteae occultae lineari-oblongae albo-farinosae. Calyx cupularis 3-4 mm. longus ad medium in lobos inaequales alios (in spica exteriores) subrotundatos atro-purpureos extra farina sparsos intra dense farinosos alios (in spica interiores) oblongos pallidos utrinque farinosos fissus. Corollae saturate violaceae tubus

_ eylindricus circ. 12 mm. longus supra leviter ampliatus exannulatus epilosus, lobi circ. 2.5 mm. longi subquadrati vix

_ emarginati. Stamina in flore brevistylo prope os inserta. _Ovarium globosum; stylus in flore brevistylo vix ad medium ubum attingens. Capsula deest. |

4

14 4

“y @&

‘Primula werringtonensis, Hort. Wallace. Descr. Forrest.

\

SmitH—NEW PRIMULACEAE. 55

“West China :—lLitang-Yalung divide, S.W. Szechuan. Alt. 12-13,000 ft. Flowers deep violet, powdered inside with white meal, which is much more abundant than in the other Muscarioid primulas met with, so that the inside of the corolla is quite white. Bracts and upper calyx lobes light chocolate. On steep shaded slopes under Rhododendron, on the Yalung side of the divide, facing N.W. Also amongst boulders and_ scrub, sheltered but not shaded. Seen nowhere else but in this one spot, some 50 plants being senate July 1921.’

Ward. No. 4386.

Very much resembling P. apoclita, Balf. f. et Forrest, but with different leaves, closely beset on both sides with a very minute indumentum and lacking the long, white hairs so characteristic of the above species and its near allies.

Microforma P. obconicac, Hance aestimanda ; foliis basi altius cordatis textura tenuioribus minus hirsutis, floribus intensius coloratis, corollae tubo longiore et angustiore, e calyce longe exserto a typo recedit; in silvis umbrosis in montibus Likiang dictis in provincia yunnanensi atque prope Muli in eisdem locis in provincia szechuanica crescitur; forma silvicola occidentalis plantae typicae esse videtur.

“West China :—Lichiang Range, Yunnan. Lat. 27° 30 N. Alt. 9-10,000 ft. Plant of 6-12 inches. Flowers rose; eye greenish-yellow, fragrant. On rocks in very shady eens in mixed mpi pes coniferous forests. June 1913.” G. Forrest. No.

“Mountains Nw of Chien-Chuan, Yunnan. Lat. 26° 50’ N. July 1914.” _ Forrest. No. 12767.

Lichiang ater ta fruit. G. Forrest. Nos. 10674, 11455, 15335-

Omphalogramma elegans, Forrest.

Affinis O. Souliei, Franch. atque O. Forrestii, Balt. f. et fortasse prioris varietas rectius habenda; ab O. Souliei foltis ad anthesin vix bene evolutis supra longiuscule pilosis, calycis lobis sublinearibus acutis divergit; ab O. minore, Hand.-Mzt. autem valde affini fructu duplo majore recedit.

Planta sub anthesin 15-25 cm. alta, in fructu ad 60 cm. basi Squamis circumdata. Folia plerumque post flores evoluta, nonnunquam fere coetanea, petiolo 3-4 cm. longo latiusculo pilis longis articulatis bene induto suffulta; lamina 4-6 cm. longa,

2-3 em. lata, ovato-lanceolata vel ovata, apice obtusa, basi subabrupte angustata haud cordata, supra pilis longiusculis albidis

——. infra sparsius ad costam nervosque,

56 SMitH—NkEwW PRIMULACEAE.

integra vel undulata ciliolata. Scapus dense pilosus, sub flore rufo-pilosus. Calyx circ. 9 mm. longus fere ad imum in lobos plerumque sex sublineares 1-1.5 mm. latos acutos fissus plus minusve dense pubescens. Corollae saturate purpureae tubus cire. 3 cm. longus evlindricus gradatim superne ampliatus extra pilosulus intra minute pubescens exannulatus, lobi obovati circ. 2 cm. longi et 1.5 cm. lati apice bilobulati fa sub-irregulariter incisi, undique et praesertim ad margines pilis articulatis Si edelogia obsiti. Stamina 6 ad medium tubum inserta fere ad os attingentia filamentis liberis cire. 6 mm. longis glabris, antheris 6 mm. longis. Stylus glaber stamina paululo superans ; stigma compresso-capitatum ; ovarium ovoideum ,glabrum 6 mm. longum. Capsula 2 cm. longa, fere 1 cm. lata, cylindrica seminibus multis disciformibus ad 3 mm. diametro pallido- brunneis; planta ad tempus fructus maturi valde elongata, ad 60 cm. alta, petiolis nune ad 12 cm. extensis, lamina anguste pines saepe 13 cm. longa.

ibet iiothinehiong, Salween divide, sarong. Lat. 280 - 12’. N. Long. 98° 24’ E. Alt. 14,000 ft. Plant of 8-10 inches. Flowers deep purple-blue, tube yellowish. Alpine pasture on the margins of scrub and peaty bogs. July 1921.” G. Forrest. No. 19979.

“The same in fruit. Plant of 2 ft. Oct. 1921.’ G. Forrest. No. 20857.

“West of Chamatong. Lat. 28° 18’ N, Long. 98° 27’ E. Alt. 12,000 ft. Plant of 6-9 inches. Flowers rich deep imdigo- purple. On moist a on the margins of and amongst dwarf

scrub. June 1922 G. Forrest. No. 21793,

In fruit. Oct « 1922. G. Forrest. No. 22812.

Omphalogramma minus, Hand.-Mzt. in Sonderabd. aus Anzeig. Akad. Wien, No. 26-27 (1923), p. ce The following appear to belong to the above recently described apsces collected in approximately the same area :

‘S.E. Tibet :—Salween-Kiuchiang divide, Tsar sarong. Lat. 28° 12’ N. Long. 98° 24/E. Alt. 14- 14,500 ft. Plant of 2-4 inches. Flowers throughout deep indigo-purple. Sheltered situations by streams amongst boulders. July agers: G:

Forrest. No. 20047.

- “The same in ~*~ Plant of 4 13 inches. Sept. 1921.”’ Ge Forrest. No. 20

: fe ieee Kiaokinns. re ide west of Chamatong. Lat. 28° 18’ N.... Long. 989: 27'-B.-/ Alt. 11- 12,000 ft. Plant of 2-4 inches. Open stony pasture amongst rocks by G. Forrest. No. 21795. Oct. 1922. G. Forrest. No. 22864.

The Leaf Glands of Dioscorea macroura, Harms. BY MATTHEW YOUNG ORR.

With Plates CXCVII-CXCVIII and Three Figures in the Text.

Dioscorea macroura, the peculiar leaf glands of which have been made the subject of an extended research by the writer, is a West African species, and a native of the tropical rain-forests in the region bordering on the Gulf of Guinea,

First collected by Zenker in the Cameroons in 1891, it was described as a new species by Harms in 1897. In his diagnosis, which appeared in the Notizblatt* at Berlin, Harms lays par- ticular emphasis upon the extraordinary development of the acuminate leaf-apices, which he refers to in the following terms :—‘‘tiberall tritt an den Blattern eine auffallig lange, schmale, mehr oder minder scharf abgesetzte, verdickte, schwanzartige Spitze hervor.’’ ‘This highly developed acumen presents its most interesting aspect, however, as a study in physiological plant-anatomy, for in its internal construction there appear features that are perhaps unique within the genus. Its chief point of interest lies in the complex glandular system, which traverses its length from base to apex, and which differs markedly from the ‘‘sunken glands”’ of other species, so fully described by Correns + in his work on the extra-nuptial nectaries of Dioscorea. Moreover, the mucilaginous secretion which fills the lumen of these glands harbours a species of bacterium, which is always present in great numbers, even in the living leaf.

This constant association of a bacterium with the secretory organs of Dioscorea macroura is not without a certain physio- logical import, and may prove to have a direct bearing on the economy of the plant. In this, and in certain other features, the glands of this species certainly show a striking resemblance to the ‘‘bakterienknoten’’ found in the leaves of other tropical plants, the essence of which, according to continental investi- gators, is a symbiotic union between the plant and a ah micro-organism. Among the recorded oe of ‘*bakteri

* See Notizbl. k. bot. Gartens u. Mus. zu Berlin, i. (1897), p. 2

+ C. E. Correns in Sitz. Ber. Math. Nat. Cl. Akad. Wien. Lege Sk: 1 (1889), p. 651. (Notes, R.B,G., Edin., No. LXVIII, November 1923.]

58 ORR—LEAF GLANDS OF DIOSCOREA MACROURA.

knoten,’’ that of Psychoiria alsophila, described by Boas,* has a particular interest in this respect, since this species was also collected by Zenker in the Cameroons. :

The. plants of Dioscorea macroura, from which the material for this investigation was obtained, have been in cultivation in the hot houses at the Royal Botanic Garden for some years; the tubers, from which the original plants were raised, having been brought from Nigeria by Dr. J. M. Dalziel, and presented by him, to the Garden, along with other plants from the same area.

By adopting methods of intensive cultivation, in a moist atmosphere at a temperature ranging from 60 to So degrees Fahrenheit, it has been possible to raise, in successive seasons, plants of ever increasing size and vitality. Last year, the climb- ing stems of this species reached a length of over 18 metres, while the laminae of the largest leaves averaged 15 cms. long by 30cms. broad, with leaf-tips fully 8 cms. in length. So far, the plants have never flowered, but, each year, they have been singularly prolific in the production of stem bulbils, and these have been made use of to propagate the species. ‘The plant also perennates by means of a ‘‘nest’’ of root-tubers, which are said to be of a poisonous nature.

Reference has already been made to the remarkable length of the leaf-tip of this species, and it will not be out of place, at this stage, to describe more fully the distinctive external features which are so closely correlated with its glandular nature.

In the mature leaf, the acuminate tip is thicker than the lamina, is rigid, and tapers gradually from its base to the pointed apex, which is usually directed downwards. It may reach a length of 8 cms. and is about 5 mm. broad at its junction with the lamina.

at the extreme tip. Between the veins. the intervening tissue is slightly swollen, and forms a series of cushion-like ridges, from four to six in number, which are yellow-green in colour

_ and quite opaque, in contra-distinction to the leaf-lamina, which is of a dark green colour and more or less translucent They are, if anything, more prominent on the under side of the acumen, while, on the upper surface, each ridge is tudinal slit, which communicates with the

ORR—LEAF GLANDS OF DIOSCOREA MACROURA. 59

enclosing the glands, the latter roofing over the interposing veins.

The innermost pair of these glandular ridges extends practi- cally the whole way from the base of the acumen to its apex, but those more lateral in position, by a repeated process of fusion with their immediate neighbours on the inner side, lose their identity in turn, until all are merged in the two innermost, persistent, glandular tracts.

Near the margins of the acumen, there project, from its upper surface, two thin flanges of tissue, which, in the older leaves, are directed outwards; but, in the younger leaves, the entire marginal portions of the acumen, from which the flanges arise, bend inwards, and completely enclose the upper surface and its longitudinal slits. As the leaf matures, the involuted margin gradually unfolds, and the slits become exposed, .

The disparity in size between the young leaf and its acumen is very marked, for example, an undeveloped lamina, 7 mm. long, may be surmounted by a tail-like tip of 40 mm. in length. It does not follow, however, in spite of this disproportion in size, that the acumen necessarily becomes physiologically functional before the lamina reaches maturity, for it can be shown that, at this stage, the glands of the acumen are still in process of formation, and they only reach their highest state of development when the leaf-blade has expanded to more nearly its adult proportions.

Before proceeding to a consideration of the anatomy of this specialised organ, it might be well, at this point, to refer briefly to some of the technical methods employed.

The fluids used for the fixation of the parts to be examined were acetic-alcohol and chromo-acetic solution, and the material was embedded in paraffin, and sectioned by means of a micro- tome. By piecing together the serial sections obtained by cutting the material in different directions, it was possible to secure a mental picture of the convolutions of the glandular svstem, which was helpful in elucidating its construction. Various stains were employed, but perhaps the best results were obtained by using safranin, iron-alum haematoxylin with a contrasting stain, or a combination of gentian-violet and eosin. For the examination of the micro-organism present in the glands, the usual bacteriological methods of staining were followed.

The general anatomy of the leaf-tip can best be illustrated by reference to Plate CXCVII, which is a photograph of the

_transverse section of a medium-sized acumen, cut from near the

apex. In this, the position of the glands, their relationship to

ae the other tissues, and the general configuration of the organ are clearly demonstrated,

60 OrR—LEAF GLANDS OF DIOSCOREA MACROURA.

The glands are the dark-coloured areas, the contrast being produced by the deeply- staining secretion, which fills the glandular ‘‘pockets.’’ It will be observed that only two of these glandular areas appear in the section, in place of the four or six referred to in the description of the external features of the acumen. ‘This reduction in number is accounted for by the process of fusion which has been shown to take place between the outer and inner glandular ridges, resulting in the formation of the two combined glands, which alone persist in the distal portion of the acumen. Each gland is an elongated ‘“‘ pocket”’

_in the mesophyll, which, in transverse section, has an extremely irregular aiiiae due to the frequent infolaiage of the wall- layers. Its truly convoluted character, however, can only be fully appreciated by following its course through a series of longitudinal sections.

This ‘‘ pocket’? communicates with the upper surface of the acumen by way of the longitudinal slit, referred to above, the epithelium of which is epidermoid in character, and is strongly cuticularised, judging by its reaction to microchemical tests, and this property is shared, but to a less degree, by the cells which form the lining of the gland cavity itself. In the apical region of the acumen, from which this section was taken, the undulating character of the upper surface is not so pro- nounced as it is nearer the base, and the ducts, which, in the section, appear as narrow, winding canals, open ies into a shallow channel formed by the two marginal flanges of parenchyma.

Internally, the opposing epithelial lavers of the duct diverge abruptly to form the roof of the gland cavity, turning upwards in their course until the upper limits of the gland are reached, and thus partially enclosing a patch of mesophyll, roughly semi-

circular in outline, with a much-indented circumference due to frequent infolding of the limiting layer. From the extremities of the gland, the epithelium is continued downwards to form the curved floor of the crypta, from which the secretory tissue o* the gland is developed The cells which compose this superficial layer, with the possible exception of those which form the basal attachments of the secretory elements, have not only their outer, but also their lateral walls cuticularised, somewhat after the manner of an

; _endodermis, while their inner walls remain unaltered. Some-

he investigate. This distribution of the cutin in the cell-walls P athe obvious in sections stained with safranin or ia blue.

OrR-—LEAF GLANDS OF DIOSCOREA MACROURA. 61

Underlying this cutinised epithelium are two or three layers of smaller, thin-walled cells, which form a compact sheath, partially enclosing the gland, and separating it from the larger, chlorophyll-containing cells of the mesophyll. This sheath does not extend beyond the upper extremity of the gland.

The actual secretory tissue of the gland consists of a large number of vermiform, filamentous outgrowths, each the product of a single epithelial cell. ‘These are produced solely from the basal epithelium, from which they project into the lumen of the gland, there becoming intertwined to form a tangled plexus, the interstices of which are filled with a tenaceous fluid, possessing a strong affinity for such stains as corallin, Hoffman’s and methylene blue. In the living state of the plant, this fluid forms a medium, which appears to be eminently suited to the growth and development of the bacterium which abounds in it.

Fic. 1,.-A portion of the secretory tissue of a mature gland. sc., vermiform secretory elements ; b., fluid content, with bacteria ; ep., cuticularised epithelium. x 250. In the mature gland, these filamentous outgrowths are sub- divided by cross septa into as many as six elongated cells, which are nearly circular in section, but, unlike the epithelial cells, they have thin walls, although the basal cell of each filament may show a certain amount of cuticularisation in its lateral walls. The nuclei of these cells are particularly prominent, and usually there are one or more large vacuoles in the cytoplasm. An enlargement of some of these features of the glandular tissue will be found in fig. 1.

62 Orr—LEAF GLANDS OF DIOSCOREA MACROURA.

The mode of origin of these cell-filaments will be touched upon when dealing with the development of the gland, after the remaining features of the mature acumen have been disposed of.

In the parenchyma, immediately adjoining the glandular tissue on its upper side, there are numerous air-channels, which follow the convolutions of the gland in a longitudinal direction, thus forming a complex aeration system. Each of these channels is enclosed by a definite sheath of parenchymatous cells.

Likewise intimately associated with the glands are the vascular bundles, which lie parallel to them, and, at frequent intervals, send out smaller, lateral veins, which encircle the lower rim of each gland, being separated from the actual glandular tissue only by the parenchyma sheath. In these small vascular strands, there is no apparent inequality in the amount of xylem and phloem developed to meet the requirements of the gland.

Ler

25

fa

oe oP

Fic, Blgrimdag ayn of the acumen of a young leaf, with developing gland. sc., cre sue, in proc See's : te f PiguiNe P ie ee f.. * flange” tion ; d., duct ; e., ‘‘corrugated epidermis ;

ge” of tissue. x 350. a Here and there in the surrounding tissue, are to be seen large a oval cells, each containing a bundle of raphides, invested with a - mucilaginous envelope. Other cells containing tannin occur in ‘ the mesophyll, just below the glands, and are particularly con- Spicuous in material fixed in chromic acid, Stomata are

Orr-—L&ar GLANDS OF DIOSCOREA MACROURA. 63

singularly infrequent in the epidermis of the acumen, and, when present, are confined, more or less, to the margins.

To investigate the origin of the glands, it was found necessary to select very young leaves from the region just behind the growing-point of the shoot, for the initial stages of development are passed through almost before the young leaf unfolds. The comparatively advanced stage represented by fig. 2 was obtained by cutting sections of an acumen, less than 4 mm. in length, and only .7 mm. broad.

The first indication of the glands is the appearance, in the upper epidermis, of a simple invagination, interposed between the two flanges which, at this stage, are very minute and close together. The cells which form the walls of this pit resemble those of the epidermis, except that the corrugations, so obvious on the outer surface of the latter tissue, are absent entirely from the superficial layer of the gland.

In the youngest leaves, the opposing walls of Me invagination aré in. close approximation to one another, but, as the acumen increases in size, the slit widens, and the epithelial cells at the base of the pit increase by radial divisions to form a flattened, flask-shaped cavity. The cells forming the floor of the aatity are rich in protoplasmic contents, and become the initials of the vermiform, secretory elements. Each initial cell, by growth in a radial divert: increases in size until it proteudes into the lumen, when division of the protoplast takes place. The pro- jecting portion of the cell is then cut off by a periclinal wall, and becomes the first cell of the filament. This process is repeated in a series of intercalary cell-divisions until each filament attains to its full dimensions.

A certain amount of fluid is present in the developing gland, even at this stage, and, in fact, it appears almost at its inception, for the growing apex of the stem, and the leaf primordia, are bathed in it, a condition which has a particular significance, since this fluid also harbours the bacterium found in the fully- developed gland. (Plate CXCVIII., fig. 1.

About this time, nearer the base of the rapidly enlarging acumen, there are indications of the formation of additional gland-cavities by a process of subdivision; the new “‘ pockets” being formed directly from the pre-existing ones, not by invagination, but by an inverse process. This is brought about by active grow th in the band of tissue immediately underlying the long axis of each concavity, resulting in the elevation of the central portion of the floor to the level of the epidermis, and thus forming a ridge of tissue, several cells across, which bisects the Original pocket longitudinally, and divides the now greatly distended duct into two. With the growth of the acumen, these

64 OrR—LEAF GLANDS OF DIOSCOREA MACROURA.

“twin’’ glands diverge, and the process is repeated until all the glands found in the fully-developed organ have come into existence. At a later period of development, the epithelium forming the roof of the gland becomes extensively folded, owing to the excessive multiplication of the cells composing it, which has the effect of increasing the internal area of the gland-cavity, while, externally, the hollows between the folds are filled in by the unequal growth of the adjoining ground tissue.

Bo eax

vA

a

Ren he

FiG. 3.—Transverse section of a portion of the glandular region at the base of a developing acumen, showing two immature glands, which have been formed by the sub- division of the original gland cavity. sc., secretory tissue; d.d., ducts; e pidermis ; f., base of ‘‘ flange.” x 350.

Meanwhile, other cells of the peripheral layer encircling the lower half of the gland undergo tangential division to form the compact, parenchyma sheath, which, together with tne epithelial layer, separates the lumen of the gland, on its under side, from the associated vascular tissues, and the adjacent less-compacted mesophyll.

The very young leaf, and its acumen, are covered with minute, multicellular scales, which are most obvious in sections of the bud, where they appear immersed in the deeply-staining fluid. They are notably evanescent, and are wanting in the fully- developed leaf. These epidermal outgrowths bear a striking resemblance to the immature secretory elements, which are,

ORR—LEAF GLANDS OF DioSCoRBA MACROURA. 65

themselves, the products of a tissue that is epidermal in origin, the chief point of difference between the two being, that while the former are temporary structures, the latter become per- manent constituents of the glandular tissue of the acumen. It is not unreasonable to assume, therefore, that the secretory filaments of the gland are merely persistent epidermal scales, which have been retained, and modified, for a definite physio- logical purpose.

Attempts to determine the chemical composition of the fluid content of the gland cavity were rendered difficult by the presence of the micro-organism, and to its influence, no doubt, may be attributed some of the unexpected results obtainell.

At no time was there any visible exudation of fluid from the external orifice of the duct, and chemical tests had therefore to be applied directly to sections of the acumen, by immersing them in a small quantity of the reagent, on a slide, or in a test tube. In this way, a study of the biochemistry of the secretory organs, and associated tissues, was accomplished, the results obtained being briefly collated below.

Fairly thick sections were treated with Fehlirg’s solution, alone, and with the addition of an acid, and were then heated, the process being conducted on a slide in the manner recom- mended by Schimper, but with purely negative results. Barfoed’s solution likewise produced no visible reaction in the tissue of the gland, or its secretion. This failure to detect the presence of sugar, in some form or other, was unexpected, although Correns remarks that he was unable to find either sugar or starch in the developing extra-nuptial nectaries of those species of Dioscorea, that he investigated. In the present case, starch was located in the cells of the mesophyll, but it was entirely absent from the glands which were coloured a bright reddish-brown by iodine, suggesting the presence of proteids.

The application of the standard tests for the recognition of proteids produced somewhat surprising results. When sections were placed in a drop of Millon’s reagent on a slide, and heated, a dense brick-red precipitate appeared within the glands. Treated in the same way with hot nitric acid and ammonia, the glands responded at once to the xanthoproteic reaction. They were coloured red by the so-called Raspail’s reagent, and were turned to a violet colour by the more delicate biuret reaction.

The glands were the only parts of the acumen affected by these reagents, and, while it is recognised that the successful appli- cation of any one of these tests does not in itself afford conclusive proof of the presence of proteids, yet, taken together, the combined results are undoubtedly indicative of the gland content _ being of an albuminoid nature.

66 OrrR—LEAF GLANpS oF DIoSCOREA MACROURA.

In this particular, the glands of Dioscorea macroura resemble the ‘‘ proteid-containing glands’’ of Ardisia, first described by Hohnel.* ‘These glands are likewise the home of a micro- organism, whose relationship to the plant, particularly in the case of Ardisia crispa, has been subjected to a critical investi- gation by Miehe.+ Other analogies exist, as exemplified by the parallel features of the glands in certain Rubiaceae, described by Zimmerman} and Boas. ‘This aspect of the glands will be discussed at a later stage.

Standard tests for the detection of tannins were applied to thick sections of the acumen, but no trace of any such astringents could be found within the glands, which is in accordance with the findings of Correns, who failed to discover any tannins in the extra-nuptial nectaries. On the other hand, certain large elements of the ground tissue gave characteristic tannin reactions with iron salts, potassium bichromate, ammonium molybdate, cupric acetate and osmic acid. With Schultze’s solution, the whole of the glandular tissue was coloured dark- yellow, and the cuticularisation of the epithelial cell-walls became apparent, but neither with this reagent, nor with iodine and sulphuric acid, could a definite cellulose reaction be obtained in the thin cell walls of the secretory trichomes. No trace of oil could be found in any part of the acumen.

The pronounced proteid character of the gland content, as disclosed by these microchemical tests, was further emphasised by a comparison of the nitrogen content of the acumen with that of the lamina. This was ascertained by the Kjeldahl method of analysis, applied to leaves taken from the living plant at intervals during the day. In every case, the tissue of the acumen was found to contain a much higher percentage of nitrogen than the tissues of the lamina. ‘The following table, in which the results of two of these nitrogen estimations are compared, will serve to illustrate this point.

Nitrogen Content in 100 grams of Leaf Tissue.

Time. amina. Acumen. Excess of N,

| Content in Acumen. 10.50 a.m. 246 mgs. 560 mgs. 314 mgs. 4.30 p.m. 173 mgs. 486 mgs. 313 mgs.

The remarkable difference between the nitrogen content of the acumen and that of the lamina, as shown by this process, could only be accounted for by the existence within the former, and Ya “Mice in Ber in Sitz. Ber. Wiener Akad., lxxxiv., abt. 1 ( 1882), pp.

tA Zimmerman a

PP- 574, 58 3°

I. xxix. i 156 and hrb. fiir wiss

nin jeer Sn a oe p. 29. )P. ? wae tain XXXVI. (1 903), p-

OrR—LEAF GLANDS oF D1osCOREA MACROURA. 67

‘couversely by its absence from the latter, of some factor capable of bringing about this increase in the amount of combined nitro- gen in the tissues of the acumen. Bearing in mind the proved proteid character of the glands, an explanation for the increase was sought for in the activities of the bacterium associated with them.

A necessary preliminary to a study of the organism and its vital processes was its isolation from the plant, and this was accomplished by adopting, with some modifications, the methods employed by Harrison and Barlow,* and others, to isolate the bacterium from root-nodules. The acumen of a living leaf was first sterilised by dipping it into a fluid containing small pro- portions of hydrochloric acid and mercuric chloride. After washing with distilled water, it was cut across with a sterilised scalpel, and by means of a platinum loop, a small quantity of the gland content, with the included bacteria, was then trans- ferred from the cut surface to a test tube containing a liquid culture medium. From this, other tubes were inoculated, and pure cultures of the bacterium were obtained. The usual pre- cautionary measures, to guard against contamination, were taken at each step in the process, and controls were set up for all subsequent cultures, and these remained perfectly sterile.

Various culture media, both liquid and solid, were used for the cultivation of the bacterium, but it was found that liquid media gave the best results in the first transference from living leaf to artificial conditions. ‘The following culture fluid, which has been used in the investigation of root-nodule bacteria, proved very successful in the initial stages of isolation.

saccharose, I.0 grm. acid potassium phosphate, 0.5 grm. magnesium sulphate, 0.02 grm. distilled water, roo ce.

Substitution of other sugars for saccharose did not produce the Same amount of growth, and glucose was particularly unsatis- factory, but excellent results were obtained by using a sterilised decoction of the leaf tissue. Only neutral, or slightly alkaline solutions, were employed throughout the experiments.

After incubation for a period of 36 hours, at a temperature of 25° C, the inoculated fluids become distinctly turbid, and at the end of 3 days, the cloudiness took the form of a dirty-white growth of a somewhat ropy nature, which collected at the bottom of the tube. Microscopic examination of an air-dried drop of this milky liquid, after staining with carbol-fuchsin, or gentian- * Harrison and Barlow in Proc, Roy. Soc. Canada, 1go6.

68 OrR—L&EAF GLANDS OF DIOSCOREA MACROURA.

violet (Gram’s method), showed that it consisted of a pure culture of rod-shaped organisms, from 1.6m to 2.4u in length, and from 64 to .8u in breadth. ‘They proved to be Gram positive, and it is interesting to note further, that the gentian-violet stain re- mained in the organism, after dehydration with amyl alcohol, but not after ethyl alcohol, resembling in this particular the similar reaction obtained by Harrison and Barlow on Pseudomonas radicicola. About 24 hours after inoculation, the organisms were actively motile in liquid media, and by careful staining, using the van Ermengen method * in a_ slightly modified form, it was possible to determine the presence of a single polar fageMuts on the bacterium cell.

Stab, smear and plate cultures were also made on agar and palatin, by inoculation from liquid media. The agar was pre- pared by adding 2 per cent. of the substance to the nutrient solution containing saccharose; in the case of gelatin, the per- centage was a little higher. The agar cultures were incubated at 25° C, but those on gelatin, were, of necessity, developed at a lower temperature. It was found, however, that within certain limits, the difference in temperature did not appear to have much influence on the rate of growth.

On saccharose agar, the colonies were round or oval, from 3-5 mm. across. Raised above the substratum, they were hemis- pherical in form, sabaceous, faintly reticulated, and had a slightly fimbriated margin. On nutrient gelatin, the colonies were ochraceous, but otherwise they resembled those produced on agar. Smear cultures, on the other hand, rapidly became agglomerated to form nodulose masses (Plate CXCVIIL., fig. 2). In stab cultures, the growth was strongest at the point of entrance, and relatively weak internally, along the edges of the stab.

On agar and gelatin, the colonies made their appearance in from 2-4 days after inoculation, the time taken varying with the composition of the medium kind as a source of infection. For example, cultures on nutrient gelatin, initiated by inoculation from solid media, became apparent in 24 hours, while, under the same conditions, cultures made from liquid nied coauined from

4-5 days to reach a similar state of development.

The culture of the organism, under artificial conditions, was maintained over a period of three months, at a mean temperature of 15° C, and during that time, cross peldctions between different media were carried out, the cultures, in all cases, remaining pure.

That the organism is aerobic was suspected from its behaviour in stab cultures, and this was confirmed by the impossibility of * See Centr. fiir Bakt. xv. (1894), Pp. 969.

Orr—LEAF GLANDS OF DIOSCOREA MACROURA. 69

inducing it to grow in a Buchner’s tube, over strong pyrogallic acid in caustic potash. Liquefaction of gelatin was brought about after a comparatively long period of growth, ranging from 18 to 28 days, the liquefied gelatin subsequently becoming cloudy. Sterilised milk, to which a culture of the bacterium had been added, ultimately became clear and translucent, and gave a slightly acid reaction with litmus. The transference of a dro of this fluid to nutrient gelatin resulted again in the production of a pure culture of the organism

In suitably stained sections ra the glands of Dioscorea macroura, relatively large, irregularly shaped bodies, resembling the bacteroids of leguminous nodules, had been observed, and these also appeared in the cultures on artificial media. Similar involution forms were found by Faber* in the bacterial leaf- glands of Pavetta and Psychotria. "This author carried out an extended research on the physiological attributes of the organism inhabiting these ‘‘bakterienknoten,’’ and, as a result of his experiments, he came to the conclusion that the ‘‘ mycobac- terium,’’ isolated from the glands, possessed the power of utilising free nitrogen, and that the inter-relationship of plant and micro-organism. was therefore of the nature of a “ bakteriensymbiose..’

A similar view of the intimate relationship existing between Ardisia crispa and its associated bacterium was expressed by Miehe, after a lengthy series of observations on the biology of the micro-organism and its effect on the growth of the plant.

In the light of these researches, and having regard to the several points of resemblance between the glands of Dioscorea macroura and those of other tropical plants investigated by Faber and Miehe, it was essential that the Dioscorea organism should be subjected to similar tests, in order to determine whether, or not, it was endowed with like properties, and capable of bringing free nitrogen into combination. Kjeldahl] determinations of the nitrogen content of cultures of the bacterium, isolated from the glands, were therefore compared with similar estimations of the amount of nitrogen present in the culture medium in sterile controls. In every case, there was evidence of an appreciable gain in the nitrogen content, presumably due to the activities of the organism.

he percentage increase in nitrogen varied according to the nature and composition of the medium employed, and seemed to be inversely proportional to the amount of combined nitrogen present in the nutrient medium at the time of inoculation, e.g a nutrient medium with a low initial nitrogen content, three weeks after inoculation, showed a percentage increase in nitrogen * F.C, von Faber in Jahrb. fiir wiss, Bot. li. (1912), p. 283, and liv. (914), p. 243.

70 Orr—L&ar GLANDS oF DioSCOREA MACROURA.,

of just over 100, while a medium of a different composition, and a higher initial nitrogen content, showed a gain of only 67 per cent., although, in the latter experiment, the culture had been maintained over a much longer period of time.

The power of nitrogen fixation possessed by the Duoscorea organism was tested up to fully 80 days after inoculation, but its effect on the nitrogen content of artificial media is typified by the following result, obtained by the Kjeldahl analysis of a 23 days culture growth.

Nitrogen Content in roo grams of Culture Medium.

In the control, 23.18 mg. In the apes medium at con-

clusion of experiment, 48.61 mg. Gain, 25.43 mg.

The increase in the nitrogen content may appear small when contrasted with the results obtained from similar experiments carried out with nitrogen bacteria from root-tubercles. It com- pares not unfavourably, however, with some of Faber’s estimations of the nitrogen increase in culture solutions, brought about by the organism isolated from Pavetta Zimmermanniana. He found that, with a nitrogen content of 12.256 mg. in the control, the gain in three separate experiments was, respectively, 9. mg., 19.592 mg. and 12.750 mg. of nitrogen per 200 Cc. of culture solution. ‘The duration of each experiment was 20 days. Using an almost nitrogen-free culture solution, Faber found that the increase in nitrogen, for the same length of time, was much greater than the amounts quoted above, while with a higher percentage of nitrogen in the control, the increase was correspondingly less.

That the bacterium found in the glands of Dioscorea macroura stands in the same relationship to the plant, as do those organisms, described by Faber and Miehe, to the plants with which they are allied, is an interesting hipothesiw. which gains support from the behaviou? of the isolated bacterium in culture media. But whatever powers of nitrogen-fixation may be possessed by the organism, when divorced from the plant, it is quite another matter to state categorically, that, within the plant, the bacterium plays the réle of nitrogen provider, in the same

way as do the bacteria in root-nodules, although the high soba oue content of the acumen does seem to postulate a certain activi

It may be merely a case of contingent symbiosis, or a fortuitous

_ association, such as that mentioned by Koorders* in his description of the ‘‘ wasserkelche’’ of tropical plants. S, H. Koorders in Ann, Jard. Bot. Buitenzorg, xiv. (1897), p. 451.

_—

OrR—LEAF GLANDS oF DIOSCOREA MACROURA. 71

On the other hand, this association of a bacterium with Dioscorea macroura has every appearance of being of the nature of a more permanent alliance, for, as it has been pointed out already, the organism is present in the secretion enveloping the bud which originates from the tuber, by which avenue it 1s apparently able to reach the developing glands of the new plant.

All attempts to grow the plant from a sterilised tuber, under perfectly sterile conditions, have so far proved abortive, and it is therefore impossible to say what effect the disassociation of bacterium and plant would have upon the growth of the latter.

Although dissimiiar in structure and mode of origin to those of Ardisia, Pavetta and Psychotria, the leaf-glands of Dioscorea macroura undoubtedly possess some of their peculiar physio- logical properties. They have been shown to differ in many ways from the extra-nuptial nectaries of other species of Dioscorea, and in one particular—their bacterial nature—they are possibly unique within the genus.

In conclusion, the writer desires to record his indebtedness to Dr J. M. Dalziel, for much helpful information regarding the plant in its native habitat, and to Mr I. H. Burkill, Director of the Botanic Gardens, Singapore, for his invaluable assistance in the matter of identification. An expression of thanks is also due to Mr R. M. Adam, for the photograph of the culture, reproduced on Plate CXCVIII., and to Mr J. J. Campbell, for the care expended on the cultivatién of the plants, which furnished the material for this investigation.

SUMMARY.

1. The leaf-glands of Dioscorea macroura form a series of elongated “ pockets”’ in the mesophyll of the prominent acumen.

2. The ‘‘pockets’’ are formed by a process of invagination followed by subdivision, and when mature communicate with the upper epidermis by way of narrow slits or ducts.

3. Each glandular ‘‘pocket’’ is lined, in part, with a secre- tory tissue composed of vermiform, multicellular trichomes, and is filled with a mucilaginous fluid which gives a strong proteid reaction.

4. The fluid content of the gland cavity is the home of a bacterium which is always present in great numbers in the glands of the living leaf. ;

5. The bacterium, when isolated in pure cultures on artificial media, is found to possess the power of nitrogen-fixation to an appreciable degree.

. The nitrogen-content of the glandular acumen 1s greatly in excess of that of the leaf-lamina, postulating a similar process of nitrogen-fixation taking place within the glands

2 OrrR—LEAF GLANDS OF D1OSCOREA MACROURA.

‘The presence of nitrogen-fixing bacteria in the glands of Dioscorea macroura would appear to constitute a parallel case to that of such a species as Psychotria bacteriophila, in which the foliicolous micro-organism is a source of nitrogen supply to the plant.

EXPLANATION OF PLATES CXCVII-CXCVIII. Illustrating Mr M. Y, Orr’s paper on Dioscorea macroura,

PLark CXCVIL. Leaf-acumen of Dioscorea macroura, in transverse sectio g.g., glands, with en secretory elements oa Srey “stained fluid c ; d.d., ducts, leading to the upper surface; p.s., paecleiad ‘sheath, en circling the d; ac., air chat inel 5 r., raphi ides; v.b., vascular br nd s.t., stoma; f., ‘‘flange’’ of tis sue Pate CXNCVIIT. Fig. 1 “Apical region 1 of a shoot of Dioscorea macroura, in tudinal ction. Big spiced eristem 5. ne containing bac - : ; t., epidermal scale; raphides. Fighs% ieee, culture of the paged ted isolated from the glands. of Dioscorea macroura, at the point of liquefaction.

NoTeES, R.B.G., EDIN. PLATE CXCVII-

Orr—Leaf glands of Diescerea macroura.

NoTES, R.B.G., EDIN. PLATE CXCVIII.

A New Species of Desmogyne. BY MARY M. B. KNAGG, B.Sc.

Desmogyne angustifolia, Knagg.

‘Species affinis D. neriifoliae, King et Prain, a qua foliis

lineari-lanceolatis longe acuminatis et calyce Jobato haud trun- cato facile distinguitur. : Frutex epiphyticus glaberrimus, ramis dependentibus gracil- ibus. Folia alterna; lamina 6-13 em. longa 0.25~-1.25 cm. lata, coriacea, lineari-lanceolata, apice longe attenuata, basi in petiolum gradatim angustata, margine integra valde revoluta, costa subtus prominente supra impressa; petiolus 0.25-1 cm. longus. Inflorescentia corymbosa floribus 2-8; pedunculi 2-6 cm. longi; pedicelli 1.5-2.5 cm. longi sursum incrassati apice in cupulam 6-8 mm. altam expansi; bracteae triangulares acutae 1 mm. longae. Calyx 3-4 mm. longus, 4 mm. latus, tubulosus, 5-lobatus, lobis circ. 1 mm. longis acutis anguste triangularibus apice valde attenuatis basi circ. 1 mm. latis. Corolla 2-2.5 cm. longa, 3 mm. lata, tubulosa, recta 5-lobata, lobis brevibus anguste triangularibus circ. 2 mm. longis. Stamina 10, epigyna, libera, filamentis brevissimis circ. I mm. longis, antherarum elongatarum erectarum apicibus ad corollae os attingentibus. Ovarium s-loculare, stylo filiformi apice indis- tincte 5-lobulato, ovulis numerosis. Fructus deest.

‘Burma :—Headwaters of the Irrawaddy, Nam Yin. Alt. 2,000 ft. Cup, calyx and lower part of the perianth dull crimson, upper part of perianth pale green with red reticulations. ‘The flowers apparently do not open. The anthers are at the bottom of the tube on a very short filament, and long appendages reaching to the apex of the tube. Epiphytic, hanging in long festoons from trees lining the river. At the base of the stem is a large woody swollen water-reservoir, from the other side of which grow the long roots.’ Coll. F. Kingdon Ward. No. 5566. 3-12-22.

This species differs from D. neriifolia, King et Prain, in having narrower strongly revolute leaves and in the distinctly *obed, not truncate, calyx. Moreover, the flowers are smaller

(Notes, R.B.G Edin., No, LKVIII, November 1923. |

74 Knacc—A New Sprcirs oF DESMOGYNE.

than those of D. neriifolia and the pedicel cup is more pro- nounced, having the pedicel itself less thickened below the cup.

There are now in the Herbarium, Roval Botanic Garden, Edinburgh, two imperfect specimens of Desmogyne. One of these is a fruiting specimen from the herbarium of J. H. Lace, collected by Capt. B. N. Abbay on the N.E. Ranker, Burma, in 1913; while the other, which has a few detached and imperfect donee: was collected in the Nam Jisang-Mali divide, Burma, by F. Kingdon Ward in 1922. These specimens agree with the species D. angustifolia in having a lobed calyx but differ from it in having a narrower pedicel cup and broader foliage. Owing to the small amount of material available and its poor condition it is, however, impossible to name these specimens at present.

Printed under the authority of His Majesty's owe ng Office by Jous PE.

“80 Wiser8)S.0. P.180 375 4/24 G7(B).

FOR OFFICIAL USE: Number Cpe

NOTES

ROYAL BOTANIC GARDEN, : EDINBURGH. NOVEMBER 1924. CONTENTS.

: PAGE Plantae Chinenses Forrestianae : Catalogue of the Plants (excluding Rhododendron) collected by George Forrest eee his fifth exploration of Yunnan and Eastern Tibet in the Years oP By the Staff of the Royal Botanic Gunde Edinburgh. é : en

_ The Royal Botanic Garden 2 i E List of Staff at November, 1924 : : : Joe Rules and Regulations : : : 5 iii Historic Notice ; : ; : ¢ : v Regius Keepers : : : : ix Principal Gardeners Sani 1756 : : : x Features of the Garden. With Key Pin é : : xi Teaching in the Garden ; : . ee Enumeration of Visitors, 1889-1923 : : . XVili Title (with date of publication of separate ees and list of Contents, Vol. XIV.

= - EDINBURGH: PUBLISHED BY HIS MAJESTY’S STATIONERY OFFICE.

SOLD AT THE GARDEN,

And to be ae obs from i J STATION ERY OFFIC B (Scorsi nee, 120 GEORGE STREET, EDINBUR RGH Or Bates any ; Bookse Her,

. ed ei

*..

Plantae Chinenses Forrestianae.

(CATALOGUE of the Prants (excluding Rhododendron) collected by Georce Forrest during his fifth exploration of

Yunnan and Eastern. Tipet in the Years

1921-1922. ) PREPARED BY ba THE STAFF OF THE ROYAL BOTANIC GARDEN,

EDINBURGH.

[Notes R.B.G. Edin,, No,, LXIX-LXX., November 1924.]

PREFACE.

THE area explored by Mr George Forrest during his 1921-22 expedition included certain parts of N.W. Yunnan, S.E. Tibet and 8.W. Szechuan. His collection for those two years con- sists of 3924 numbers of which a large proportion is represented by the genus Rhododendron. The present catalogue is an enumeration of the plants collected excluding Rhododendron. The materials of the latter genus have not yet been fully worked out and will require considerable time for their elucidation.

The majority of the specimens were collected within the con- fines of Yunnan; the remainder in those regions of Szechuan and Tibet which border on N.W. Yunnan. The areas lie be- tween lat. 24° 35’ N. and 28° 48’ N. and between long. 97° 35’ E. and 101° 6’ E.

This fifth expedition supplements the work carried through on previous journeys, and extends the area botanically explored further west and north. On the west Mr Forrest went beyond the Salwin-Kiu Chiang divide, while on the north he extended his survey along the same divide, as well as along the Mekong- Salwin divide. Considerable collections were made in the Salwin valley up to lat. 28° 48’ N.

The Royal Botanic Garden, Edinburgh, owes the possession of this valuable collection to the generosity of . Williams, Caerhays Castle, Cornwall, and of Mr Reginaid Cory, Duffryn Castle, Cardiff, for whom Mr Forrest undertook these two years of botanical exploration. .

io PO ee epee ee

Catalogue.

THE numbers are Mr Forrest’s field numbers as attached to his dried specimens, and the descriptions are copied from his

field labels

19334. Sterculia principis, Gagnep. ‘Tree of 70-90 ft. Flowers creamy yellow, fragrant. Burmese frontier below Tengyueh. Lat. 25° N. Alt. 3~4,000 ft.

19337- Rhodoleia Championi, Hook. f. ‘Tree of 40-50 ft. ‘lowers deep rose-crimson. Open mixed forests on the eastern flank of the N’Maikha-Salwin divide. Lat. 25° 50’ N.. Long. 98° 4o’ E. Alt. 9,000 ft. March 1921. N.W. Yunnan.

1y340. Ostodes Katharinae, Pax, vel aff. Shrub or tree ot 20-40 ft. Flowers creamy-white. In open forests on the eastern flank of the N’Maikha-Salwin divide. | Lat. 25°. 50’ N: Igongi:'98° 40’ E..; Alt.:'8,000 ft. March 1921. N.W. Yunnan.

19341. Lacaitea (Trichodesma) calycosa (Coll. et Hemsl.), Brand. Half shrubby plant of 3-4 ft. Flowers pale greenish-yellow. Open, dry situations on the mar- gins of scrub in the Salwin valley. Lat. 25° 10’ N. Long: 98° 50’ E. Alt. 6—7,000 ft. April 1921. N. W. Yunnan.

19342. Cyrtolepis Buchanani, Roem. et Sch. Semi-scandent trailing shrub of 10-15 ft. Flowers green, fragrant. On scrub and in thickets in the Salwin valley. Lat. 25° 10'N, Long. 99° E. Alt. 7,000 ft. April 1921. N.W. Yunnan.

19343. Caragana Chamlagu, Lam. Spinous shrub of 2-4 ft. Dry situations in thickets on the hills east of Tengyueh. Flowers pale yellow, shaded purplish- brown. Lat. 25° N. Long. 98° 30’ E. Alt. 6-7,000 ft. April 1921. N.W. Yunnan.

19346.

19347-

19348.

19349.

19350.

19351.

19353-

19354-

PLANTAE CHINENSES FORRESTIANAE

. Acer amplum, Rehd. var. ‘Tree of 4o-50 ft. Foliage

only. Mixed forests on the Shweli-Salwin divide. Lat. 25° 10’ N. Long. 98° 45’ E. Alt. 10-11,000 ft. April 1921. N.W. Yunnan.

Primula sino-mollis, Balf. f. Plant of 6-10 inches. Flowers pale rose-purple, eye yellow. Shady thickets in the Salwin vallev. Lat. 25° ro’ N. Long. 98° 50’ E. Alt. 7,000 ft. April-rg2t N.W. Yunnan.

Elaeocarpus Braceanus, Watt. ‘Tree of 40-50 ft. Flowers dull creamy-yellow. In mixed forests in ravines on the eastern flank of the N’Maikha-Salwin divide. Lat. 25° so’ N. Long, 98° 25’ E. Alt. 89,000 ft. March 1921. N.W. Yunnan.

Phyllanthus. Emblica, Linn. Shrub of 2-3 feet. Flowers pale yellow. Dry, rocky, grassy slopes in the Salwin valley. Lat. 25° ro’ N. Long. 98° 50’ E. Alt. 6-7,000 ft. April r921. N.W. Yunnan.

Indigofera stachyodes, Lindl. var. Shrub of 4-6 ft. Flowers greyish-rose. Open stony slopes amongst scrub in the Salwin valley. Lat: 25° 10’ N. een 98° 50’ E. Alt. 6-7,000 ft. April 1921. N Yunnan.

Viola serpens, Wall. Semi-prostrate plant of 3-9 ims. Flowers white, stained purple at base. Shady moist pasture amongst willow scrub on the Tengyueh- Shweli, divide. Lat. 25° 5’ N. Long. 98° 30’ E. Alt. 8,o00 ft. April 1921. N.W. Yunnan.

Viburnum congestum, Rehder. Shrub of 4-8 feet. Flowers fragrant, waxy-white flushed pale rose-purple on exterior. Open scrub on marshy uplands on the Shweli-Salwin divide. Lat. 25° 10’ N. Long. 98° 45’ E. Alt. 10,000 ft. April 1921. N.W. Yxonan.

Dichroa febrifuga, Lour. Shrub of 4-7 feet. Flowers immature, green (!). Margins of shady pine and mixed forests on the Shweli-Salwin divide. Lat. 25° 10' N. Long. 98° 45’ E. Alt: 89,000 ft. April 1921. N.W. Yunnan. .

Olea dioica, Roxb. Spinous-leaved shrub: of 15-20 ft. Flowers green. Amongst scrub and in thickets on ee" hills east of Sha-yang. Lat. 259-20’ N. Long. 99° 40’ E. Alt. 7-8,000 ft.. April 1921. N.W. Yunnan.

19355.

19356.

19357-

19358.

19359.

19360.

19361.

19362.

NUMERICAL CATALOGUE 79

Jasminum dispermum, Wall. Scandent shrub of 8-14 ft. Flowers white, deeply flushed purple-rose on ex- terior, fragrant. On scrub and in thickets in side valleys on the hills east of Sha-Yang. Lat. 25° 20’ N. Long. 99° 40’ E. Alt. 7-8,000 ft. April 1921. N:W. Yunnan.

Euonymus : cornutoides, Loesener. Shrub of 12-20 ft. Flowers dull purple. In thickets and mixed forests on the Shweli-Salwin divide. Lat. 25° 10’ N. Long. 98° 45"E. ‘Alt. 9,000 ft. Aprikag2t. N.W. Yun- MASE .

Buddleia heliophila, W. W.Sm. Shrub of 3-5 ft. Flowers pure pale soft rose-pink, drying purplish, fragrant. Dry scrubby banks on the margins of pine forests on the hills around Beta-pu. Lat. 25° 30’ N. ‘Long. 99° 48’ E. Alt. 8-9,000 ft. April 1921. N. W. Yunnan.

Ligustrum Coryanum, W. W. Sm. Shrub of 3-6 ft. Flowers creamy-white, fragrant. Margins of dense serub and-on-banks in the Hsia-kuan_ ho valley, west- ern 'flank.of the Tali range. Lat. 25° 25’ N. Long. 100°. 10! 1B. ‘Alt’: 6;000 ft... April;r921. N.W. Yunnan.

Machilus robusta, W. W. Sm. Shrub of 20-25 rt: Flowers deep soft olive-brown. In shady thickets by streams on the Salwin-Pu-piao divide. Lat. 25° 10" N. Long. 99° E. Alt. 6-7,000 ft. April 1921. N.W. Yunnan.

Roxb. Shrub or ‘tree of 20-40 feet.

In openings and on the

Yung-

N

Olea _ dioica, Flowers green, fragrant. margins of mixed forests on the range between Chang and the Mekong valley. Lat. 25° 12 Long. 99° 20’ E. Alt. 910,000 ft. April 1921. N-W. Yunnan.

Cryptomeria japonica, D. Don. Tree of 50-70 ft. In mixed forest on the Salwin-Pu-piao divide. Liat. 25° 10! N. Long. 99° E. Alt. 8-9,000 ft. April 1921. N.W. Yunnan.

Hymenodictyon flaccidum, Wall. Shrub: of 20-40 ft. Foliage only, deep purple. In thickets in the Salwin valley. Lat. 25° ro’ N. Long. 98°50° EB. Alt. 6,000 ft. April 1921, N.W. Yunnan.

/

19363.

19364.

19365.

19366.

19367.

19368.

19369.

19370.

PLANTAR CHINENSES FORRESTIANAE

Bryum. Matted plant of 1 inch. Moist clayey slopes in shady mixed forest on the Shweli-Salwin divide. Lat. 25° 10’ N. Long. 98° 45’ E. Alt. 10,000 ft. April 1921. N.W. Yunnan.

Quercus. Shrub or tree of 30-40 ft. In mixed forest on the mountains around Sha-yang. Lat. 25° 20’ N. Long. 99° 40’ E. Alt. 7-8,000 ft. April 1921. N. W. Yunnan.

Myrica esculenta, Ham. Shrub of 6-9 ft. In imma- ture fruit. Fruits edible. Native name—Yeh-meh. Dry hillsides amongst scrub on the hills around Chu- tone. at. 96° 24’ Ni Ee gg tay US. AR. 7-8,o00 ft. April 1921. N.W. Yunnan.

Glochidion. Matted shrub of 12-15 ft. Branches arched. Flowers olive-green (!)._ Amongst scrub and in thickets on the hills east of Sha-yang. Lat. 25° 20’ N. Long. 99° go’ E. Alt. 7-8,000 ft. April 1921. N.W. Yunnan.

Randia sp. aff. sinensis, Roem. et Sch. Spinous, semi- scandent shrub of 9-15 ft. Flowers white, fragrant. In thickets and scrub in side valleys on the hills east of Sha-yang. Lat. 25° 20’ N. Long. 99° 40’ E. Alt. 7-8,o00 ft. April 1921. N.W. Yunnan.

Vaccinium scopulorum, W.W Sm. Shrub of 2-4 ft. Flowers greenish-white. Shady banks in dry pine forest on the hills east of Sha-Yang. Lat. 25° 20! N. ter 99° 40’ E. Alt. g-10,000 ft. April 192!1.

N.W. Yunnan.

Prunus majestica, Koehne. Shrub of 20-40 ft. Flowers (?) probably pink. In fruit. Fruits black- crimson, bitter. Open situations in ravines on the hills east of Sha-yang. Lat. 25° 20 N. Long. 99° 40' E. Alt. 7-8,000 ft. April 1921. N.W. Yunnan.

Prunus majestica, Koehne. Shrub or tree of 30-45 ft. In fruit, fruits black-crimson, bitter. Flowers pre- cocious, large, deep pink in colour and produced in great abundance. Flower Feb. and March—or even earlier according to season. Tali valley. Lat. 25° 40’ N. Long. 1009 12’ E. ° Alt. 6,700-7,000 ft. April 192r. N.W. Yunnan.

[The Tali Cherry. ] |

(9371;

19372.

19373.

19374.

19375.

19376.

19377.

19378.

19379.

NUMERICAL CATALOGUE 81

Millettia sp. aff. pachycarpa, Benth. Shrub of 4-6 ft. Flowers bright rose-pink, standard darkest with dark crimson markings. Open dry hillsides and in ravines on the hills west of Yung-chang. Lat. 25° 10’ N. Long. 99° 12’ E. Alt. 7-8,o00 ft. April 1921. N. W. Yunnan.

Docynia indica, Decne. Shrub of 20-25 ft. Flowers creamy-white. Amongst scrub and in thickets on the hills east of Sha-yang. Lat. 25° 20’ N. Long. 99? 40’ E. Alt. 9,000 ft. April 1921. N.W. Yunnan.

Begonia Acetosella, Craib. Plant of 3-5 ft. Flowers rich pink. Shady moist situations by streams in mixed forests in the Salwin valley. Lat. 25° 10’ N. Long. 98° 50’ E. Alt. 6,000 ft. April 1921. N.W. Yunnan.

Aucuba chinensis, Benth. Shrub of t10-r5_ feet. Flowers creamy-white. Shady situations by streams in the Salwin valley. Lat. 25° 10’ N. Long. 98° 50’ E. Alt. 5-6,o00 ft. April 1921. N.W. Yunnan.

Viburnum cordifolium, Wall. Shrub of 6-9 feet. Flowers creamy-white. Open thickets in side valleys on the Tali Shan. Lat. 25° 40’ N. Long. 100° 12’ E. Alt. 10,000 ft. May 1921. N.W. Yunnan.

Salix cathayana, Diels. Shrub of 2-4 feet. Open meadows on the Tali Shan. Lat. 25° 40’ N. Long. 100° 12’ E. Alt. 9,000 ft. May 1921. N.W. Yun- nan,

Pinus sinensis, Lamb. var. yunnanensis, Shaw. ‘Tree of 30-70 ft. Open pine forests on the eastern flank of the Tali Shan. Lat. 25° 40’ N. Long. 100° 12’ E. Alt. 9-10,000 ft. May 1921. N.W. Yunnan.

Abies Delavayi, Franch. Tree from 40-90 ft. Form- ing open forests on the eastern flank of the Tali Shan. Lat. 25° 40’ N. Long. 100° 12’ E. Alt. 9-11,000 ft. May 1921. N.W. Yunnan.

Sabia yunnanensis, Franch. Shrub of 6-9 ft. Flowers green. Amongst scrub on the margins of meadows on the eastern flank of the Tali Shan. Lat. 25° 40’ N. Long. 100° 12' EB. Alt. 10-11,000 ft. May 1921. N.W. Yunnan.

19382.

19356.

19387.

19388.

19389.

19390.

19391.

19393-

PLANTAE CHINENSES FORRESTIANAE

. Myrica. Shrub of 4-6 ft. Flowers olive-yellow. In

immature fruit. Open scrub in dry side valleys on the eastern flank of the Tali Shan. Lat. 25° 40’ N. Long. 100° 12’ E. Alt..9,000 ft. May 1921. N.W. Yunnan.

Trillium Tschonoskii, Maxim. Plant of 10—18 inches. Flowers white. Shady situations by streams in side yes on the eastern flank of the Tali Shan. Lat. 25° ao’ N. Long. 100° 12’ E. Alt. g—10,000 ft. May 1921. N.W. Yunnan.

Tenge yunnanensis, Mast. Shrub or tree of 30-60 ft. In mixed forest in side valleys on the eastern flank of the Tali Shan. Lat. 25° 40’ N. Long. 100° 12! E. Alt. to-11,000 ft. May 1921. N.W. Yunnan.

Salix cathayana, Diels. Shrub of 6-9 ft. Open thickets on the eastern flank of the Tali Shan. Lat. 25° yo! N. Long. 100° 12’ E. Alt. 1o+rf,ooo0 ft. May 1921. N.W. Yunnan.

Abies Faberi, Craib. Tree of 50-70 ft. In Abies forest on the eastern flank of the Tali Shan. Lat. 25° 40’ N. Long. 100° 12’ E. Alt. 10-11,000 ft. May 1921. N.W. Yunnan.

-Bulleyia yunnanensis, Schltr. Epiphytic plant of

18-20 inches. In fruit. On ledges of cliffs in side valleys on the eastern flank of the Tali Shan. Lat. 25° 4o’ N. Long. 1009 12’ E. Alt. 7-8,000 ft. May 1921. N.W. Yunnan.

Primula sinolisteri, Balf. f. Plant of -4-8 inches. Flowers pale rose or white. Ledges of cliffs in shady side apts on the eastern flank of the Tali Shan. Lat. 25° 40’ N. Long. 100° 12’ E. Alt. 9—10,000 ft. May 1921. N.W. Yunnan.

Isatis tinctoria, Linn. Plant of 2-3 ft. In fruit. In and on the margins of wheat fields in the Lang-kong valley. Tat. 269 12’ N. Long. 100° 10’ Alt. 7,000 ft. May 1921. 'N.W. Yunnan.

Vitis Thunbergii, Sieb. et Zucc. var. adstricta (Hance), Gagnep. Scandent shrub of 6-12 ft. Flowers olive- green. Amongst scrub on the margins of thickets on the ascent of the Sung-kwei pass from the Lang-kong valley. Lat. 26° 12! N. Long. 100° ro’ E, Alt. 8,000 ft. May 1921. “NW, Yunnan,

19393+

19394.

19395.

19396.

19397.

19398.

19399.

IQ4or,

NUMERICAL CATALOGUE 83

Berberis levis, Franch. Spinous shrub of 2-3. ft. Flowers yellow. Open, dry situations amongst scrub on the western flank of the Sung-kwei, pass. Lat. 26° 12’ N. ‘Long. 100° 12’. E, Alt..g,o00 ft. May 1921. N.W. Yunnan.

Leptodermis glauca (Franch.), Diels. Dwarf matted shrub of 1 ft. Foliage strongly malodorous, flowers white, faintly flushed rose on exterior. Open, ‘dry situations amongst scrub on the western flank of the Sung-kwei pass. Lat. 26° 12/"N. Long. 100° 12’ EB. Alt. 9-10,000 ft. May 1921. N.W. Yunnan.

Artemisia annua, Linn. Plant of 2-2% ft. Foliage only, strongly aromatic. Margins of begs streams in the Teng-chuan valley. Lat. 26° 4’ N. Long. 100° 6!'E. Alt. 7,000 ft. May 1921. N.W. Yunnan.

Rosa. Spinous shrub of 2-4 ft. Very free-flowering. Flowers bright rose. Open situations by streams in the Teng-chuan valley. Lat. 26° 4’ N. Long. 100° 6’ E. Alt. 7,000 ft. May 1921. N.W. Yunnan.

Pyrus. Shrub of 20-30 ft. In fruit. Open thickets on the western flank of the Sung-kwei pass. Lat. 26° 12/N. \Long..100° 12’ E. Alt. 8,000 ft. May 1921. N.W. Yunnan.

[Said to be the wild stock on which ‘most of the cultivated pears of W. and N.W. Yunnan are grafted or budded. |

Clerodendron fragrans, Wall. Shrubby plant of 3-6 ft. Flowers white, fragrant. Shady situations on the margins of thickets on the western flank of the Sung- kwei pass. Lat. 26°12’ N. Long. yoo® 12/ EK. Alt. 8,000 ft. May 1921. N.W. Yunnan.

Rosa. Spinous shrub of 3-5 ft. Very floriferous. Flowers deep rose. Open situations by streams 1m the Teng-chuan valley. Lat. 26° 4’ 'N. Long. 100° 6’ E. Alt.:7,000 ft. “May 1921. ‘N-W. Yunnan.

Primula sinolisteri, Balf. f. Plant of 3-8 inches, Flowers pale lavender-rose with green eye, fragrant. Shady, moist situations in thickets on the western flank of the Sung-kwei pass. Lat. 26°12! N. Long. 160° 19! E. Alt. 9,000 ft. “May 2921.)..N.W. Yun- nan,

84

19402.

19403.

19405.

19406.

19409.

19414.

19416.

19417.

PLANTAE CHINENSES FORRESTIANAE

Viburnum cordifolium, Wall. Shapely shrub of 6-10 it. and as many feet in diameter. Flowers pure white with grey anthers. In open pine forests and on the margins thereof on the western flank of the Sung-kwei pass. Lat. 26912’N. Long. 1009 12’ KE. Alt. 8,500 ft. May 1921... N.W. Yunnan.

Rosa. Spinous shrub of 9-18 ft. Branches arched. Flowers white, strongly fragrant. Margins of thickets in dry situations on the western flank of the Sung- kwei pass. Lat. 26° 12! N. Long. 1009 12’ KE. Alt. 8,500 ft. May 1921. N.W. Yunnan.

Rosa. Spinous shrub of 2-4 ft. Flowers bright rose. Open situations in the Langkong valley. Lat. 26° 12’ N. ‘Long. 100° 10’'E. Alt. 7,000 ft. . May 1921. N.W. Yunnan.

Scrophularia Delavayi, Franch. Pasture plant of 10~ 14 inches. Flowers bright ‘yellow. Open alpine meadows on the summit of the Tali Shan. Lat. 25° 4o'N. Long. 100° 12’ E. Alt. 14,000 ft. May 1921. N.W. Yunnan.

Prunus serrula, Franch. var. Shrub or tree of 25—50 ft. Flowers white. In thickets and mixed forests on the Tali-Shan. Lat. 25° 4o’ N. Long, 100° 12’ E. Alt. 10-11,000 ft. May 1921. N.W. Yunnan.

Skimmia melanocarpa, Rehd. et Wils. Shrub of 8-16 ft. Flowers greenish-white, anthers yellowish. In thickets in ravines on the eastern flank of the Tali- Shan’? Lat: a5? 4go!'N: ’ “Bérigs 160742" EB. ATE: 8,000 ft. May 1921. N.W. Yunnan.

Lonicera adenophora, Franch. var. Shrub of 6-9 ft. Flowers deep dull crimson. Amongst scrub on rocky slopes in ravines on the eastern flank of the Tali Shan. Lat. 25° 40’ N.. Long. 100° 12/ EK. Alt. 9,000 ft. May 1921. N.W. Yunnan.

pruinosa, Franch. Dwarf spinous shrub of

1-2 ft. Flowers yellow. Ledges of cliffs and rocky

slopes in ravines on the eastern flank of the Tali Shan.

- Lat. 25° go’ N. Long. 100° 12/ E. Alt. 9,000 ft. May 1921. N.W,. Yunnan,

19419.

19420.

19421.

19422.

19423.

19424.

19425.

19426,

19427.

NUMERICAL CATALOGUE 85

Acanthopanax gracilistylus, W. W. Sm. Spinous shrub of 6-10 ft. Branches arched. In immature fruit. In thickets on the ascent eastern flank of the Li-ti-ping, Yangtze-Mekong divide. Lat. 27° 12! N. Long. 99° 4o’ E. Alt. 9-10,000 ft. June 1921. N. W. Yunnan.

Quercus fenestrata, Roxb. Shrub or tree of 20-50 ft. In mixed oak forest in the Yangtze valley. Lat. 27° 6’ N. Long. 99° 54’ E. Alt. 6,000 ft. June 1g2r. N.W. Yunnan.

Corylus chinensis, Franch. Tree of 30~50 ft. Flowers (?). Fruits (?). Bark scaling in long strips. In mixed thickets and forests on the western flank of the Li-ti-ping, Mekong-Yangtze divide. Lat, 27° 12’ N, Long. 99° 35’ E. Alt. g-10,000 ft. June 1921. N. W. Yunnan.

Quercus variabilis, Blume. ‘Tree of 40-60 ft. Jn mixed oak forest in the Yangtze valley. Lat. 27° 6’ N. Long. 99° 54’ E. Alt. 6,000 ft. June 1921. N.W. Yunnan.

Rosa. Spinous shrub with arched branches of 6-12 ft. Flowers white, fragrant. Margins of thickets and by streams on the western flank of the Li-ti-ping, Mekong-Yangtze divide. Lat. 279 12’ N. Long. 99° 35’ E. Alt. 9,000 ft. June 1921. N.W. Yunnan.

Quercus variabilis, Blume. Tree of 60 ft. In mixed oak forest in the Yangtze valley. Lat. 27° 6’ N. Long. 99° 54’ E. Alt. 6,000 ft. June 1921. N.W. Yunnan.

Castanea vulgaris, Linn. Widely-spreading symmet- rical tree of 50-60 ft. Flowers creamy-yellow. Mixed and oak forest in the Yangtze valley. Lat. 27° 12’ N. Long. 99° 48’ E. Alt. 6-7,000 ft. June 1921. N. W. Yunnan.

Quercus dentata, Thunb. Widely-branched handsome tree of 40-60 ft. In mixed oak forest in the Yangtze valley. Lat. 279 6’ N. Long. 99° 54’ E. Alt. 6,000 ft. June 1921. N.W. Yunnan.

Quercus variabilis, Blume. Tree of 40-50 ft. In mixed oak forest in the Yangtze valley. Lat. 27° 12’ N. Long. 99° 48’ E. Alt. 6-7,000 ft. June 1921. N,W. Yunnan,

19429.

19430.

19431.

19432.

19433-

19434.

19435.

PLANTAE CHINENSES FORRESTIANAE

. Quercus ‘fenestrata, Roxb. Tree of yo-s5o ft. In

mixed oak forest'in the Yangtze valley. Lat. 27° 12 N.. . Long.:.g9° .48’:E. . Alt..6-7,000 ft. ‘June 1921. N.W. Yunnan.

‘Symplocos crataegoides, Ham. Shrub of 12-20 ft.

Flowers bright yellow. In thickets on the eastern flank of the Lji-ti-ping, aceite we divide. «Lat 27° \12' N. Long. g9°;4o’ E. Alt. 8,000 ft. June 1921. N.W. Yunnan.

Androsace dissecta, Franch. Semi-procumbent plant of 10-16 inches. Flowers white. Shady moist situa- tions in open pine and mixed forest on ‘the eastern flank of the Li-ti-ping, Yangtze-Mekong divide. Lat. 27° 12' N.. Long. g9°.40’.E.. Alt. 10,000 ft. June 1921. N.W. Yunnan.

Prunus sp. aff. Dielsiana, Schneid. Shrub of 20-40 ft. In immature fruit. Stipules and. bracts very gland- ular, glands marginal, the whole plant malodorous when fresh, with a strong acrid smell. In thickets by streams on the western flank of the Lji-ti-ping, Mekong-Yangtze divide. Lat. 27° 12/ N. Long. 99°

35’ E. Alt. g-10,000 ft. June 1921. N.W. Yunnan.

Prunus conadenia, Koehne. Shrub of 20-35 ft. In immature-fruit. Flowers (?). In open thickets by streams on the western flank of the Li-ti-ping, Mekong-Yangtze divide. Lat. 279 12’ N. Long. 99° -35’-E. ‘Alt. 9-r10,000 ft. “June 1921. ‘N!W. Yunnan.

Viburnum atrocyaneum, Clarke. Shrub of 4-7 ft. Flowers green. Amongst scrub in a ravine on the La-shi-pa, Yangtze divide. ‘Lat: 269 '54/°N. Long. 100° '6/‘E. “Alt. 9-10,000 ft. May 1921. N.W. Yunnan.

Lonicera Henryi, Hemsl. var. Semi-scandent shrub of 8-ro ft. Flowers deep golden-yellow, non-fragrant. Amongst scrub in a ravine on the La-shi-pa- Yangtze Mivade. »:cLat..26? 54’ AN. +Leng. roo?’ E. A 9-10,000 ft. May 1921. N.W. Yunnan.

Rosa. Weakly spinous shrub of 6-7 ft. Branches arched. -Flowers deep rose, white at. base. ~On cliffs $y: aystsniee eg, the divide between La-shi-pa and the Yangtze valley. Lat. 26° 54’ N. yong. 100° 6’ E. Alt. 9-10,000 ft. May 1921. N.W. Yunnan,

19436.

19438.

19439.

19441.

19442.

19448.

19449.

19451.

NuMERICAL CATALOGUE 87

Ulmus: Tree of 30-50 ft. Flowers .(?). Fruits. (?). Main branches and»stems covered with a corky fluted epidermis. Open situations in ravines onthe La-shi- pa-Yangtze. divide. Lat. 26°54’ N: Long. 100° 6’ E. Alt. 9,000 ft. May 1921. N.W. Yunnan.

Androsace: axillaris; Franch. ‘Trailing plant of 1-2 ft. Flowers white. Shady, grassy situations amongst rock and scrub on the: banks of the Mekong river. Lat; 27° 24’ N. Long. 99° 6’ E.. Alt. 6jo00 ft. June 1921. N.W. Yunnan:

praises pinnata, Benth. Surculdse plant of\ 1-2 nches: Flowers white. Shady banks and_ rocks ae grass-on the banks of the Mekong river. Lat: 27° 24’N. Long. 99° 6''E. Alt! 6,000 ft. June 1921. N.W. Yunnan.

Pterocarya Forrestii, W.W. Sm. Tree of 30-40 ft. Flowers green. In immature fruit: Open forest: on the eastern slope of the Li-ti-ping, Mekong-Yangtze divide. Lat. 27° 12’ N: Long. 99°40’ E. Alt. 8,000 ft. June 1921. N.W. Yunnan.

Actinidia callosa, Lindl. Scandent shrub of 10-20 ft. Flowers crimson with a broad. yellow margin. scrub in thickets on the eastern slope of the Li-ti-ping, Mekong-Yangtze divide. Lat. 279 12’ N. Long. 99° 40’ BE. Alt. 8,000 ft. Jane 1921. NsW. Yun- nan.

Hypericum patulum, Thunb. var. Henryi, Hort. Kew. Shrub. of ‘2-4 ft. Flowers orange-yellow. By o streams on the Lichiang plain. Lat. 26° 54’ N. Long. 100° 12' E. Alt. 8,200 ft. May 1921. N.W. Yunnan.

Corydalis cheirifolia, Franch. Plant of 12-18 inches. Flowers bright blue, tipped white. Boggy pasture‘on: the margins of pine forest on the Li-ti-ping, mares Vangtze divide. Lat. 279 12’ N. Ling. 99° 38’ E. Alt. 11,500 ft: June 1921. N.W. Yunnan.

Anemone Davidi, Franch. Plant of 6-12 inches. Flowers white. Open pasture on the margins of pine forest on the Li-ti-ping, SS divide. Lat. 279.12’ N. Long» 99° 38’ sis Alt) 11,500 ft. June 1921. N.W. shame

19453-

19454.

19457.

19459.

19461.

19463.

19464.

19465.

PLANTAE CHINENSES FORRESTIANAE

Rubus. Spinous shrub of 3-6 ft. Flowers (?). Form- ing dense thickets on the margins of pine forests on the Li-ti-ping, Mekong-Yangtze divide. Lat. 27° 12’ N. Long. 99° 38’ E. Alt. 11,500 ft. June 1921. N.W. Yunnan.

Stranvaesia Davidiana, Decne. Shrub of 6-9 feet. Flowers grey-white. Open situations amongst scrub on the western flank of the Lji-ti-ping, Mekong- Yangtze divide. Lat. 27° 12’ N. Long. 99° 35) E. Alt. 9,000 ft. June 1921. N.W. Yunnan.

Pedicularis. Plant of 6-12 inches. Flowers bright rose with beak crimson. Forming dense carpets beneath pines in the open, never spreading beyond the shade on the Li-ti-ping, Mekong-Yangtze divide. Lat. 27° 12’N. Long. 99° 38’ E. Alt. 12,000 ft. June 1921. N.W. Yunnan.

Ilex yunnanensis, Franch. Shrub of 10-12 ft. Flowers grey-white, fruits scarlet. In thickets by streams on the western flank of the Li-ti-ping, Mekonig- Yangtze . divide... yat:/ 279° 12’ N... Long. 98°..35/°E.. Alt. 8-9,000 ft. June 1921. N.W. Yunnan.

Prunus vestita, Wall. Shrub or tree of 20-45 ft. Flowers small, creamy-white. In open mixed forests and thickets on the western slope of the Li-ti-ping, Mekong-Yangtze divide. Lat. 27° 12’ N. Long. 99° 35’ E. Alt. 89,000 ft. June 1921. N.W. Yunnan.

Enkianthus. Symmetrical shrub of 18~20 ft. In fruit. In thickets by streams on the western flank of the Li-ti-ping, Mekong-Yangtze divide. Lat. 27° 12’ N. Long. 99° 35’ E. Alt. 8-9,000 ft. June 1921. N W. Yunnan.

Stranvaesia Davidiana, Decne. Shrub of 20-30 ft. Flowers white-creamy. In open thickets on the west ern flank of the fet poe. Mekong-Yangtze divide. Lat..27° 12’ N. Long. oon 35’ E. Alt. g-10,000 ft. June 1921. N.W. Yun

Primula Poissonii, Franch., forma. Boggy, peatv meadow on the Ciena, i Mekong: Vangtee divide. Lat. 27° 12’ N. Long. 99° 38’ E. Alt. 12,000 ft. June 1921. N.W. Yunnan.

19466.

19469.

19470.

19471.

19472.

19473.

19474.

19475.

NUMERICAL CATALOGUE 8g

Physalis peruviana, Linn. Plant of 1-2 ft. Flowers green-yellow. Stony ground by streams in the Me- kong valley. Lat. 27° 20 N. Long. 99° 6’ E. Alt. 6,000 ft. June 1921. N.W. Yunnan.

Salvia Bulleyana, Diels, var. Plant of 6-12 inches. Flowers pale orange with lip deep rosy purple. Moist meadows on the Li-ti-ping, Mekong-Yangtze divide. Lat2.27° 12’ N.. Long. 99°..38).E. .;Alts.12,000 ft. June 1921. N.W. Yunnan.

Paederia tomentosa, Blume. Scandent shrub of 6-9 ft. Flowers soft purple with grey-purple ihroat. Most malodorous (!).. Open situations on scrub in the Mekong valley. Lat. 27° 25’ N.. Long. gg° 6’ E. Alt. 6,000 ft. June 1921. N.W. Yunnan.

Cinnamomum Delavayi, Lec. var. Aromatic shrub of 20-25 ft. Flowers olive-brown, fruits red. In open thickets on the western flank of the Li-ti-ping, Mekong-Yangtze divide. Lat. 279 12’ N. Long. 99°

35’ E. Alt. 9,000 ft. June 1921. N.W. Yunnan.

Picea brachytyla, Pritzel, forma rhombisquamea, Stapf. Shapely tree of 40-70 ft. Forming forests on the summit of the Lji-ti-ping, Mekong-Yangtze divide. Lat. 27° 12’ N. Long. 99° 38’ E. Alt. 11,500-12,000 ft. June 1921. N.W. Yunnan.

Oreorchis foliosa, Lindl. Plant of 4-12 inches. Flowers golden olive-brown, fragrant. Open pasture on the Li-ti-ping, Mekong-Yangtze divide. Lat. 27° 12'N. Long. 99° 38’ E. Alt. 11,500 ft. June 1921.

Berberis Wallichiana, DC. vel aff. Spinous shrub of 4-5 ft. Fruits glaucous purple. Open scrubby streams on the eastern flank of the Li-ti-ping, Mekong-Yangtze divide. Lat. 27° 12’ N. Long. 99° 4o' E. Alt. 10,000 ft. June 1921. N.W. Yunnan.

Cephalotaxus sinensis, Rehd. et Wils. Shrub of 20- 4o ft. Open thickets by streams on the eastern flank of the Li-ti-ping, Mekong-Yangtze divide. Lat. > a 12/N. Long. 99° 40’ E. Alt. 8-9,000 ft. June 1921. N.W. Yunnan,

go

19476.

19478.

19482.

19483.

19454.

19485.

19486.

19487.

PLANTAE CHINENSES FORRESTIANAE

Cotoneaster turbinata, Craib. Shrub of 10-16 ft. Flowers creamy-white. Open situations on the mar- gins of thickets in the Mekong valley. Lat. 27° 25/ N. Long. 99° 6’ E. Alt. 6-7,000 ft. June 1921. N.W. Yunnan.

Prunus crataegifolius, Hand.-Mzt. Shrub of 6-9 feet. Flowers rose-pink. On cliffs and bouldery slopes by streams a scrub, on the neta -Salwin divide. Lat. 27° 54’ N. Long. 989 50’ E. Alt. g-10,000 ft. June 1921. N.W. Yunnan.

Lindera kariensis, W.W.Sm. Shrub of 20-25 ft. In immature fruit. Flowers (?) probably green. In mixed thickets by streams on the Mekong-Salwin divide. Lat. 27° 54’ N. Long. 98° 50’ E. Alt. 8-9,000 ft. June 1921. N.W. Yunnan.

Actinidia purpurea, Rehder. Scandent shrub of 12-16 ft. Flowers fleshy, pure white. On scrub and rock in ravines on the Mekong-Salwin divide. Lat. 27° 54/ N. Long. 98° 50’ E. Alt. 8-9,000 ft. June 1921. N.W. Yunnan.

Pyrus. Shrub of 16-20 ft. Flowers rose. Shady situations amongst scrub in ravines on the Mekong- Salwin divide. Lat. 27° 54’ N. Long. 98° 50’ E. Alt. g-10,000 ft. June 1921. N.W. Yunnan.

Scindapsus. Plant of several feet. On rocks and trees in shady ravines on the Mekong-Salwin divide, Salwin valley. Lat. 27° 54’ N. Long. 98° 50’ E. Alt. 6-7,000 ft. June 1921. N.W. Yunnan.

Veratrilla. Bailloni, Franch. Plant of 12-24 inches. Flowers dark olive-green. Lush mountain meadows on the Mekong-Salwin divide. Lat. 27° 54’ N. Long. 98° so! E.. Alt. 12-13,000 ft. June 1921. N.W. Yunnan.

Cremanthodium Forrestii, Diels. Plant. of 6-12 inches. Flowers golden-yellow. Stony, grassy slopes on the Mekong-Salwin.divide. Lat. 27° 54’ N. Long. 98°. 50’. E.. Alt. 13,000 ft: June 1921.. N.W. Yun- nan.

19488.

19489.

19490.

19493.

19494.

19495-

19498.

19499.

NuMERICcAL CATALOGUE gI

Saxifraga pallida, Wall. var. typica, Engl. et Irmsck. Plant of 2-6 inches. Flowers creamy-white. Ledges of cliffs and moss-covered boulders on the Mekong- Salwin divide. Lat. 279 54’ N. Long. 98° 50’ E. Alt. 13,000 ft. June 1921. N.W. Yunnan.

Lillium Davidi, Duch. Plant of 2-3 ft. Flowers deep orange-red, spotted deep purple. Stony, grassy slopes in and on the margins of thickets in side valleys on the Mekong-Salwin divide. Lat. 27° 54’ N. Long. 98° 50' E. Alt. 8,000 ft. June 1921. N.W. Yunnan.

Diplarche multiflora, Hook. f. et Thoms. Shrub of 6-12 inches. Flowers white. Open, stony slopes and meadows and on rocks and cliffs on the Mekong-Salwin divide. Lat. 27° 54’ N. Long. 98° 50’ HE. Alt. 13,000 ft. June 1921. N.W. Yunnan.

Diapensia brevistyla, W. EK. Evans, MSS. Mat plant of many inches diameter, 6-12 or more. Flowers white. Moist, stony meadows on the Mekong-Salwin divide. ;) Lat. 27° sq’ N.~ Long..98° 50’ B. Alt. 14,000 ft. June 1921. N.W. Yunnan.

Berneuxia thibetica, Decne. Ledges of cliffs and humus-covered boulders in shady situations on the Mekong-Salwin divide. Lat. 27° 54’ N. Long. 98° 50’ E. Alt. 10,000 ft. June 1921. N.W. Yunnan.

Cassiope selaginoides, Hook. f. et Thoms. Shrubby plant of 2~5 inches. Flowers white flushed rosy-pink towards base. On stony slopes, boulders and cliffs on the Mekong-Salwin divide. Lat. 27° 54’ N. Long. 98° 50’ E. Alt. 13-14,000 ft. June 1921. N.W. Yunnan.

Euonymus tingens, Wall. Shrub of 20-25 ft. Flowers creamy-white, lined and marbled purple. Open thick- ets in side valleys on the Mekong-Salwin divide. Lat. 279 s4'N. Long. 98° 50’ E. Alt. 9-10,000 ft. June 1921. N.W. Yunnan.

Enkianthus. Shrub of 6-12 ft. Flowers pale yellowish- rose, lined a deeper shade. In and on the margins of open thickets by streams on the Mekong-Salwin divide. Lat. 27° 54’ N. Long. 98° 50’ E. Alt. 9-10,000 ft. June 1921. N.W. Yunnan.

92 PLANTAE CHINENSES FORRESTIANAE

19501. Vaccinium oreotrephes, W. W. Sm. Duplicate of 1917. Shrub of 6-18 inches. Flowers rose or white flushed rose. Open, moist, peaty, stony pasture or on rocks on the Mekong-Salwin divide. Lat. 27° 54’ N. Long. 98? 50’ E. Alt. 13-14,000 ft. June 1921. N.W. Yunnan.

19504. Actinidia callosa, Lindl. Scandent shrub of 6-10 ft. Flowers pale crimson, margined creamy-white. On scrub in side valleys on the Mekong-Salwin divide. Lat 28° N. Long. 98° 50’ E. Alt. 8-9,000 ft. June 1921. N.W. Yunnan.

19505. Betula utilis, D. Don, var. Tree of 30-50 ft. Bark red (!). In mixed forests in side valleys on the Mekong-Salwin divide. Lat. 28° N. Long. 98° 50’ EK. Alt. 9-10,000 ft. June 1921. N.W. Yunnan.

19506. Pyrus Vilmorini (Schneider). Shrub of 20 ft. In immature fruit. Mixed thickets, rocky situations in ravines on the Mekong-Salwin divide. Lat. 28° N. Long. 98° 50’ E. Alt. 910,000 ft.. June 1921. N. W. Yunnan.

19507. Machilus. Shrub of 15-30 ft. Foliage and bark faintly aromatic. Flowers olive-brown. ‘Thickets in ravines on the Mekong-Salwin divide. Lat. 28° N. Long. 98° 50’ E. Alt. 9,000 ft. June 1921. N.W. Yunnan.

19508. Senecio acerifolius, C. Winkl. Surculose plant of 2-2% ft. Flowers orange-yellow. Moist pasture by streams in side valleys on the Mekong-Salwin divide. Lat. 28° N. Long. 98° 50! KE. Alt. 10,000 ft. June 1921. N.W. Yunnan,

19509. Lonicera_ trichosantha, Bur. et Franch. Erect branched shrub of 9-18 ft. Flowers pale golden- yellow, non-fragrant. Thickets by streams in side valleys on the Mekong-Salwin divide. Lat. 28° N. Long. 98° 50’ E. Alt. 9,000 ft. June 1921. N.W. Yunnan.

19510. Pyrus Vilmorini (Schneider). Shrub of 10-20 ft. Flowers (?) probably crimson. Open, shady situa- tions on bouldery slopes on the Mekong-Salwin divide. Lat. 28° N. Long. 98° 50’ E. Alt. 10,000 ft. June 1921. N.W. Yunnan.

IQ5I1.

19513.

19514.

19516.

19518.

19519.

19520.

19521.

NUMERICAL CATALOGUE 93

Prunus brachypoda, Batal. ‘Tree of 40-50 ft. Flowers creamy-white. In mixed thickets on the Mekong- Salwin divide. Lat. 27° 54’ N. Long. 98° so’ E. Alt. 8-9,000 ft. June 1921.. N.W. Yunnan.

Paraquilegia microphylla, Drumm. et Hutch. Cushion plant of 10-16 inches diameter. Flowers pale bluish- lavender, anthers and nectaries orange. Firmly rooted in crevices of shady dry cliffs in ravines on the Mekong-Salwin divide. Lat. 27° 54’ N. Long. 98° 50’ E. Alt. 11,000 ft. June 1921. N.W. Yunnan.

Primula bella, Franch. Plant of 1-2 inches. Flowers rose-purple—drying blue-purple—with throat white. Moist, peaty, stony, alpine meadows and on boulders on the Mekong-Salwin divide. Lat. 279 54’ N. Long. g8° so’ E. Alt. 13-14,000 ft. June 1921. N.W. Yunnan.

Primula silaensis, Petitm. Massed plant of 2-4 inches. Flowers deep blue-purple, fragrant. Moist, stony, peaty meadows on the Londre pass, Mekong-Salwin divide. Lat. 28° 12’ N. Long. 98° qo’ E. Alt. 14-14,500 ft. June 1921. N.W. Yunnan.

Vaccinium modestum, W. W. Sm. Dwarf shrub of 4-14 inches. Flowers dull rose, tinged green. Open peaty, moist, stony pasture and on rocks on the Londre pass, Mekoug, Salen divide. Lat. 28° 12’ N. Long. 98° 40’ E. Alt. 13,000 ft. June 1g21. N.W. Vantaa

Cassiope palpebrata, W. W. Sm. Mossy plant of I-1% inches. Flowers creamy-white. Forming mats on rocks and moist, stony meadows on the Londre pass, Mekong-Salwin divide. Lat. 28° 12’ N. Long. 98° 40’ E. Alt. 13-14,000 ft. June 1921. N.W. Yunnan.

Salix Lindleyana, Wall. Prostrate shrub of a few inches. On moist, stony pasture on the Londre pass, Mekong-Salwin divide. Lat. 28° 12’ N. Long. 98° 40’ E. Alt. 14,000 ft. June 1921. N.W. Yunnan.

Berberis sp. aff. diaphana, Maxim. Spinous shrub of 3-5 ft. Flowers pale orange-yellow. Scrub on the earn, of — forests on the Mekong-Salwin divide. Lat. 279 54’N. Long. 98° 50’ E. Alt. 10-11,000 ft, June 1921. N.W. Yunnan.

19522.

19523.

19524.

19525-

19527.

19528.

T9539.

19532.

PLANTAE CHINENSES FORRESTIANAE

Primula lepta, Balf. f. et Forrest. Plant of 3-6 inches. Flowers deep blue-purple, fragrant. Open, moist

Lat. 28° 12’ N. Long. 98° 4o’ E. Alt. 13-14,000 ft. June 1921. N.W. Yunnan.

Primula Wardii, Balf. f. Plant of 6-10 inches. Flowers lavender-rose, with grey eye. Open, moist pasture by streams on the Londre pass, Mekong- Salwin divide. Lat. 28° 12! N. Long. 98° 4o’ E. Alt. 13,000 ft. June 1921. N.W. Yunnan.

Stellaria decumbens, Edgew. Mat plant many inches in diameter. Flowers white. On moist, peaty ground on the Londre pass, Mekong-Salwin divide. Tat. 289 12’ N. Long. 98° 40’ E. _ Alt. 14,000 ft. June 1921. N.W. Yunnan.

Primula amethystina, Franch. Plant of 4-9 inches. Flowers deep purple-blue. Moist, peaty, stony mead- ows on the Londre pass, Mekong-Salwin divide. Lat. 28° 12' N. Long. 98° 4o’ E. Alt. 14,000 ft. June 1921. N.W. Yunnan.

Nomocharis euxantha, W. W. Sm. et W. E. Evans. Plant of 3-5 inches. Flowers golden-yellow, very faintly spotted and tinged crimson on base interior. Stony meadow on the margin of rhododendron scrub on the Londre pass, Mekong-Salwin divide. Lat. 28° 12’N. Long. 98° go’ E. Alt. 14,000 ft. June 1921. N.W. Yunnan.

Nephrodium Filix-mas, Rich. var. Specimens un- developed (!). Amongst rocks in ravines on the Mekong-Salwin divide. Lat. 279 sa’! N. Long. 98° 50° E. Alt. 10,000 ft. June 1921. N.W. Yunnan.

Primula stragulata, Balf. f£. et Forrest. Clustered plant of r inch. Flowers deep rosy-purple, eye white. Moist, peaty moorland on the Bei-ma Shan. Lat. 28° 18’ N. Long. 99° 10’ E. Alt, I4-15,000 ft. June 1921. N.W. Yunnan.

Geranium Robertianum, Linn. Plant of 12-14 inches. Flowers pale rose. Dry <