American journal of botany

TES ae sv sant os

yet

at t € H Sea Fh ae ee = Ml oS Ny .* ~~

fa pees. +
at bere We Mayer Ags Mey
cee es ser ag UE hhh BEG | s#
yy Toes sta Ninn aad PRPS. “3S nee “= : | ra www sap

et PRES Oe r CaS Ay -~ = get ft ae an,

\ ; Pa a yf: N ' ene 54 i | a’ I. ~ ie &

Paige "4 Padapitate itr tl hae, A tah aaa cay
vy i} ‘a, \ ayy! ¥\-

elle me ima fe Chl Apecreoew a Am a Aa ~ z & ba x \e ba hve as =

~ a ws S64 | Z bt-jileok ach ; 2 a: NFR =~ oe “Gk =e ba = 3 beat 47>

™ misy {hd ye ov? Tee or & f nig ak aie “yy” © Auny ie iit ite ye sy SNe
~ B, aol val al % co 5 “ : > <a vty * We OEE A, wy JN, Pas et 2 use pee
“Aig Err SR toe csi pennies ef 1 “Aa Mit dates AT aoa iil unbbaoinan
| Bayle. . a ie pal vy eee ky iggeers toon ran | Ss iy Ba \ Mt
Wa A Reene es amay iNT HLEPT Hal Waa" iw a { PS ey i]
®

Ee . wma F Zaper rk Ye Benes:

SAP, BD torr We Me SVN Nady, rNaarhant “Ay .
bingo HSyrerssg, 9 07 et gi A ervathin
Se get a verte. ome, i] AAD ww bil “Deak | |

py OAR gg eT cee VS aia TE EGG 2 | slp Ey - AY « A> RELLNL YN
wee dai Wal ALCON he é aie a ap bs E aan! i} Nha wer Eb
gi Lt Pitt t | LLL b;? Bi ar esateaehh “AP od 14#UNe@ AA or VN, Ene Aun ~~ wih
“yy sagt | it. Aa’ r+
‘ey “ty wi he a agent Hd paw an mIMAGNS ~™

a ine ananae

a = eS 3

Me 88S, an ty Weed gar VPNs

yee a ey ee mea ad Oe ee ee tier ial Int

SONATE ay ‘egey y
OAC a sale teh Naa ape teah?

- "i ke be
eb EEE “AMSG ply, iq"
oa - t id ay

AGA teat’ Ay i Ttag. Wega: HTT * Bh eens ane, ANBAR

eS srag ra, 4 WNyrgs T Bl we

wR ANN A , f.. (Tt iw wT THATCH RS Oe Poy y\
pA: AN am figment “A: pny \e “Wiadh/ VV takZ ied Cele HU “AN gpeaps * Jf ttrrent MI #89 Wed
abs nr rh uy, ys oGee Qhe~ vom ren ve
site we gfrnmregest “Wy minnnta | TAS aT WS, Hays 2 iv eae of™ Ve neetriteel:
11 {| AACA oy 7 Big Ce Ae SCAN @ SNARES: © Orr ot See e a
cieegeee MAL | | Latah) | | | “1. / | iS iS es on ics ve ieee ee ts weenie ation he A gente me
iw WO, ; | { : x og
lp x Sf ar ft» MO aa ‘A Vv Tecra Pole MG a ANAL A A» 4 fy, id heder
f ySseumndae a iz Hate, {| 7 teaeal c. 2 iabetl SA) a SOA i ae nine
Neem er hag nage ary, ‘ SB eB Aa NERS AAS i
vee ~@ pa Vi) i: ot M2 “ er 4 ame F ™ ow ‘y ee a od? tt v’ Ani 7 S

SEA Jot 07

Bee le maed~ ine? yw" 1]
a ae “ eae vps Lr age? r) sah assay fT re sidan aby

; 2 apy,
"he To pe | AR ae . 144 yee dary
Ang - ‘rr. peers ‘ wi - ,, Ueow Cvyy- Ay mY - ‘
tian | Heyl ye Thy
Mas POR Sly, : re .
DURAN andes UUM ana rttcees rity
\ Ney) LY \W ry ha — wet, _y Ae aw* ane FAAo. Ta J>ORy. 7 au BA wy
dviuy fn gs Mas iN] Vi baecere Ett Td dy rages - yr” _w! 7 at
Hatt tt te | | ¥ ia aH “Ct -o~ iil
wi Ls a | oo { -' |eé
me TTT TTT | vate male alee weetnil Maen yy Loa sb Mt etRA GS nan Sah
Ty WB. ae a Sak Ltn gy Se a renee rom gM O AR Ca MG
soem PNT 8 NN Ag wn, Wg Md ao nt} ven WANG Oy tee
de ti wtih Wee Pe Lat ets eN Ns oh EPP Acton etka. Adey wy = *Wyiia) sey 4n
| nants Byala ids DAL Pee ees ein oe Lette tt | ¥,
mgt aes me

"Vij. Ls ty
Ty Vy “ x 5 : j a . y q 4 iy. }
| Va N | geso": Wes wet Ae et Brin r h wt ' é
oa a Nias ‘ Wayeatic, } rence

o%:
Wegycsaee® gurcatveeedtee dong danneae”

, RAT | od vwe
bye, eras, wuytiit ha WUISY "it Lan ae ri nee oN 1 & 1 pam, cht Ay ull
; ss De a

St aeepes, og 2 eA a ART TAY
. hoi _

4c oeee 6M TEM Qe *

& § tis a wks

wre ak,

ts iS UENS 2 EAN can oempage Vita. sina hil
* ripped. pasting ht’ oe. wt SAA od | eeeel Wis Prades || VUE ny Buy
, 7 TS OT nh lapger far aaah iy aK An: Mmm a "ty
pth | _s6URR “Te. \y We Pd Oe Lala Ne opapee” “prea rlesovied
ami ey mene ‘iatee.t Wee Wi Maat II weeeRitine,- no tae 4rd VL T |
poprapre ° aN fy "ins ene ey rer ta aucdf Hine an af

470°%O°% a WA, Ayu! : Nite Por, \ wT
rn ~s | Jt) | 3 r ihe ve ES TaN ites Nvbaiannnanaedien ee LN, 1 %\
Ahk SPN ie rcs sacha Werba” rWtnaaensal ON Aiea Fain een Nan.

ey Pr w ii Po en Oi os

KUT sity Me Manne ae Pi wana. VOW, fissansmracinvevonses
y Vege” 4)

| | 1 Oe She owe wees shiny
hit TT Pil iT re ~e ar. sah sedaheanshabal
J - Sa adhe y

1 rgrnart mTth |

Aah, ei Neowse,

a eebe:: WAP Nanay, ~ @ ;

ww

on

4 Aan, we Sl 4. .
htt ett Ni r

} = \ A
" pany eViey ee oars ariieye: “Mine
wo at

| nny a lye
te meer Wow lay wy Ledbi ade oe PERN CS, Mase,
Ad Li) Mine meron pe
Pia Nea WwW

rimmed yay Foe ee AMEN ALAN G
My } Terenas iindaiaaa al Nir \)
,. Paka ay se a tats il TIT He Tee eeaiqe Man
ae ee On Cn aE Ih Mtge 2 evened Te ¥ a
ii etugeees wm, “Anas

aguaratvorntie reer t Tal Lah | WiLT eat
} Wy tye, Meascersey oy cigs steel gw IAT Hay
mig Wyaltw: NAS | =

| ry

{

|

|
i
wf

oe meg ed

SEL

vim

—e

———

fe AMERICAN
. JOURNAL OF BOTANY

OFFICIAL PUBLICATION OF THE

BOTANICAL SOCIETY OF AMERICA

EDITORIAL COMMITTEE

F. C. NEwcomseE, Editor-in-Chief
University of Michigan

C. STUART GAGER, Business Manager R. A. HARPER,
Brooklyn Botanic Garden Columbia University
IRVING W. BAILEY A. S. HitcHcock
ae Bussey Institution Bureau of Plant Indusiry
{ H. H. BARTLETT EpGAR W. OLIVE
University of Michigan Brooklyn Botanic Garden

A. G. JOHNSON, University of Wisconsin
(Representing American Phytopathological Society)

VOLUME TV—1917

WITH THIRTY-ONE PLATES AND NINETY-FOUR TEXT FIGURES

we ;
RU RL Ore
@

S Ny
;
s

24588

National Museu

PUBLISHED
IN COOPERATION WITH THE BOTANICAL SOCIETY OF AMERICA
BY THE

BROOKLYN BOTANIC GARDEN

Ar 41 Norru Queen Street, Lancaster, Pa.

PRESS OF
THE NEW ERA PRINTING COMPANY
LANCASTER, PA,

@SStHES

PABLESOE CONTENTS VOLUME IV, 1917

Nout, JANUARY

PAGE

The influence of certain climatic factors on the development of

Endothia parasitica (Murr.) And. (with three text figures)
NEIL E. STEVENS
Gunnera petaloidea Gaud., a remarkable plant of the Hawaiian
{| SVS te 0 OA a eR a ND VAUGHAN MACCAUGHEY
An interesting modification in Xanthium (with three text figures)
CHARLES A. SHULL

Endophyllum-like rusts of Porto Rico (with plates I-III)

E. W. OLIVE AND H. H. WHETZEL

No. 2, FEBRUARY

Fifteen- and sixteen-chromosome Oenothera mutants (with nine

RM MMIMOMIGCS eg eee. Gey ahs 4 APR ap tee a ANNE M. Lutz
The influence of temperature on the growth of Endothia parasitica
Gvibisone text figure)... 60 cess ec. NEIL E. STEVENS

No. 3, MaArcuH

Matroclinic inheritance in mutation crosses of Oenothera Reyn-
oldsii (with four text figures)
Cart’ D. La RuE AnD HH. H> BARTLETT
Duration of leaves in evergreens (with thirteen text figures)

VINNIE A. PEASE

The relation between evaporation and plant succession in a given

aikea (vith nine-text figures) . ../...5 2... FRANK C. GATES
No. 4, APRIL

The relation of some rusts to the physiology of their hosts (with
[Dice IN MrAII GI Vai ech, amen mee ee Tit E. B. MarIns
The development of some species of Agarics (with plates VI-XI)

112

119

145

161

179

A. W. BLIZZARD 221

il

lv TABLE OF CONTENTS

The origin and development of the galls produced by two cedar
rust fungi (with one text figure and plates XII-XVIJ)
J. L. WEIMER

No. 5, May

The perennial scapose Drabas of North America
EDWIN BLAKE PAYSON

The osmotic concentration of the tissue fluids of Jamaican mon-

tane rain-forest vegetation
J. ARTHUR HARRIS AND JOHN V. LAWRENCE
The viability of radish seeds (Raphanus sativus L.) as affected
by high temperatures and water content (with one text figure)
H. G. WAGGONER

No. 6, JUNE

@he taxonomy of. the Agaricaceaes 7 WILLIAM A. MURRILL

Observations on forest tree rusts (with two text figures)
JAMES R. WEIR AND ERNEST HUBERT

Endothia pigments |. (with six text figures)

Lon A. HAWKINS AND NEIL E. STEVENS
Observations on an Achlya lacking sexual reproduction (with
plates OVA) ess, 5 ey a anata eave cat Wm. H. WESTON
The rusts occurring on the genus Fritillaria (with three text

HeUTES hohe Be ca he ee CHARLES C. REES 2

INO 7) 4) Wiley

Fertility in Cichorium intybus: the sporadic occurrence of self-
fertile plants among the progeny of self-sterile plants (with
CWO: text shores!) i; 2 emcees ale Oe eee A. B.: Sroux

Inheritance of endosperm color in maize... .. ORLAND E. WHITE

The influence of light and chlorophyll formation on the minimum
toxic concentration of magnesium nitrate for the squash
(with two text figures)..R. B. Harvey AND R. H. TRUE

The use of the vibration galvanometer with a 60-cycle alternating
current in the measurement of the conductivity of electrolytes
(withsone. text -fioure)y).. 2) ee NEWTON B. GREEN

Immunochemical studies of the plant proteins: proteins of the
wheat seed and other cereals. Study IX..R. P. WODEHOUSE

The toxicity of galactose and mannose for green plants and the
antagonistic action of other sugars toward these (with four
[ESRI OUC etsy esa age aa Stat A mR RE LEWIs KNUDSON

241

253
268

299

oe
396

407

AII

A17

430

TABLE OF CONTENTS

No. 8, OCTOBER

Taxonomic characters of the genera Alternaria and Macrosporium
(with eleven text figures and plates XIX and XX)
Joun A. ELLIoTT

Crown-rot of fruit trees: histological studies (with plates XX [-—
RIK WWD SRE aE SP oN RA ca eo i J. G. GROSSENBACHER

No. 9, NOVEMBER

Effect of soil temperature on the growth of bean plants and upon

their susceptibility to a root parasite..... DONALD REDDICK
The development of Cortinarius pholideus (with plates XXVIII
EIN CIO ret Ne et te Syl ce yee W. H. SAWYER, JR.

Leaf-structure as related to environment (with six text figures
g
HERBERT C. HANSON

No. 10, DECEMBER

The phytogeography of Manoa Valley, Hawaiian Islands (with
foumeeen text figures)......5..... VAUGHAN MACCAUGHEY

Revision of the Hawaiian species of the genus Cyrtandra, section
Cylindrocalyces Hillebr. (with five text figures)

439

477

513

520

00

501

JosErpH F. Rock 604

On the distribution of abnormalities in the inflorescence of Spiraea
Vanhouttei (with four text figures and plates XXX and
DOOGI ee Ge Be re ea J. ArTHUR HARRIS

ikniemamoun Ole TVer eee Pe Pecan eis a

624

(Mates of publication: No. 1,,Feb. 3; No. 2, Feb, 17; No: 3,.Mar.
2a, No. 4, May 8; No. 5, May 31; No. 6, June 29, No. 7, July 14;

Nie. 8, Oct. 2; No. 9, Nov. 24; No. 10, Dec. 12.)

PRRATA,|VOUUNEES TY:

Page 12, (‘) at end of paragraph 2 should be transferred to end of
third line above, after the word freezing.”

Page 13, insert f and i respectively as the first letters,of the third
and second lines from the bottom.

Page 371, for text Fig. 2, read*Fig. 3.

Page 371, in explanation of figure, for infrequens, read Fritillariae.

Page 372, tor text Pig. 3, read Hig a2:

Page 372, in explanation of figure, for Fritillariae, read Miurae.

Fi

2

yo

Lee , Stan Gace, Busitess Manager ’ 2

“m epreaeitina American Photopatoloi feu cy)

is cs to all ry Sr except Mex éx i 20

‘
Y

is should be oe :

Akane at e08t rates for soni e-< ‘anid tab glar ‘fr natter n excess

'
Wada

rmatter 5 ce nti ee

os ee hhored oes ae is retu

gp PART yee

3 cover ane en Poe be =

y

ceo must be added conllct

ss

fave been. lost in the

AMERICAN
JOURNAL OF BOTANY

——__—_ a.

VoL. IV JANUARY, I917 No. 1

THE INFLUENCE OF CERTAIN CLIMATIC FACTORS ON
THE DEVELOPMENT OF ENDOTHIA PARASITICA
(MURR.) AND.

NEIL E. STEVENS

The chestnut blight is at present common from the northern limit
of the chestnut, that is, southern New Hampshire and Vermont, to
central Virginia. The area which it occupies includes the northern
limits of growth of two native species of Endothia, Endothia gyrosa
(Schw.) Fr. and Endothia radicals (Schw.) De Not. It is also a
transition region for several important plant diseases. In the southern
portion of this territory bitter rot is one of the commonest and most
destructive diseases of apples; in the nothern portion it is a botanical
curiosity; and pear-blight, which is so abundant in the more southerly
portions of this area, is hardly known from the northern states of New
England. Apple scab, on the other hand, is more important in the
northern portion than in the southern.

In order to gain more complete knowledge of the behavior of
Endothia parasitica through this range and if possible to throw some
light on the factors which limit the growth of these other fungi, the
writer has undertaken a quantitative comparison of the growth and
fructification of the fungus with the weather conditions, as far as data
are available. While the work is not yet complete, enough data have
accumulated to warrant the publication of results. This seems
especially desirable in view of the fact that two of the stations, Wil-
mington, Delaware, and Hartford, Connecticut, must now be aban-
doned because of the general infection of the chestnut. 7

PLAN OF EXPERIMENTS

Previous observations on the growth and reproduction of Endothia
parasitica have been confined chiefly to single localities, with little
[The Journal for December (3: 527-593) was issued January 6, 1917.]

. I

2 NEIL E. STEVENS

opportunity for comparison. Consequently in this work special care
was taken to have inoculations made in the same way and on trees
as nearly similar as possible but in different localities. It was desired
of course to make observations at stations climatologically as different
as possible. The actual location of inoculations was however governed
by practical considerations. In order to avoid spreading the chestnut
blight beyond its present range it was necessary to confine work to
regions where the disease was so well established as to leave no hope
of eradication. A quantity of healthy chestnut was obviously neces-
sary. The latter consideration excluded the entire region between
Philadelphia and New Haven, Connecticut. With the exception of
one locality (Overlook Mountain) the inoculations were all made near
regular U. S. Weather Bureau observation stations.1

The stations selected were Concord, N. H.; Williamstown and
Amherst, Mass.; Hartford, Conn.; Wilmington, Del.; Van Bibber,
Woodstock, and Frederick, Md.; Washington, D. C.; and Fairfax and
Charlottesville, Va. The distance from Concord to Charlottesville is
about 500 miles, or about 5 degrees of latitude. In addition to regular
inoculations at these stations inoculations were made at various
elevations on Overlook Mountain in the Catskills in order to determine
whether difference in altitude would make any perceptible difference
in the growth and fructification of Endothia parasitica. Overlook
Mountain was selected as being the only place known to the writer
where chestnut grows through a considerable range of elevation and
where the chestnut blight is present. Graylock Mountain near
Williamstown, Mass., was first selected but chestnut was not found on
this mountain above 1,500 feet.

Work was begun in the spring of 1914 and each station visited once
in five or six weeks during the summer of 1914 and twice during the
_ summer of 1915. At each visit ten or more inoculations were made on

‘ healthy chestnut trees and the condition of previous inoculations
noted. The trees inoculated were uniformly second growth and as
far as possible were from 6 to 8 inches in diameter. Wherever these
conditions were not met the fact is indicated in the report of obser-
vations. The inoculations were made by cutting through the bark
with a sharp knife and inserting a quantity of mycelium and spores
from a pure culture, usually on corn meal, with a freshly cut twig.

1JTn the selection of these stations, as well as in the interpretation of weather
data, the writer had the advice of Mr. L. M. Tarr, local forecaster, U. S. Weather
Bureau, New Haven, Conn.

INFLUENCE OF CERTAIN CLIMATIC FACTORS 3

PREVIOUS WoRK ON RATE OF GROWTH

Anderson (1, p. 16)! conducted experiments on the growth of
Endothia parasitica on Castanea dentata at Charter Oak, Pa., during
the summer of 1912, and Rankin (9) during the same summer at
*Napanoch, N. Y. Both these writers give the average growth for
each month during the summer and Anderson gives it for the entire
year. The average annual growth? at Charter Oak, Pa., for the year
ending June I, 1912, was 15.97 cm. according to Anderson (I, p. 575),
while Rankin estimates 12 cm. about the average amount of a season’s
growth at Napanoch.

Rogers and Gravatt (10) made an intensive study of the spread
of the chestnut blight over a small area near Bluemont, Va., and give
6.35 inches (15.87 cm.) as the average annual diameter growth of
cankers at this point. They found the average growth on Castanea
pumila near Leesburg, Va., for the year ending August 15, 1914, to be
6.8 inch (16.08 cm.). There is fairly close agreement among the
results from Virginia and Pennsylvania even though they were taken
in different years. The growth in New York is, however, consider-
ably less.

RATE OF LATERAL GROWTH

Since Endothia parasitica kills its host by girdling the parts attacked,
vertical growth is of no importance so far as its parasitic qualities are
concerned, consequently in this work the rate of lateral growth alone
was measured. As careful comparative measurements for various
periods of the same year have already been given by Anderson (1)
and Rankin (9), special attention was paid to determining the amount
of growth for one year at the various points. On this account no cuts.
were made in the cankers until they were one year old. All measure-
ments made previous to that time were taken from the sunken area
in the bark.

Table I gives the annual lateral growth (determined by cutting
away the bark) of cankers at the various stations for the years ending
in May and in August, 1915, so far as the data are complete. Each
figure represents an average of all the normal cankers; that is, cankers
which developed only on one side of the cut were not included. These
averages are expressed in the nearest centimeter, as that seems to the

1 Reference is made by number to “Literature cited,” p. 31.
* All measurements are for lateral growth.

4 NEIL E. STEVENS

writer to represent about the degree of accuracy with which a number
of cankers can be measured. These measurements are not exceptional
in any way and in all probability represent about the average growth
at those points during the year. In general, the growth for the year
ending in May is about the same as that for the year ending in August.
‘This is not true of inoculations made at all seasons however.

Experiments during two seasons (1912-13 and 1913-14) indicate
that inoculations of Endothia parasitica on Castanea made in the fall
do not develop until the following spring. Those made in Maryland
during November, 1912, showed no evidence of development until
early in the following May. A similar series made early in November,
1913, showed no growth until spring and cankers from inoculations
made in April, 1914, developed throughout the summer as rapidly as
those made the fall before. These results agree with those of Anderson
(1, p. 8) and Rankin (9, p. 244).

TABLE I

Lateral Growth of Cankers of Endothia parasitica in Various Localities

Elevation Year Year

Locality (in Feet) Ending | Cm. Ending Cm.
IQIs IQI5

Concord: (NewHiL Were 350 May 18 | 14 | Aug. 19 14

Williamstown, Mass....... 711 (900) 22 | 15 16 15

Amberst, Miasss< sc. eee 222 17, | 36 17 15

(2 stations).

Hartiord, Connz. 4.) ee 159 (350) 15) 16 18 16

Woodstock, Nv. Yerre eae or 1,000 2A eels II 16

Wilmington, Delo... 22a. 86 I4 | 19 10 20

Van Bibiber,“Mdioe..., ate 100 14: |.2024) Octien 7 18

Woodstock Mid... 728. ce 392 Apr, 27 | 20) GeAtiges.9 20

Frederick Midis. 77 she an 275 (325) 27a 28 9 |(Sprout gird-
led. No rec-
ords.)

Washineton, DSC... ... 2... II2 (300) 227) 220, “ejulys23 on

aiid Naseeige etnies: ace ake 300 jines 6. 23 4 2)

Charlottesville, Va.......... 854 Apr. 20 | 25 | (Forest fire; no later

records. )

As is shown by the table, there is a more or less regular increase in
the annual growth from Concord, N. H., to Charlottesville, Va. So
great is this difference that it must obviously be due to the difference
in climate and not to a variation in the trees. The record is unfor-
tunately not complete at Frederick, Md., or Charlottesville, Va. At
Frederick the trees inoculated in August proved too small and were

INFLUENCE OF CERTAIN CLIMATIC FACTORS 5

girdled before the year was complete. At Charlottesville a forest fire
destroyed the inoculated trees some time during the last week in

April, 1915.

SI w .
x > & N S \ K S
ly

s x : S N y 8 R S t : g
es SS RS SS NG 8G 8
de Sele Bek ete SG Aa oe >
Se KX; 86 Ny Sq S89 Rd Ks Ne SX VE Ss

RY Sq KE FE LT FQ Ro TE FT OR

Fic. 1. Graphs showing the growth of Endothia parasitica on Castanea dentata
and climatic data for the year ending April and May, 1915.

The relation of the amount of growth at the various stations is
best seen from the curves (Figs. I and 2), where the amount of growth
is expressed in percentage of that of Charlottesville. The amount at
Charlottesville has been used as standard for comparison of all data
in making curves, since this is the most southerly point and is near the
center of the chestnut belt. This will also make comparisons easy
in case points further south are studied as the chestnut blight advances.

6 NEIL E. STEVENS

For comparison with data just given the amount of annual growth
for the years ending in May and August at various elevations on
Overlook Mountain (Ulster Co., New York) is given (Table II).
While the writer has no accurate data as to the temperature at these
various elevations it is interesting and significant that in general

wy

. : : S

S S caveats Ni 9

p N) © GS y 9 Q K N

Fo og 8 8 8) 8 8k
SSS Ra OS Se
Cy VAS as qd § G'S. RS NY Sena
Se See RCE NVR RR a We Gk eS
Se Fy KO MS SS SN SR gS QS aN oae
BS BX Sq CE SB LTE §$Q VE PS CS ssa

QR OK LAS YT ae
Ppt AL
CCEERIZAL DAL
CEES NSS
|e oe

Fic. 2. Graphs showing the growth of Endothia parasitica on Castanea dentata
and climatic data for the year ending August, 1915.

the amount of growth decreases with increased altitude. The only
exceptions to this rule are in the case of Station O 4 which, being on
the south side of the main ridge of Overlook, showed more growth

INFLUENCE OF CERTAIN CLIMATIC FACTORS 7

than Stations O 2 and O 3 on the north side of the ridge. Station O6
is also an exception to the general rule since though at an elevation of
only 1,500 feet it showed but 10 cm. growth for a year. The writer
is quite unable to explain this condition beyond the possibility that
this reduced growth may be due to the fact that the trees inoculated
were in a rather deep and shady ravine.

TABLE II

Lateral Growth of Cankers of Endothia parasitica at the Various Stations on Overlook
Mountain, Woodstock, N. Y.

Station (in Fee Endine Cm. Ending Cm.
IQIs IQIs

(CF oo cu US ieee ae ne aoe ea 600°, | May 25°) 15. Aug. 11 I5
ETE ee a ee es 1,000 2On ters II 16
Si En cose ple gee ee een eer 1,500 2An eas: II 15
Sy oi bk Be a ea 1,500 222 oi. II
D Es 2 je pee anne nee nA ear 1,500 25h LO! Ngee 2, | 10
Re Ce ck. . coe yk eh al 1,900 27. |. TA 12 1g)
© 2 (north side of ridge)... . 0.5. .6.. 2,500 2On i Te se TT
O 4 (South side of ridge)............. 2,800 26/713 1 a Ma
@O ae(north-side of range).............. 2,900 26544 Gi oa 6)

CLIMATOLOGICAL DATA

In comparing the growth of this fungus with climatic conditions
the highest degree of accuracy could be obtained only by carrying ona
complete series of meteorological observations in each locality. This
procedure, which would have required an observer stationed at each
point, was impracticable. Consequently, it was decided to depend
entirely on the data regularly furnished by the U. S. Weather Bureau.
This, of course, necessitates neglecting certain factors known to be
important to plant life. The writer is of the opinion, however, that
if progress is soon to be made toward understanding the climatology
of plant disease a serious attempt must be made to utilize the meteoro-
logical data already available.

While the climatic data available from the Weather Bureau records
are not all that might be desired, all the stations except Van Bibber,
Md., and Fairfax, Va., furnish daily maximum and minimum tem-
peratures and amount of precipitation, as well as the number of clear,
partly cloudy, and cloudy days, and the prevailing direction of the
wind for each month. The date of last killing frost in spring and first

STEVENS

NEIL E.

“SUOI}E}S SNOLICA 94} 1e PeAIasqoO JsIy 91am Sarodsoose YOIYM UI YIUOUL dy} SazeOIpPUT Ysi1a4se uy

69°6¢ 9L'8v 6£°9P 99°CV zzov | Cg°6v LOCE gg'St
peas gc Ant 6 ‘Sny | O1 ‘Sny | 11 ‘sny | gi ‘sny | ZI ‘sny | 91 ‘sny | 61 ‘sny
gL°9£ QI°gt PERE 96'S¢ 1b-zv CLS of '6z of LE V9'6z VOrw2
ze ‘ady | ee ‘idy | 4ze -ady | Lz ‘ady | vi Aey | ve Aeyy | Si Avy | Li Avy | zz Avy | gr Avy
OCT ewe) ele hare 6.0 0.635 0 Gt ATG ese lege Odeo
eel, oo'Z 6r°6 LL°8 ze'9 v6°Z £g°9 ~82'8 eave "979
Ese eilcee 6S°¢ Ga e 6f-z Vz'g L6°9 £16 Lo°6 67°01
ze°s gS'9 bez, 0z'9 To'9 Cg°z IS'I oo'’ CL 6L'I
bYyz gle pega Qc n83P mde pesnc oz'I gv! 66°0
6v'0 ,,06°0 oZ'o Sri Sic’ lOve gS 66°¢ alae Z9°S%
c6°0 Lo'l Lo'l oh aae | HET Qz'O 6z°0 ra xe) Ivo del L
160°V 09'°f go's bev 1z'9 ory oft zo'L viv 6Q°2
06°S bo-9 Coz GLL fg°s 99 oL¢ zS'9 cre Lv?
LEG 6r'v gc'V CoV 06°S £6°¢ Cee 63°c 61 LOS
96°z 90°% IQ‘ 96°I CLI bo’ Qtz ORC Tics EGE
91'S CQ'l OL'I oS"! 61°z coe Coe 60°z CL ZI‘I
Li'l 99'0 Iv zO'l 980 Ze'O 0z'O zS'O £G'o 1z'O
Cl-z 00°9 96°S 089 Cie voz 96°! 11° o's 16°S
loz Cee 96°z Qc LIL Cle of'y eG’ veew 6b'e
Ql°z 0z'°9 LQ LE C6°z oz oL'I Ceo: Lgic 6£°z
Qv'r ay a Coz bS't 63°0 ov ELeg gS’ v6'I bo‘
eve ooze LoD CoS tov og't bee 6S°9 C9°S Lge
“BA ‘aT[IA “dq ‘u01 ‘PIN “PIN ‘Pa “ACN ‘O48 T “uu07) “SSCL “SsSey, “UMOF ‘H'N
-soqyjopreys | -Suryseay | ‘yOUeporq | YOoyspoo A |‘UOVsUIWTIAA| AUOYOTT ‘pl0ojysey qs1squUYy -SUIETIIIAA ‘paoau07d

"u0ynj1g19ad4 1040 I,
Ij] alavy

"+ SuIpus 1e9h 10jJ [COT

"+ *Sulpua Jedd JOj [e}OT,

on Fagen owe iepel evettel ie: sf © Azeniqaq

coo OND O O40 Olaet DLO tee Azenue{
SIOL

ose oe yaquiaceq

Pere ae yaquiaAaoyy

Creat ei et aCe arO ery ChEICHC 19q0390Q
Oe rr Jaquia3zdas

INFLUENCE OF CERTAIN CLIMATIC FACTORS

La

—l

COMO eH NOATO
=

eA IA
-sayjojreyy

Gy) ‘ad ~uo1

-SUIYSe MA

vz1

“PIN
“YOMapety

*PIN

v6
OI ‘sny

z6

‘Pda

‘yOo0Ispoo M |‘UOUITOTI AA

c8
II ‘sny

69

DMOMnOOANAA KRY mmAO OY

yuoyosy

HCIN oa

ea
VI
ZI
Ll

LI

*uu0d
‘piopiepy

VII

cOl

VI

“SSC
‘yyy

LI ‘sny | 91 “sny

Li Avy | zz Avy

OII gol
61 ‘sny
vol 16
gI Ae
CI CI
VI QI
II | OI
Q 6
6 8
9 O
6 8
II OI
Q OI
II 8
¢ 9
¢ v
II Auk
cI II
OI 8
6 9
VI VI
“SSBINT ‘UMO} “HH “"N

~SUICITIT AMA ‘pioouo7)

** SUIPUD ead JOj [eIOT,

** SUIpua IedA IOj [eIOT,

a eltettci atic tele ve ael-eme Areniqay
Fa. Creat, OF One) chet one ‘Arenue [{

jaquiasaq
Sen GO Oso a doo JaquiaAoN

oc ee eo eo eo ow **** 18qoIIQ
Gide Oe oc 0. Dep Jaquisa}dag

adopt 40 YIUT IO°O UoNDIdiIALg YIM Skog {0 AIQUNnyAT

Al

alav iL

Io NEIL E. STEVENS

killing frost in autumn is also available for most of the stations and the
regular Weather Bureau observation stations give the percentage of
possible sunshine each day as well as atmospheric pressure and direc-
tion and velocity of wind.

Among the climatic elements recorded, any direct relation between
atmospheric pressure and growth is very difficult to trace. Wind
velocity and light while undoubtedly important for a green plant
probably have little relation to the growth of Endothia parasitica,
especially since the advancing edge of the mycelium is under the
unbroken bark of the host tree. A careful study of the Weather
Bureau data shows no correlation between amount of growth and
either the prevailing direction of the wind or the number of clear
days during the period investigated. The writer’s laboratory experi-
ments also have failed to demonstrate any relation between the amount
of light and the growth and fructification of the fungus even when
growing on the surface of culture media.

PRECIPITATION

As regards precipitation, there are plainly two elements to be
considered: the amount of rainfall and its distribution. Table III
gives the monthly precipitation for each station during the course
of the investigations, Table IV the number of days with more than
.OI inch precipitation for each year during the same period. Careful
examination of rainfall data fails to show any relation between either
amount or frequency of rain and the amount of growth of the fungus.
Amherst, Mass., had practically the same rainfall as Charlottesville,
Va. Yet the growth at the latter point averaged nearly four inches
greater. Williamstown, Mass., on the other hand, had a much
smaller rainfall’ than Amherst but showed about the same amount of
growth. Amherst, Mass., and Hartford, Conn., had much greater
rainfall for the year ending in August than for that ending in May,
yet the amount of growth was practically the same.

While the different localities show considerable variation both in
the amount of rainfall and in the number of days with rain this seems
to have no relation to the amount of growth. This is probably best
shown by the curves (figs. I and 2) of rainfall and number of days with
precipitation. The various points of these curves are expressed in
percentage of the rainfall and number of days with rain at Charlottes-
ville, Va. The irregularity of the rainfall curves as compared with

INFLUENCE OF CERTAIN CLIMATIC FACTORS II

the curve of growth makes it seem almost impossible that either total
amount of rainfall or number of days with rain has any direct effect
on the growth of the fungus. This is theoretically very probable
since the growing edge of the fungus is in or near the cambium of the
host under the bark and its moisture supply must come from the host
itself.

It is conceivable that a fungus might be susceptible to changes in
the water content of those portions of its host in which it grows, so
slight as not to produce a perceptible effect on the host. There is,
however, no evidence that such is the case in Endothia parasitica. On
the other hand Rankin (9, p. 245) who investigated the relation of
the growth of Endothia parasitica to the water content of the bark of
Castanea dentata during the summer of I912 at Napanoch, N. Y..,
failed to demonstrate that the “variation in the physiology of the
tree which results from drought conditions alters to any great degree
either the susceptibility of the chestnut tree or the rate of progress
of the mycelium in the bark.”

With his conclusion the writer’s observations entirely agree. In
the course of three years’ inoculation experiments and field observa-
tion the writer has been unable to obtain any evidence that the rate
of growth of this fungus is affected by external dryness which does not
produce a perceptible withering effect on the host.

LENGTH OF FROSTLESS SEASON

The only remaining factor seems to be that of temperature. Zon
(13) has emphasized the necessity of considering the length of the
growing period in plant climatology and the advisability of tabulating
climatic data separately for the period of growth and the period of
rest. While his contention is undoubtedly correct for green plants
it is apparently not true in the case of Endothia parasitica which has,
strictly speaking, no resting season. | Field observations and laboratory
experiments both show that Endothia parasitica will grow whenever
the temperature rises above its minimum for growth, which, as Shear
and Stevens (11, p. 7) have determined, is about 8° or 9° C. This is
apparently true regardless of the previous temperature and whether
the host is dormant or not. |

Anderson and Rankin (3, p. 574) conducted experiments separately
at Charter Oak, Pa., and Napanoch, N. Y., and agree that the chief
growth of Endothia parasitica occurs between March and October but

12 NEIL E. STEVENS

that the fungus continues to grow during mild periods of winter.
During January, 1913, an average lateral growth of 0.51 cm. was
recorded for cankers at Charter Oak, Pa., while no growth whatever
was recorded in November, December, or February. In this con-
nection Anderson and Rankin call attention to the fact that during
January there were ten different days in which a temperature above
9° C. was recorded. As Rankin (9, p. 244) states, “‘cessation of growth
of the mycelium in the bark during fall and winter as well as negative
results of inoculations at this time of the year is explained purely on
the basis that the temperature is too low for the vegetative activity
of the fungus.”’

Experiments made by the writer with plate cultures of Endothia
parasitica in the laboratory agree with these field observations. When
such cultures were kept at temperatures below their minimum for
growth, that is, 7° C., 3° C., and 1° C., for twenty-four hours: and
then removed to room temperature for twenty-four hours, they grew
practically as much while in the warm room as did cultures which had
never been in the ice box. So quickly does the fungus recover from the
effect of the low temperatures that plate cultures which were kept in the
refrigerator for twenty-two hours and at room temperature for only two
hours each day showed a measurable growth at the end of a week.
Spring weather, with warm days and cool nights or even a warm period
in midwinter would then permit growth. In fact plate cultures kept out
doors at Washington, D. C., during January, 1915, madea total growth
of 1.5cm. for the month. Anderson and Rankin further state (3, p. 575)
that ‘‘the mycelium does not seem to be injured in the least by freezing,
but remains alive in all parts of the canker during the winter. These
investigators report that cultures kept frozen for a month at a time
renewed growth naturally on being brought back into the laboratory.”’

This being the case one would expect to find little connection
between the length of frostless season and the amount of growth in
the various localities. Table V gives the length of frost-free period
in days during the time of the experiment at the various stations.
There is of course in general a decrease in the length of frost-free
period from Charlottesville northward. This is, however, not regular,
since the length of frost-free period is greater at Hartford, Conn., than
at Van Bibber or Woodstock, Md., while the growth is of course
greater at the latter points. Williamstown, Mass., had a considerably
shorter frost-free period than Hartford, Conn., and on the other hand

INFLUENCE OF CERTAIN CLIMATIC FACTORS 13

a much longer frost-free period than Amherst, Mass., while the amount
of growth at these points is practically thesame. ‘The curves, figure 2,
in which the length of frost-free period at the various points is indi-
cated in percentage of the period at Charlottesville, show that while
there is in general a falling off in the length of frost-free period from
Charlottesville to Concord, the agreement between this curve and the
curve of growth is not such as to indicate any direct causal relation.

TABLE V
Frost Data for Various Localities
: First Killing Frost | Last Killing Frost | Length of Frost-free
Locality in Autumn, 1914 in Spring, 1915 Period in Days
Pomeord, NaH... see cee September 29 May I5 137
mepmerst, Wass... .0.0......8 8) 25 20 131
Williamstown, Mass............ 29 April 22 160
ilagtiord (Conn... 2. 2... 60 s,s: October 27 13 197
Mohonk ILake,-N. Y..........-. 14 14 183
Witmieton, Wels. ............. 28 4 207
Wanibibper, Md... 05.2006. u. September 29 15 167
nwoodseock,; Mids... 20. eee. 29 15 167
mreaerick, Medio... oe Le. | October 28 15 196
Mrasmineton; Di Coo... ek ee | 28 3 208
@harlottesville, Va.....000...... | 28 5 206

TEMPERATURE

In measuring the effectiveness of temperature in plant climatological
studies annual or monthly means are obviously of very little signif-
icance. As has been frequently pointed out, localities with similar
mean annual temperatures may have actually very different climatic
conditions. Among the methods of measuring temperature more
satisfactorily probably the most used is that of direct summation of
daily mean temperatures. Merriam (6) was the first to apply this
method in preparing a chart of the climatic zones of the United States.
Briefly, the method is as follows: A certain minimum temperature is
assumed as a starting point and the amount added to the summation
each day is the number of degrees above the assumed minimum which
represents the mean temperature for that day. The minimum is
sometimes the freezing point but often a somewhat higher temperature.

Recently the Livingstons (5, p. 353) have called attention to the
act that although these temperature summations have in many
nstances furnished data consistent among themselves and constituting
an apparently reliable criterion for the measurement of the intensity

I4 NEIL E. STEVENS

and duration aspects of the temperature factor it is improbable that
any fundamental or general principle regarding the influence of tem-
perature in a plant is derived from the relations thus brought out.
They suggest as more satisfactory for measuring temperature effect-
iveness a method of calculating temperature efficiencies based on the
well-known chemical principle of van’t Hoff and Arrhenius, that
within limits the velocity of most chemical reactions doubles or some-
what more than doubles for each rise in temperature of 10° C. On
this basis the Livingstons (5, p. 366) have prepared a table of approxi-
mate efficiency indices for temperatures in whole degrees from 40° F.
to 99° F., assuming the efficiency to be unity at 40° and to double with
each rise in temperature of 18°, and have prepared maps of the United
States comparing temperature summations with the temperature
efficiencies calculated according to their tables. The results of the
two methods show a rather close general agreement but there are
numerous discrepancies in detail.

For purposes of comparison both methods have been used in the
present work. In all calculations the mean for each day was deter-
mined by the formula: Mean = $(maximum + minimum). The
calculations have been made in the Fahrenheit scale, not because this
scale is as convenient as the Centigrade but because all Weather
Bureau data are so published.

DIRECT TEMPERATURE SUMMATIONS

The direct temperature summations. have been calculated for all
‘the stations where complete data are available.* Forty-five degrees
F. has been regarded as the zero point, since it is undoubtedly slightly
below the temperature at which Endothia parasitica is able to grow (11).
The amount added each day is then one half the sum of the maximum
plus the minimum temperature as given in the monthly reports of
climatological data issued by the Weather Bureau. The sum of these
amounts for the 365 days for which the growth of the canker was
measured is the temperature summation for the year. Table VI gives
these summations for the various localities and the percentage of each
when the summation at Charlottesville for the year ending April 23,
IQI5, is considered 100 percent.

With the single exception of Wilmington there is a fairly regular

4 The writer is indebted to Mr. Anthony Merryman for much assistance in
calculating weather data.

INFLUENCE OF CERTAIN CLIMATIC FACTORS r5

decrease in the temperature summations from Charlottesville north-
ward. A comparison of the curves of growth and temperature
summation (figs. I and 2) shows that there are some irregularities and
that the temperature summation falls somewhat more rapidly north-
ward than does the amount of growth.

TABLE VI
Temperature Summations

Locality Rete Summation | Percent Endne Summation| Percent
Charlottesville, Va....; Apr. 20 | 6,412 100.00
Washington, D.C.... 22 5,153 80.0 July 28 | 4,941 77 ok
Frederick, Md........ 27 5,005 78.1 ,
Woodstock, Md...... 27" 5.024. 73.3 Aug. 9 | 4,742 73.8
Wilmington, Del...... May 14 | 5,443 84.9 10 | 5,169 80.0
Mohonk Lake, N. Y.. Zhe. 3,622 56.5 II 3,465 54.0
Hartford, Conn... .. . 15 | 3,943 61.0 Eon 357.70 59.0
Amherst, Mass....... D7 | 3,504 55-9 D7 347.0 54.3
Williamstown, Mass. . 223, Ole7. 47.0 16 | 2,970 46.3
Concord, N..H......°-. 18 | 3,045 47.5 TO e2sOe4. 45.6

TEMPERATURE EFFICIENCIES

In calculating temperature efficiencies the Livingstons’ method
was adopted with no change except in the zero point. That is, it was
assumed that the efficiency doubled with each rise in temperature of
18° F., since this assumption seems to agree most nearly with the
work of the numerous investigators who have sought to determine the
application of the van’t Hoff-Arrhenius principle to physiology.®

There is, of course, no direct evidence that growth in the case of
this particular fungus is accelerated by rise in temperature at the
rate assumed. The calculations were made rather to determine how
closely the general law would apply to this organism under field con-
ditions.

Efficiency was assumed to be unity at 45° F. The writer is however
inclined to believe now that 47° might be even more accurate. This
makes the formula for calculating efficiency

Loe
Bony

when e = the efficiency and ¢ the daily mean temperature. A table
was prepared on this basis and used in calculating the temperature

é€é=2

° For a brief résumé of the literature on this point see 5, p. 356-359.

16 NEIL E. STEVENS

efficiency for each day. The table is obviously the same as that given
by the Livingstons (5, p. 366) except that it assumes 45° instead of
40° to equal unity. Whenever the mean daily temperature was
below 45° the efficiency was considered zero. The efficiency index of
each locality for a year is the sum of the daily indices.

Table VII gives the temperature efficiencies for the various localities
studied and the percentage of each based on the temperature efficiency
of Charlottesville as 100 percent. This table should be compared
first of course with the table of temperature summations. As the
figures of the efficiency index at Charlottesville approximately equal
the first two figures of the temperature summation at that point a
rough direct comparison is possible. In general, it is evident that the
temperature efficiency indices fall off less rapidly in amount from
Charlottesville northward than do the temperature summations.
This is shown more strikingly by the percentages and as is indicated
by the figures the curve of temperature efficiency follows the curve
of growth more closely for the northern localities than does the curve
of temperature summations. The former falls slightly less rapidly
than does the growth curve; the latter somewhat more rapidly. The
only serious exception is Wilmington which has higher temperature
summation and efficiency indices than the other Maryland stations
. or even Washington, D. C., without a corresponding increase in
amount of growth. This discrepancy the writer is wholly unable to
explain.

TABLE VII
Temperature Efficiencies

Locality YearEnding| Efficiencies| Percent |YearEnding; Efficiencies; Percent
Charlottesville, Va....| Apr. 20 635 100.0
Mairiax, Vas i tunes 2). ee ee eae ate. ant: seat:
Washington, D. C....| Apr. 22 594 94 July 28 574 rere)
Brederick, Md... .:... 27, 586 93 tees a bier
Woodstock, Md...... 277. 586 92 Aug. 29 562 89
Van Bibber, Md...... ae ie eo, ae ieee Ae
Wilmington, Del......) May 13 632 99 Aug. 10 605 95
Hartford; Gonn.. a1. : 15 481 76 18 463 73
Mohonk Lake, N. Y.. 24 435 68 II 421 66
Amherst, Mass....... 17 431 68 17 421 66
Williamstown, Mass. . 22 396 62 16 381 60
Goncord ON blew, 18 384 60 19 371 58

When the extent of the territory covered and the necessarily ap-
proximate nature of the data and their calculation are considered the

INFLUENCE OF CERTAIN CLIMATIC FACTORS Ly

degree of correlation between the curves of growth and of temperature
is remarkably close. In general the correlation is slightly less perfect
when the effect of temperature is expressed by efficiency indices than
by direct summation. In either case, there can be little doubt that
under climatic conditions in which the optimum temperature of the
fungus is rarely greatly exceeded (11, p. 9) the amount of growth made
by Endothia parasitica depends directly on the amount and duration
of heat available. If this conclusion is correct the chestnut blight
may be expected to spread somewhat faster in the future than it has
in the past unless other factors intervene to check its growth. For
instance, the temperature summation for Corinth, Miss. (year ending
June 1, 1915), where there is still some chestnut and where Endothia
fluens mississippiensis was first collected, is 6,561 or 102.0 percent of
the summation at Charlottesville. The efficiency at that point is
764 or 120.3 percent of that at Charlottesville. The chestnut blight
should then be able to make at Corinth a growth somewhat greater
than that at Charlottesville and considerably greater than that at
any of the northern points.

At first glance the statement that the amount of lateral growth of
Endothia parasitica is dependent directly on temperature may seem
so simple an explanation as to be artificial. A consideration of the
conditions under which the advancing edge of the mycelium lives in
the host shows, however, that the biological conditions are unusually
constant and that the fungus is very little influenced by many factors
of great importance to green plants.

The environmental factors most used in such a classification of
plants as that given by Képpen (4), for instance, are many of them
negligible. The chemical nature of the medium in which the fungus
grows parasitically must be fairly constant since it is always the same
portion of the same host species. Certainly the difference between
individual trees of this species is so slight that as yet no tree resistant
to this fungus has been found.

Light, so important in the growth of green plants, is negligible
here. The writer has thus far been unable to demonstrate that light
had any effect on the growth or reproduction of this fungus under
laboratory conditions and in all probability no light whatever reaches
the advancing edge of the mycelium under the bark.

The fungus has, moreover, no resting season. It is almost inde-
pendent of external moisture supply since it lives in the portion of the

18 NEIL E. STEVENS

host where moisture is most abundant, and where evaporation is very
slight, if indeed it occurs at all.

If the biological relations of the fungus are correctly understood.
it is, while growing as a parasite in or near the cambium of its host,
uninfluenced by any environmental condition except that of temper-
ature, at least in the territory it now occupies in this country. And
the influence of temperature itself is restricted to an increase or de-
crease of the amount of growth rather than any permanent cessation
of growth such as is brought about by heavy frost in the case of
green plants.

ASCOSPORE PRODUCTION

In studying the relation of climatic conditions to reproduction in
Endothia parasitica attention was concentrated on the production of
ascospores. The time necessary for the development of pycnidia is
so short that to determine the factors involved would necessitate an
intensive study of a few adjacent localities, with much more frequent
visits than were possible in covering so large an area as was involved in
the present study.

Previous observations on the production of ascospores have been
isolated rather than comparative. Murrill (8, p. 187), in his original
description of the fungus, stated: ‘“‘The winter spores [ascospores]
mature in late autumn .. . and germinate the following spring.”

Anderson and Babcock (2, p. 36) made several hundred inoculations
on various dates from May 29 to July 12, 1912, and recorded the date of
appearance of pycnospore horns and perithecia. They conclude that
(p. 37): “In general it may be said that under natural conditions in
the summer time the spore horns will be developed in from three to
six weeks, and that the winter or ascospore stage will develop in ten

weeks or more. The fact that the perithecial stage on all these plots
appeared in September and October should not be interpreted as
indicating that the approach of winter had any influence in bringing
about this stage.”

Rankin (9, p. 249) made inoculations at Napanoch, Ulster Co.,
N. Y., each month during the summer of 1912 commencing with May
and observed that stromata were not produced on any of the cankers
until about the second week of September (p. 254), and that they
appeared as quickly on cankers produced by inoculations of July 4
as on those made May 1. Cankers produced from inoculations made
at different times from May 1 to August 1 showed uniformly mature

INFLUENCE OF CERTAIN CLIMATIC FACTORS I9

perithecia and ascospores by the middle of November. He refers to
the perithecial stromata developing ‘‘abundantly in the autumn
around the old pycnidia.”’

Rogers and Gravatt (10, p. 45) report that in their inoculations
at Leesburg, Va., made on July 21 and August 16, 1912, pycnidia with
spore horns were developed by October 6. Although the cankers
were examined in March and again in August, I914, no perithecia
were found.

That unfavorable conditions may delay for a long time and perhaps
entirely prevent the production of ascospores was first brought to the
writer’s attention by inoculations of Endothia parasitica on chestnut
sprouts near Washington, D. C. These inoculations were made in
July, 1913, and produced abundant pycnidia within two weeks.
Sections of the stromata made in September, 1913, showed numerous
fundaments of perithecia. The inoculations were conveniently
located and as they were from the first material sent from China by
Meyer were frequently examined. The cankers continued to grow
normally and in most cases girdled the sprouts and formed numerous
stromata with abundant pycnospores and fundaments of perithecia.
Up to December, 1914, however, when the sprouts were destroyed by
fire, no ascospores had developed.

FIELD OBSERVATIONS

When this work was begun it was expected that ascospores would
be produced in the fall as had been the case in the work of Anderson
and Babcock and of Rankin and other investigators. Actually,
however, at none of the stations was a single canker in the entire series
of inoculations found which had produced ascospores or even mature
appearing perithecia during the season for 1914. In 1915, however,
quite different climatic conditions existed. Perithecia and mature
ascospores were found in abundance not only on cankers arising from
inoculations made in 1914 but from those made in May, 1915. The
problem then became not so much a comparison of the fructification
at different localities as a comparison of the fructification during
different seasons at the same locality.

Table VIII gives the results of observations at the various localities
on the development of perithecia and mature ascospores. It is evident
that no perithecia were produced during the season of 1914 at any of
the localities. Observations made December, 1914, at stations as

20

Observations on the Development

Locality Inoculations Made

Charlottesville, |Apr. 20, May 21,
Va. july 435 Aue:
Lt, and Oct. 2
1914.

Apry2n,; Junelo;
July 4, Aug. 1,
and Oct. 24,
I9I4.

Fairfax, Va.

Vienna, Va. Apr. 2, May 14,

June 6 and
July 18, 1915.

Washington, Apres22,. «Nilay.
DAG: 28, June 25,
July 28, and

Oct. 21, 1914.

Frederick, Md. |Apr. 27, May 30,
Aue. Oo wand

Oct. 19, 1914.
Woodstock, Md.

Apr. 28, May 14,
June 1, July
6; Aug, 10;
Oct: 7, 197A:
May 14, 1915.

Van Bibber,
IN Glaee ee ake

Apr. 29, May
14; june 1,
july652 rand
Aug. I0, I914.

Wilmington,
Del.

NEIL E. STEVENS

TABLE VIII
of Perithecia and Mature Ascospores at Various
Localities
Tecial Up dl betiitiectaey
fo and In; |Ascospores First Observed| Additional Notes
cluding
DE Cr HBR Cae ce oS ore ee Destroyed by fire
1914. last week in Apr.,
IQI5.
Dec. 23, |June 6, 1915. OnjOne of the trees in-
IQT4. cankers from all} oculated Apr. 21,
inoculations ex-| 1914, died during
cept those of Oct.| that summer and
24, 1915 and Aug.| no ascospores
6, 1915, from in-| were produced on
oculations of Oct.| this tree.
24, 1915.
Aug. I, |Sept. 21, 1915. A-|Perithecia more
I9I5. bundant from all] abundant on in-
inoculations. oculations of
June 6 and July
18 than previous
ones.
Dec. 25, |A few mature peri-|.. 25-4
IQI4. thecia, Apr. 22,
1915, from inocu-
lations except
those of Oct. 21,
1914. Abundant
July 28, 1915, on
all;
Dec. 20, |May 14, I915. AjIn 1914 there was
I9I4. few from inocula-| less development
tions of Apr. 27| of | pycnospore
and May 30, 1914.| horns from May
Dec. 27, |Aug. 9, 1915. Nu-| 30 “tom Aue le
1914. merous on all in-| than in the
oculations. month preceding
May 30.
Dec: 28)-|May 14, 19015) from). .2 2.6 eee
IQ14. all inoculations ex-
cept those of Oct.
7) TOTA VOCE.” 75
1915, from all in-
oculations, includ-
ing those of May
14, 1915.
Dec. 28, |May 14, 1915. Nu-|No perithecia were
I9I4. merous from in-| developed from

oculation of April} the inoculations

29. Fewfromin-| made Aug. Io,
oculations of May,| 1914, but the
June, and July.| number from the

Aug. 10, 1915. Nu-| earlier inocula-
merous onallmen-| tions were greatly
tioned above. increased.

INFLUENCE OF CERTAIN CLIMATIC FACTORS

TABLE VIII—(Continued)

Locality

Hartford, Conn.|/May 15, June 8,

Amherst, Mass.

Williamstown,
Mass.

Concord, N. H.

Locality

Stations on} Eleva-

Overlook | tion in
Mountain. | Feet
O7 600
Sig 1,000
3-1 1,500

25

Inoculations Made

July 15, Aug.
18, and Sept.

24.

May 17, June 8,
July 15, Aug.

No Peri-

thecia Up

to and In-
cluding

Sept. 23,

I9QI4.

May 17,
IQI5.

17, sept. 24,]|,

1914; May 17,
1915.

May 22, June 9,
July 14, Aug.
16 and Sept.
25, 1914, and
May 22, 1915.

May 18, June
11, July 17,
Aug. 20, and
Sept. 22,
1914, and
May 18, IQI5.

Inoculations Made

On or about
May 25, June
12, . july ~ 10;
Aug. I2, and
Oct. 31,, TO14:

Do.

On or about
May 25, June
E2,., July- io,
Aug. 12, and
Octy 1) 1914:
and May 25,
I9I5.

May 22,
1915.
Occasion-
al imma-
ture peri-

thecia
were
found on
this date.
May 18,
1915.

No Peri-
thecia Up
to and
Inciuding

Oct.,
I9I5.

Oct: “1,
IQI5.

May 24,
IQI5.

Perithecia with Mature
Ascospores First Observed

May 15}. 1915. A
few mature peri-
thecia. Aug. 18,
1915. Abundant
from all inocula-
tions.

Aug. 17, 1915. From
all inoculations,
including those of
May 17, IQI5.

Aug. 16, 1915. Nu-
merous from all in-

oculations except
those of May 22,

IQI5.

Aug. 19, 1915. Pre-
sent in cankers

on all dates, in-

May 18, IQI5.

Perithecia with Mature
Ascospores First

Observed
May 15, 1915. A-
bundant.
May 24, 1915. A-

bundant near the
center of one can-
ker.

Auge. 14, 1orss), A-
bundant in all can-
kers.

AugeTt; 1915, A-
bundant on all, in-
cluding inocula-
tions of May 24,
IQI5.

Additional Notes

Gets o.ce‘e ‘eles 8 6 © 0 © 0, ere

from inoculations!

cluding those of;

escee eee eee ec oO eo oO oO o

The trees inocu-
lated May 22,
1915, had unus-
ually thick bark.

Additional Notes

oe ee © © © © 8 8 8 8 eo ow 8

Old cankers which
had perithecia
were producing
pycnospore ten-
drils in large
quantities often
in the same stro-
mata.

22

NEIL E. STEVENS

TABLE VIII (Continued)

Locality :
No Peri- | Perithecia with Mature
‘Stations on| Eleva. | Inoculations Made thecia ve ' Ascospores First Additivnal Notes
Overlook | tion in Tene Observed
Mountain Feet
C 1,500 |On or about} May 23, |Aug. 11, 1915. Pre-|Perithecia less nu-
| May 25, June} I915. sent in cankers| merous than at
| 12; July 10} from all inocula-| the lower sta-
Aug 12, and tions. tions.
Oct. I, 1914.
{O56 1,500 Do. Oct. 1, |May 26, 1915. Pre-|.-. 3). .233 eee
IQI5. sent in cankers
from: inoculations
of May and June
I9I5. JeNibhee 108
1915, abundant in
all.
O1 1,900 Do. May 26, |Aug. 13). 1914. *°A-| a. 3a eee
I9I5. bundant.
O 2 2,500 Do. AUG 12; bebe wee eae. oh nee
: IQI5.
O4 2,800 /On or about] May 26. |Aug. 13. Maturel.....0. a4 9a
May 25, June} Nearly| perithecia from
12, July 10,| mature.| one inoculation of
Aug. 12, and May 14.
Oct. a" TOI
and May 25,
I9I5.
O 3 2,900 |On. or about| Aug. 13, |... .. 22... 2.cn |e eee
May 25, June] I9I5.
2. july 0;
AUS 2am
Oct. I, I914.
far north as Wilmington failed to show any perithecia. Perithecia

did, however, develop during the late winter and spring as far north
as Hartford, Conn., and up to an altitude of 1,500 feet on Overlook
Mountain. Perithecia developed also at both northern and southern
stations during the summer of I915 although they were somewhat
less abundant at Wilmington than at other stations and were found
at only one of the three highest stations on Overlook Mountain and
here only rarely.
‘TEMPERATURE

On comparing these data (see Table VIII) with the Weather
Bureau records it is evident that perithecia may be produced under
quite different temperature conditions. In our investigations they

were produced between December 25 and April 22 at Washington,

INFLUENCE OF CERTAIN CLIMATIC FACTORS 23

D. C., at Woodstock, Md., Van Bibber, Md., and Wilmington, Del.,
and they were developed during the period between December 26 and
February 15 at Washington Junction, Md. They were also produced
in small number at an elevation of 1,000 feet on Overlook Mountain
between October 1 and May 2. On the other hand, perithecia have
been produced in mid-summer at all stations from Concord, N. H.,
to Vienna, Va. Perithecia were not produced during the winter or
spring north of Hartford, Conn., and low temperature may in this
case have been a limiting factor. Certainly perithecia are developed
through a considerable range of temperature.

If
TABLE IX
Monthly Temperature Efficiency Indices for Various Localities

nee Wash- | Fred- | Wood- ae Hart- |Mohonk| Am- ee Con-
le ington, | erick, | stock, ca ford, Bake; }eherst, clea, cord,
Val D.C. Md. Md. él. Conn. | N.Y. | Mass. Mass. Neo.
eS Se eh eo eae ee |e ee
IQI4
Wha 6 3. 85 76 75 Fhe “a is Ae a ee 28
jie ae 89 94 96 93 95 7 66 66 63 60
iivecis 3... : Oj |e LOO. | 5080) “TOT 106 | 82 75 79 75 75
ENUCTISE Ss. S.. E17 OSmerO2) || TOs" etOO Mn) 87 79 82 74 Vhs
September...| 76 72 67 68 FO 65 65 58 51 | 57
October. .... 57 56 52 57 61 Ke) 46 A3 Bor Th 27,
November...| 28 25 22 22 a7, 15 m0) 9 ml 6
December ... 4 6 5 7 5 2 4 I 4 fe)
IQIs
JanwWary ..°. 5. 3 5 3 5 4 2 fe) O I O
February... LOW | e.9 4 6 10 3 O O I O
March ..... 6 I 2 2 5 2 fe) O e) fe)
2.00) n) Ae 61 56 55 52 52 36 39 36 22 29
MESS er = 63 61 G2. 'a-a7 44 45 42 39
che ere ; 84 mak 80 | 68 63 65 64 58
fautlis Se 105 98 | I10 | 86 Gal 82 76 aa.
Aueust...... 98 95 | I00 | 80 71 76 67 71

It has been rather generally believed that low temperature was a
determining factor in the production of ascospores by pyrenomycetes
and ascospores have often been loosely referred to as ‘“‘ winter spores,”
a term used indeed in connection with Endothia parasitica (8, p. 187).
That low temperatures are not necessary for the production of asco-
spores by Endothia parasitica is shown by the fact that they developed
before September 21, 1915, from inoculations made July 18, 1915, at
Vienna, Va., during which time no temperature below 54° was re-
corded, and the mean temperature was well over 70°. That high

24 NEIL E. STEVENS

temperatures on the other hand are not necessary is shown by the fact ~
that ascospores developed at many stations between December, 1914,
and May, 1915, and at Washington Junction, Md., between December
26, 1914, and February 15, 1915. Certainly (see Tables IX and X)
the difference between the summer temperatures of 1914 and of 1915
is so slight that the failure of perithecia to develop in the first summer
and their abundance in the second summer cannot be due to the dif-
ference in temperature.

TABLE X
Monthly Temperature Summations for Various Localities

ene Wash- | Fred- | Wood- | a Hart- |Mohonk} Am- ne Con-

alle ington, | erick, | stock, fon fod, Lake, | herst, vonen cord,
Wall DAC: Md Md. Del. | Conn. | NOY. | NS Yo ongecengie vanes

1014
May2nbrc co): 7054). 717) ||) 603 5029 716| 540 | 489 | 439 | 408 | 345
June wares 921 | 870 | 870 | 858 880! 618 | 599 ! 590 | 557 | 539
July... .... |, 1,000] 960") «952 4934 970 | 757 | 702 | 737 \7o7 4)sosce
AI ISt se: 1,048 | 892 | 931 | 952 998 | 825 | 745 | 770 | 676 | 676
September...| 672| 643 | 584 | 591 712} 541 | 549 | 488 | 404 | 441
October... ....)| «4804-481 4) 43a2" | 443 507| 341 || 347 1 313 7A \\ 282
November...| 188] 138 | 100 | 125 143| 67 26 29 18 23
December... . 25) = 26 31 36 29 8 14 3 II (a)
O15

Januarye: 3. 8 18 6 £2 17 10 O O 2 e)
February ..... 40'| 33 12 19 35 I ce) fe) 2 oO
Marcha 3st. 18 4 2 5 II 2 O fe) oO O
April 4a. ke: 501 | 449 | 432 | 430 426| 228 | 281% | 228 | 238°) sree
Mialy epee es: 604} 549 | 512 | 523 539 | 338 | 285 || 200 |) 232)55 gos
JUNE Ae ate 760! 774 | 754 | 718 736| 640 | 561 | 590 | 591 | 514
ulys, eae. 993 | 973 || 963 | 869 | 1,014 | 807 | 716.) 779) 727 \272e
PRUSUSE Gea. 899| 905 | 881 | 885 923! 743 | 641 li 703 1) 1622))| Rone

MOISTURE

There seems, however, to be a fairly constant relation between
the appearance of perithecia and the amount of precipitation, or more
properly the amount of moisture in the air. For convenience in-
reference a * has been placed in Table III to indicate the month in
which ascospores were first observed at the various stations. At
many localities perithecia were first noted in the spring, a season which
of course is characterized by high humidity. In each case in which
ascospores were produced during the summer the preceding months
were characterized by abundant rainfall. July, 1915, at Concord,
Williamstown, and Amherst, showed over 9 inches of rain and the
perithecial production was correspondingly abundant.

INFLUENCE OF CERTAIN CLIMATIC FACTORS 25

OBSERVATIONS IN ULSTER County, NEw YORK

Perhaps the most complete records regarding the appearance of
perithecia and ascospores are the inoculations at Ulster Co., New York.
As stated above, Rankin (9) found at Napanoch, N. Y., that asco-
spores were produced by the middle*of November from inoculations
made at different times from May 1 to August I.

TABLE XI
Monthly Climatological Data for Three Seasons at Mohonk Lake, N. Y.

Temperature Precipitation
Mean Summation © Efficiency In Inches Pee Ore:

TOT2.
2) a aan 58.6 440 54 3.99 II
| Ci ere 65.6 626 70 1.30 4
US a ae 70.6 802 86 3.42 II
EMUIOMISES 3/055 <.s 64.8 624 67 3.88 12
September..... 60.8 475 56 3.28 14
Qctober........ 53.9 mee ies 4.50 8

IQr4.
Ue son 61.2 489 er 4.10 k}
PME. ccs 64.8 599 66 2.40 7
|[CSee a eee 67.0 702 75 3.75 10
AUeuSt kk. 68.8 745 79 2.54 6
September..... 63.2 549 65 0.32 2
Octoberm...... 55.7 347 46 3.55 2

IQI5
1! 53.6 281 44 2.54 9
|| Cis ee ener 63.5 561 63 2.65 of
Waly ee ek 68.2 716 77 8.24 18
Proust... Ss 65.4 641 7a 7.94 10

The writer made a somewhat similar series of inoculations during
the summer of 1914 at Woodstock. Inoculations were made each
month in ten different localities on Overlook Mountain. None of
these produced perithecia during the season of 1914, but most of them
as well as inoculations made in May, 1915, produced perithecia abun-
dantly by the middle of August, 1915. As Rankin made over 1,500
inoculations and the writer made more than twice that number the
results were probably not due to chance but to a difference in the
weather conditions.

The nearest weather station to these two localities is at Mohonk
Lake, in Ulster County, elevation 1,245 feet. Mohonk Lake is between
Napanoch and Woodstock, about equidistant from them and has about

26 NEIL E. STEVENS

the same elevation. Observations made at this point while not ab-
solutely identical with conditions at either of the stations would un-
doubtedly approximate the conditions at both. This was certainly
true in the seasons under consideration for the Monthly Weather
Reports of that section indicate that the weather conditions recorded
at Mohonk Lake prevailed generally over the Eastern Plateau region.

Table XI gives the monthly precipitation, monthly mean tem-
perature, temperature efficiency, and temperature summation, for
the growing seasons of 1912, 1914, and 1915, at Mohonk Lake, N. Y.
Comparison of the data for the three seasons shows only slight dif-
ferences in temperature. June and July were warmest in 1912, August
and September warmest in 1914. These differences are, however,
slight, and can hardly have been significant in preventing ascospore
production in I914, since ascospores have been produced elsewhere at
higher as well as lower temperatures.

There is on the other hand a considerable difference in the pre-
cipitation of the three years. 1915, when ascospores were produced
abundantly before August 15, had much heavier rainfall in July than
either of the other years. In I912 ascospores were produced in
November; in 1914, on the other hand, no ascospores were produced.
It is then probably significant that August, September, and October,
1912, had a total precipitation of 3.88, 3.28, and 4.50 inches re-
spectively, as against 2.54, 0.32, and 3.55 inches for the corresponding
months in 1914, a difference of over 4 inches for the three months in
favor of 1912. This difference is best seen from the graphs, figure 3.
Distribution of rainfall is probably more important to the fungus than
its total amount since most of the moisture for the growth of the
fruiting bodies of the fungus must come from the outside. The
three months under consideration had 34 days with more than 0.01
inch of rainfall in 1912 and only Io in 1914.

Even this difference, however, does not give an adequate idea of
the difference in the two years, or of the extent and severity of the
drought of September, 1914. In August, 1912, the 3.88 inches of rain
came mostly after the middle of the month, the 14 rainy days in Sep-
tember were well distributed and October had a rainfall nearly an

$ The number of days with rain is of great importance to all vegetation in such
a region as that on Overlook Mountain where the run-off is very great and com-
paratively little moisture is left in the soil. The writer has discussed the run-off
of this region in another connection (12, p. 265).

INFLUENCE OF CERTAIN CLIMATIC FACTORS

191A SDS

18 19/13

19/13

19/1

I9/3?

FPREECTFITATION AT WASHINGTON O.C.

9
8
7
é
2
/
oO

.S-

191A (9K:

19/12

See Cs TENT Oi) ON Oy WO <0
TIEAGIM CY HIM 10° YIAO HALIM SAPO

19/4 /9/15

19/2

© y t %) \ \ A)
RQAHIN NN VIAN bY

PRECIATATION AT (MOHONA LAKE, ULSTER CO. Mir
9
o
7

Graphs showing monthly rainfall and number of days with rain for the growing seasons during three

3.

years in Ulster County, New York, and Washington, D. C.

27

28 NEIL E. STEVENS

inch above normal. Quite different conditions prevailed in 1914.

There was no rain in August after the 21st, only 0.32 inch in Sep-
tember, and no rain in October until the 16th, when two days’ rain
gave the 3.55 inches of rain recorded. It will be seen then that during
almost two months from August 21 to October 16 there were only
two days with appreciable rain and these totaled only 0.32 inch, while
from August 21 to November 1 there were only four days with any
rain. It is of course by no means certain that this extreme drought
was the cause of the total failure of the numerous cankers of Endothia
parasitica to develop perithecia. The condition is, however, very
suggestive, and it seems highly probable that a causal relation exists.

OBSERVATIONS NEAR WASHINGTON, D. C.

The number of inoculations made near Washington, D. C., is
much smaller than of those made in Ulster County, New York. The
data available, however, indicate a similar relation between climate
and ascospore production. Table XII gives the climatological data
for the seasons of 1913, 1914, and 1915 at Washington, D. C. There

TABLE XII
Monthly Climatological Data for Three Seasons at Washington, D. C.

‘Temperature Precipitation
Mean Summation Efficiency In Inches Deve en OES

TOT
May cAce eres 64.4 602 67 4.55 12
Une: 2eeysee 73.0 815 gI 1.81 10
sliwlyy, en cge veka. 78.0 IOII 112 2.24. II
AUIGUIS ta ee 74.0 909 97 5.43 10
September..... 67.0 676 75 2.41 6
October, 25450... 59.0 417 53 2:37 12
November..... 48.0 130 25 2.20 8

Ior4
Mae. hist te oe 67.0 Waka 76 T.72 5
HUNG shee tas veeee cat. 73.8 870 94 6.20 II
a italy. sia ete 75.9 960 106 222 8
AUOUSE, 25%. a neves 76.4 892 98 6.00 II
September... 2. = 66. 643 72 0.66 5
October: #05057 60.0 481 56 1.56 9
November..... 45.4 139 25 4.49 4

1915
IDEN Saigon oak, scr 62.5 549 63 ZiiNis) 10
June Saree. 70.6 Hafan 84 6.58 10
| DLAD coat he Marien cit see 76.1 973 105 BaP 12
NUS USty eee 74.0 905 98 7.00 15

INFLUENCE OF CERTAIN CLIMATIC FACTORS . 29

is little difference in the temperature of the three summers, although
1913 was somewhat warmer than the others. Both 1913 and 1914,
the years in which no perithecia were produced, had a decided drought
in the fall months. 1915, on the other hand, when perithecia appeared
abundantly by September, had 7 inches of rainfall in August. In this
locality, as in Ulster County, New York, perithecia appeared following
a period of abundant rainfall and failed to appear in dry weather. It
is somewhat surprising that perithecia failed to appear in August, I914,
since the months in Washington had a larger rainfall than the fall
months of 1912 in Ulster County. On the other hand, the temperature
was much higher in Washington during August, 1914, and the humidity,
therefore, presumably lower. This would indicate that it is humidity
rather than rainfall as such that determines the production of peri-
thecia. These data are in accord with the assertion originally made
by Metcalf (7) that in dry weather spore production was reduced and
that dry seasons checked the progress of the chestnut blight.

On comparing the climatological conditions at the two stations
for the three years during which observations were made, it is evident
that those years in which most ascospores were produced were the
years of most abundant rainfall and largest number of days with rain
regardless of temperature. If these conclusions are correct, tem-
perature has very little relation to the production of ascospores by
Endothia parasitica, whereas amount of moisture in the air has a
determining relation. This is probable on theoretical grounds since
perithecia develop on the dead tissues of the canker separated by a
considerable distance from any living tissues of the host, so that
moisture which reaches the developing perithecia must necessarily
come from the air. |

SUMMARY

A quantitative comparison of the available climatic data with the
growth and fructification of Endothia parasitica at various points from
southern New Hampshire to central Virginia has been made.

The area covered includes the northern limits of growth of other —
species of Endothia and is a transition region for several important
plant diseases. |

Eleven stations, extending through five degrees of latitude, were
chosen, as well as a series of stations at different elevations on Overlook
Mountain in the Catskills.

The stations were visited regularly during the summer of 1914.

30 > NEIL E. STEVENS

At each visit ten or more inoculations were made on healthy chestnut
trees and notes taken as to the growth of the previous inoculations.

The average annual lateral growth was found to be least at the
most northern locality, Concord, N. H., and to increase gradually
southward. The growth at Charlottesville, Va., was nearly twice as
great as that at Concord, N. H.

A similar relation was found among inoculations made on Overlook
Mountain, the amount of growth at elevations of 600 to 1,000 feet
being from 20 to 25 percent greater than that at elevations of 2,500
tO 2, Q00MeEt a:

The stations were all located near regular U. S. Weather Bureau
observation stations and no meteorological observations were taken.
This necessitated neglecting evaporation entirely, though evaporation
is probably less important in the case of a parasitic fungus growing
under the bark of a tree than in the case of most green plants.

The difference in the amount of growth of Endothia parasitica at
the various stations seems to bear no relation to the amount or fre-
quency of rainfall. Ambherst, Mass., and Charlottesville, Va. had
practically the same rainfall, yet the growth at the latter point aver-
aged nearly eleven cm. greater. On the other hand, stations differing
widely in rainfall showed practically the same amount of growth.

The length of frostless season is apparently unimportant, as the
fungus has no dormant season. Low temperature retards or even
prevents its growth, but growth is resumed as soon as favorable tem-
perature returns. Cultures kept at temperatures as low as 1° C. for
twenty-four hours resumed growth almost immediately when removed
to room temperature and grew as rapidly as cultures which had never
been chilled.

The amount of growth at the various stations is very closely related
to the duration and intensity of favorable temperatures.

In tracing the relation between temperature and growth, temper-
atures were calculated by direct summation as well as by the method
of temperature “‘efficiencies’’ suggested by Livingston and the results
of the two methods compared. The methods give nearly parallel
results, though temperature summations agree slightly more closely
with amount of growth than do temperature efficiencies.

The time necessary for the development of pycnospores is so short
that the climatic factors involved could not be traced.

The fungus in some cases continued to grow parasitically for over
eighteen months without producing ascospores.

INFLUENCE OF CERTAIN CLIMATIC FACTORS 31

No mature perithecia were developed at any of the stations during
“TO14.

Perithecia and ascospores were produced in abundance at many
stations during the late winter as well as the spring and summer of
IQI5. Ser

Air temperature had very little relation to the development of
ascospores. ‘They were matured both in midwinter and in midsummer
near Washington, D. C., in 1915.

There is a fairly constant relation between the development of
ascospores and the amount of atmospheric moisture.

Perithecia were frequently first observed in the spring, a season
characterized by high humidity.

The abundant rainfall during the summer of 1915 was accom-
panied by abundant ascospore production.

The results obtained by Rankin in Ulster County, New York,
during the summer of I912 agree with those obtained by the writer in
IQI5. :

A comparison of the climatological conditions of Ulster County,
New York, and Washington, D. C., for three seasons shows that years
in which ascospores were produced were the years of most abundant
rainfall and largest number of days with rain regardless of temperature.

‘During the period under investigation dry weather has certainly
tended to reduce the spread of the chestnut blight by reducing spore
production.

From the data presented in this paper the chestnut blight may be
expected to spread somewhat more rapidly in the Southern States
than it has in Pennsylvania and the states farther north.

BUREAU OF PLANT INDUSTRY,
WASHINGTON, D. C.

LITERATURE CITED

1. Anderson, P. J. Morphology and Life History of the Chestnut Blight Fungus.
Comm. Invest. and Control Chestnut Tree Blight Disease in Penn. Bull. 7.
1913.

2. Anderson, P. J., and Babcock, D. C. Field Studies on the Dissemination and
Growth of the Chestnut Blight Fungus. Penn. Chestnut Tree Blight Comm.
Bull. 3, 1913. (Literature cited, p. 45.)

3. Anderson, P. J., and Rankin, W. H. Endothia Canker of Chestnut. N. Y.
Cornell Agr. Exp. Sta. Bull. 347: 531-618, f. 77-101, pl. 36-40. 1914. (Bib-
liography, p. 611-618.)

32

10.

Il.

122

12;

NEIL E. STEVENS

. Képpen, W. P. Versuch einer Klassification der Klimate vorsugeweise nach

ihren Beziehungen zur Pflanzenwelt. 1901.

. Livingston, B. E., and Livingston, Grace J. Temperature Coefficients in Plant

Geography and Climatology. Bot. Gaz. 56: 349-375, f. 3. I913.

. Merriam, C. H. Laws of Temperature Control of the Geographic Distribution

of Terrestrial Animals and Plants. Nat. Geogr. Mag. 6: 229-238. pl. 12-14.
1895. :

. Metcalf, Haven. The Present Status of the Bark Disease of the Chestnut.

Science, n. ser. 31: 239. I9I0.

. Murrill, W. A. A New Chestnut Disease. Torreya 6: 186-189, f. 2. 1906.
. Rankin, W.H. Field Studies on the Endothia Canker of Chestnut in New York

State. Phytopathology 4: 233-260, f. 1-2, pl. IZ. 1914.

Rogers, J. T., and Gravatt, G. F. Notes on the Chestnut Bark Disease. Phy.
topathology 5: 45-47. I9I5.

Shear, C. L., and Stevens, Neil E. Cultural Characters of the Chestnut-Blight
Fungus and Its near Relatives. U.S. Dept. Agr. Bur. Pl. Ind. Circ. 131:
21S.) 1903:

Stevens, Neil E. Notes on the Structure and Glaciation of Overlook Mountain.
Ann. N. Y. Acad. Sci. 22: 259-266, fi:I-4. 1912.

Zon, Raphael. Meteorological Observations in Connection with Botanical
Geography, Agriculture, and Forestry. Mo. Weather Rev. 42: 217-223
fit.” T9014.

BUREAU OF PLANT INDUSTRY,

U. S. DEPARTMENT OF AGRICULTURE

GUNNERA PETALOIDEA GAUD., A REMARKABLE PLANT
OF THE HAWAIIAN ISLANDS

VAUGHAN MACCAUGHEY

A distinctive. feature of the Hawaiian flora is the prevailing en-
demicity of the rain-forest species. About 85 percent of the flowering
plants of the islands are endemic, and the bulk of these are character-
istic of the rain-forest zone. ‘This zone lies between the elevations of
2,000-6,000 ft. The mountains of Kauai, Oahu, East Molokai,
West Maui, and the Kohala Range on Hawaii, rise to heights of 3,000-—
6,000 ft., and thus their summits are covered with dense rain-forest
vegetation. The great valleys of erosion have eaten back into the
very hearts of these mountain masses, so that the summit regions
abound in knife-edged ridges and great precipices. Many of the
summit ridges are only three or four feet wide at the crest; many of
the precipices are 800—1,800 ft. high. The rainfall in these regions is
torrential, and much of the vegetation is of the most pronounced
hygrophytic type.

One of the most characteristic and conspicuous plants of these
humid summit regions is the endemic halorrhagaceous Gunnera
petaloidea Gaud.! In the little hanging valleys that abound in this
zone, on the precipices as well as in the steep stream-beds, are masses
of this titanic herbaceous-perennial. The gigantic leaf-blades are
three to four feet in diameter, peltate on fleshy petioles two or more
feet long. The petioles arise from a creeping or erect rhizome, which
is fleshy, green, and four or five inches in diameter. The huge crown
of leaves springs from the apex of the rhizome. As the latter is often
branched, the total mass of foliage was spread over an area of ten or
twenty square feet, with a height of six or eight feet. In places where
they have not been disturbed by the landslides that are common in
these regions, these gigantic herbs often cover areas fifty to a hundred
feet long and twenty or more feet wide, as on the upper slopes of a
precipice, where they form a beautiful mural tapestry.

1 See bibliography.

2 The blades of the Chilean G. manicata, the largest of the genus, are 5-10 ft.
in diam., on petioles of 6-7 ft.; these are used in Chile for tanning hides.

33

34 VAUGHAN MAcCCAUGHEY

The rhizome is very soft, and can be severed by a single machete
stroke. It contains a considerable quantity of crude starch, together
with numerous conspicuous fibers. It frequently contains colonies
of endophytic alge. It is closely pressed to the wet soil, but is not
subterranean; it roots freely along the entire undersurface. The
older, naked portions of the rhizome are green and conspicuously
marked with the large petiole scars. No bark is developed. The
apical region, 18-24 inches long, is usually more or less erect, depending
upon the situation; sometimes, as near a stream-bed, the rhizome stands
erect to a height of three or four feet. The entire length of the rhizome
is generally not over six or eight feet; its frequent branching and the
decay of the older parts tend to separate an old rhizome into several
shorter new individuals. This vegetative reproduction, quite similar
to that of many ferns, is the common mode of propagation after
the plant has once established itself.

The petioles are thick, fleshy, and curiously muricate; they are
three to four inches in diameter, and two to four feet long. The
broad, fleshy stipules, 1-14 in. long, are adnate to the base of the
petiole. The blade is orbicular or rounded-reniform. It stands at
right angles to the petiole.. Its attachment is peltate, but there is a
broad, open, basal sinus. It is very thick and fleshy, and deeply
rugose. Gray states that the blades are ‘15-2 ft. in diam. when
full grown”’; Hillebrand, that they are ‘‘2-3 ft. in width’’; both of
these are underestimates, and evidently based upon the examination
of herbarium material, rather than a knowledge of the plant in the
field. Leaves that are fully expanded are commonly three to four
feet in diameter, and Bryan records a diameter of five feet.

The blade is more or less conspicuously eight- to ten-lobed, the
lobes being very shallow, rounded and coarsely dentate. On its
upper surface the blade is covered with coarse, short hairs; the under
surface has a strong network of prominent veins. There are five
large veins, pedately arranged; the venation is dichotomous, and more
or less hispid with short, coarse hairs. A variety beta, collected by the
U. S. Exploring Expedition on Kauai, and described by Asa Gray,
has nearly glabrous foliage, with ‘‘bracts ovate or oblong, 6-8 mm.
long.”

The main flowering season is mid-summer, although there seems to
be considerable variation. The panicles are terminal. The rachis is
2-3 ft. tall, hirsute and scabrous, branching from near the base, and

GUNNERA PETALOIDEA GAUD. 35

grooved. The branches or spikes are 4-9 inches long, undivided,
crowded but lax and spreading; they are covered throughout with
clustered or scattering sessile flowers. The bracts of the inflorescence
are linear, scarious, I-13 in. long. The flowers are bisexual and not
bracteolate. The calyx is globular with adnate tube; there are two
lobes, one anterior and one posterior; these are persistent, each I-2
mm. long, broadly ovate or triangular, with broad or truncate apex,
denticulate or 3-toothed, with a raised line along the inner face. The
petals are two, alternate with the calycine lobes and 2 or 3 times as
long; cucullate, enclosing the stamens before anthesis; broadly ovate
or cuneate, retuse, obscurely glandular on the back, thickish in
texture, epigynous; tardily deciduous. The stamens are two, epigy-
nous, opposite the petals; filaments very short; anthers large, about
2 mm. long, emarginate at each end, somewhat didymous, fixed
by the base, introrse, the two cells opening longitudinally. Pollen
grains four-lobed. Styles two, opposite the stamens, and nearly
twice their length; linear-subulate, hispid, slightly united at the base.
Ovary one-loculed, with a single anatropous ovule suspended from
the summit of the locule. Drupes ovoid-globose, yellow, reddish or
purple, 2-4 mm. long, crowned with the short and incurved calycine
lobes; the calyx tube forming the fleshy sarcocarp; endocarp small in
proportion; acheniform, lenticular, 3- or 4-angled, crustaceous. Seed
conformed to the endocarp; testa very thin and delicate; embrvo
minute, near the hilar extremity of the fleshy and oily albumen, sub-
cordate, the radicle superior.

Schindler’s monograph of the Halorrhagaceae in Engler’s Pflanzen-
reich contains the following detailed description of petaloidea:

“‘Statura maxima, metrali vel ultra; rhizomate crasso, haud stolonifera, folia
subpauca rosulata apice procreante, ligulis chartaceis, glabris, + 45 mm. longis,
latissime ellipticis apice obtusis, pluries divisus induto, scapos floriferos complures
axillares proferente.

“Folia maxima, petiolo validissimo, + 0.6 m. longo, canaliculato, basi laxe piloso
superne glabro vel. glabrato, hinc inde aculeolis brevibus sueto fere punctiformibus
instructo stipitata; lamina depresse reniformi sat latiore ac longa, latissime cordata,
circuitu in lobos subaequales sueto 9 late rotundatos vel obtusos divisa, margine
grosse crenata, dentibus junioribus apice apiculatis senioribus obtusis, supra plana
nec prominenti-areolata, praeter nervos nervillosque supra perlaxe subtus densius
pilis crassis conspersos glabra, usque as 0.5 m., lata mihi visi.

“Inflorescentia scapo crasso, lineatim angulato, arcuatim adscendente, brevi
sed semper manifesto, juniore saltem dense pilis crassis conicis consperso elata, quam
folia bervior, © flora, densior laxiorve, optime thyrsoidea, apice breviter acuta,

36 VAUGHAN MacCAUGHEY

+ 0.4 m. longa, axi primario crasso, piloso, bracteis primariis conspicuis quidem sed
tamen quam ramuli axillares multo brevioribus, linearibus, apice subrotundatis,
glabris, integerrimis vel basin versus obscure denticulatis, 15 mm. vix excedentibus,
ramulis basi brevissime sterilibus, suberectis, pilosulis, primum dense demum in
florum statu @ elongatis laxius quaquaverse flores » gerentibus, usque ad 150 mm.
longis mihi visis sueto brevioribus; flores sessiles, 5 mm. longi, glaberrimi; ovarium
laeve, breviter lateque cylindricum, apice vix constrictum, minute 4-lineatum; sepala
brevissima, late squamiformi-triangularia, apice acuminulata, 0.5 mm. longa; petala
glabra, ex ungue brevi late lineari in laminam haud multo latiorem, cucullatam, apice
obtusam producta, + 2 mm, longa; stamina quam petala sat breviora, crassa, an-
theris fere orbicularis, apice obtusis, laevibus, quam filamenta brevissima crassaque
longioribus; styli crasse cylindrici apice acuti, dense papilloso-villosi.
‘‘Bacca exsucca, globosa, laevis, + 2.5 mm. diam. metiens.”’

The family Halorrhagaceae Schindler? comprises seven genera.
The family includes aquatic and terrestrial perennial herbs of widely —
diverse habit; some are minute, others, like the Hawaiian species, are
titanic in size. The flowers are small, axillary or in terminal racemes
or panicles, bi- or uni-sexual, regular; sepals usually 4, petals usually
4 or 0; stamens 8, the outer opposite the petals, or 4, rarely fewer;
ovary inferior, 1—4-loculed, each locule one-ovuled; fruit nut-like,
often crowned by the calyx.

The representation and geographic distribution of the genera is as
follows:

1. Loudonia Lindl.—3 species; Australia.

2. Halorrhagis Forst.—about 60 species; Australia, Tasmania, New
Zealand, Chatham I., New Caledonia, Chile, Juan Fernandez,
China, Lower India.

3. Meztella Schindler—1 species; Australia, aquatic.

4. Laurenbergia Berg.—18 species; Africa, Mauritius, Bourbon,
Ceylon, East Indies, Java.

5. Proserpinaca L.—2 species; ‘‘Mermaid Weed”’; North America,
Canada to Guatemala, in standing and slow-running water.

6. Myriophyllum L.—36 species; ‘‘Parrot’s Feather’’; cosmopoli-
tan, all continents, including Australia and many islands.

7. Gunnera L.

The last genus, Gunnera, was named in honor of Ernst Gunner,

a Swedish bishop and botanist (1718-1773), who wrote a local flora.
In Gunnera the leaves are radical, ovate or orbicular, and often gigan-
tic. The flowers are perfect, or rarely imperfect monoecious or poly-

’ Britton and Brown use the spelling Haloragidaceae, and include the genus
Hippuris, making eight genera.

GUNNERA PETALOIDEA GAUD. 27,

gamous; small, greenish, in simple or branched spikes or panicles, the
staminate flowers on the upper branches; flowers often packed on a
great cob-like spike; petals 2 or 3 or none; calyx none or with 2 or 3

lobes;

stamens I, 2, or 3; ovary I-loculed, bearing 2 filiform styles;

fruit a drupe; plants rhizomatous.
The geographical distribution of the known Gunnera species is as
follows: |

0 ON DUNHPWN

10.

II

12.

13;

14.

15.
16.
Ey.
ne,
1O:
20.
oA
22),

Sub-gen. I. Milligania (Hook. f. emend.) Schind.

. cordifolia Hook. f.; Tasmania.
. monoica; New Zealand, Chatham.

mixta Kirk; New Zealand.

. strigosa Colenso; New Zealand.

. prorepens Hook. f.; New Zealand.

. densiflora Hook. f.; New Zealand.

. dentata Kirk; New Zealand.

. arenaria Cheeseman; New Zealand.
. hamiltoniu Kirk; New Zealand.

Sub-gen. II. Misandra (Comm.) Schind.

lobata Hook. f.; extreme S. America.

. magellanica Lam.; high mountains of S. Amer., Colombian

Andes, Chile, Patagonia, Ecuador, etc.; alpine.
reichet Schind.; Chile (1,800 meters elev.).

Sub-gen. III. Pseudo-Gunnera (Oerst.) Schind.

macrophylla Blume; New Guinea, Celebes, Java, Sumatra,
Philippines, in high mountains.
perpensa L.; S. and E. Africa, Madagascar.

Sub-gen. IV. Panke (Mol.) Schind.

petaloidea Gaud.; Hawaiian Islands only.

bracteata Steud.; Chile, Juan Fernandez.

glabra Phil.; Chile, Juan Fernandez.

pyramidalis Schind.; Chile, Juan Fernandez.

peltata Phil.; Chile, Juan Fernandez.

pilosa Kunth.; Colombia and Ecuador, high mountains.
boliviana Morong; Bolivia.

apiculata Schind.; Bolivia, high mountains.

38 VAUGHAN MacCAUGHEY

23. rheifolia Schind.; Peru.

24. brephogea Linden; Colombia and Ecuador.

25. manicata Linden; Colombia.

26. bertero1; Phil.; Chile, high mountains.

27. chilensis Lam.; Chile, high mountains.

28. brasiliensis Schind.; Brazil.

29. vestita Schind.; Chile.

30. commutata Blume; Chile.

31. insignis (Oerst.) DC.; Costa Rica.

32. wendlandu Reinke; Costa Rica.

33. insularis Phil.; Juan Fernandez.

It is extremely significant to note that G. petaloidea is one of a
number of endemic Hawaiian plants that have very close affinities
with the Andean flora. It has been suggested that at one time in
the history of the Pacific there existed a land-bridge or its equivalent
connecting the now-remote Hawaiian archipelago with the South
American continent. Considerable evidence could be brought forth
to substantiate this view.*

Some of the typical habitats of this remarkable herb are: Wai-ale-ale
Swamps, Kauai (4,000-5,000 ft.); Ka-ala and Kona-hua-nui summit
ridges on Oahu (2,500-4,000 ft.); Pele-kunu Pali, Molokai (3,000 ft.)
East and West Maui mountains (3,000-5,000 ft.); and the Ko-hala
Range of Hawaii (4,000-5,000 ft.). It is thus evident that the
Hawaiian Gunnera occupies a distinct ecological zone—2,500—5,000
ft.—which in general is characterized by steep declivities and torrential
precipitation. It is never known to occur above or below the limits
of this zone, although its drupes could be easily carried by birds, and
it has abundant opportunity to descend mechanically to the lower
levels. A striking peculiarity for a plant of such magnitude is its
strong ‘‘ preference’ for very steep slopes, upon which it maintains an
apparently precarious footing. ‘These slopes have the advantage of
maximum illumination, but are constantly subjected to landslides.
In many of the regions enumerated above, Gunnera forms a tapestry
on inaccessible and nearly vertical cliffs. Field studies of Gunnera
give the impression that it has attained a relatively static condition,
with reference to range, and is neither markedly spreading nor losing
ground.

4 The ecology of Gunnera indicates that it has been a member of the Hawaiian
flora for a very long period of time; it belongs to the primitive flora.

GUNNERA PETALOIDEA GAUD. 39

The Hawaiian name for this plant is A pé or Apé-A pé; so far as is
known the natives did not utilize this plant in any way. Some of the
Gunneras of other regions are used horticulturally to produce luxuriant
foliage effects, for which purpose they are admirably adapted. The
Hawaiian species has not been utilized in this way; it is associated
only with the fog-swept precipices of Hawaii’s beautiful rain-forests.

BIBLIOGRAPHY

Gaudichaud. Freycinet, Voy. Bot. 512. 1826.
A. DC. in DC. Prodr. XVI, 2: 5907. 1868.
A. Gray. in Wilkes U. S. Expl. Exped. 151: 629-30. 1854.
Wm. Hillebr., Flora of Hawaiian Islands, 123. 1888.
A. K. Schindler. Halorrhagaceae, in Engl. Pflanzenreich, 23, IV, 225: 117. 1905.
L. H. Bailey, Standard Cyclopedia of Horticulture.
Wm. A. Bryan. Natural History of Hawaii.
Vaughan MacCaughey. Vegetation of the Hawaiian Summit Bogs. Amer. Botan-
ist 22: 45-52. I916.
COLLEGE OF HAWAII,
HONOLULU, HAwall

AN INTERESTING MODIFICATION IN XANTHIUM

CHARLES A. SHULL

Two years ago I received through the kindness of Mr. F. F. Creve-
coeur, of Onaga, Kansas, some burs of Xanthium which show a very
unique and interesting modification. The ordinary burs of Xanthium
are too common and familiar to need description. Normally they
enclose but two ovaries, and possess only two beaks which arise con-
jointly from the outer end of the bur. Through these beaks the styles
protrude at the time of pollination.

These modified burs, however, enclose a considerable number of
ovaries, usually between twenty and thirty. The beaks are corre-
spondingly increased in number, and are arranged in two or three
concentric rows about a central depression which occupies the central
part of the distal half of the bur. - Figure 1 shows the burs about
natural size. The form of the bur is probably determined by the
cessation of growth by the centripetal portion of the receptacle, while
the, centrifugal zone continues to develop, and imbeds a number of
flowers which are apparently arranged in more or less concentric rows.
Figure 2 shows one of the burs with the outer wall of the receptacle
removed so as to show the outer row of seeds, each of which is enclosed
in its black ovarial wall.

The exact structure of the bur is most easily understood from an
examination of the cross section of the bur taken slightly above the
equator, just beyond the bottom of the depression previously men-
tioned. Such a section is shown in Figure 3. Not all of the burs
had the same number of ovaries, but the general structure of all was
the same. In this particular bur there were twenty-six ovarial cavities
in the receptacle, twenty-three of which contained. the remains of
ovaries. The position of the cavities which contained ovaries are
indicated by small circles. The other three cavities contained no
trace of ovaries, but their position indicates clearly enough that they
correspond to a third row of florets.

There is a very strong tendency to sterility, apparently, for many
of the ovaries were empty. Of the twenty-three ovaries found in the

4O

AN INTERESTING MODIFICATION IN XANTHIUM 4I

bur shown in Figure 3, only twelve contained seeds. And a small box
of burs coming from direct descendants but two generations removed
from the original plant showed complete sterility, entire absence of
seeds.

Fic. 1. Xanthium canadense var. globuliforme Crevecoeur. Burs about
natural size.

The history of this interesting race of cockleburs extends through
three generations, beginning in 1909, at which time the original parent
was discovered growing in a corn field, by a farmer living near Onaga,

42 CHARLES A. SHULL

Kansas. This parent plant was given to Mr. Crevecoeur, who planted
some of the burs in his flower garden in the spring of 1910. He secured
a number of plants from them, and reports that ‘‘a portion of the
plants bore the same kind of burs as the well known X. canadense,

Fic. 2. Bur with wall removed to Fic. 3. Diagrammatic cross section
show seeds. of a bur showing arrangement of seeds.

while the major portion bore the same kind as the seed planted.’’
This would seem to fix the relationship of the type to X. canadense,
and Mr. Crevecoeur labelled the specimen in his herbarium X. cana-
dense var. globuliforme.

The following year, 1911, seeds of the plants grown in 1910 were
sent to Miss Grace Meeker, of Ottawa, Kas., who secured a third
generation. All of the plants in this third generation were of the
globuliforme type, but they produced burs which were small, and
devoid of seeds. With these plants the group became extinct.

Mr. Crevecoeur sent me quite a number of burs, and many seeds
were planted, but all were non-viable. In some cases patches of cells
in the cotyledons were still alive, but not a single hypocotyl showed
signs of life. The early loss of viability in this case was partly due,
no doubt, to the fact that the plants of the second generation, I910,
were destroyed while the burs were somewhat immature, in order to
prevent possible escape from cultivation. It is quite natural that one
sholud not desire the survival of a cocklebur producing twenty or more
seeds to the bur!

In the spring of 1915 two of the burs still remaining from the original
parent were planted at Onaga, but the seeds did not germinate. The
possibility of studying the inheritance of the bur characters in crosses

AN INTERESTING MODIFICATION IN XANTHIUM 43

and self-fertilized strains is thus precluded so far as this local ap-
pearance of the variety is concerned. From two other sources have
come vague reports of the same varieties in other localities, but in-
vestigation has failed to uncover them. However, it is possible that
intelligent observation by field botanists might lead to their re-
discovery.

Nothing is known regarding the cause, or manner of origin, of the
globuliforme type. The character of the modification is such that it
could hardly result from hybridization, although splitting was noted in
the 1910 generation. The cause of the sterility is merely conjectural
and might be due to various factors. Sterility of pollen, if it really
occurs, would not necessarily indicate a case of hybridization.

It seems more reasonable to consider it a mutation from X. cana-
dense Miller. Whether it is progressive, a new condition, or retro-
gressing toward remote ancestry, one cannot tell. But in view of Farr’s
recent studies on the origin of inflorescences and dicliny in Xanthium,!
the latter possibility is particularly significant. Farr reaches the con-
clusion that the bur is a reduced capitulum, in which the florets now,
of course, are reduced to two. If this globuliforme type is a reversional
mutation, it gives a concrete idea of the kind of capitulum from which
the reduction has occurred. Such a concrete picture is a distinct
advantage in any attempt to depict the lines along which such an
evolutionary advance has proceeded.

THE UNIVERSITY OF KANSAS,

LAWRENCE, KANSAS

1 Farr, Clifford H., The Origin of the Inflorescences of Xanthium. Bot. Gaz.

59: 136-148. I9Q15.

ENDOPHYLLUM-LIKE RUSTS OF PORTO RICO
E. W. OLIvE AND H. H. WHETZEL

The writers recently spent a little over two months in Porto Rico,
from February 23 to April 26, 1916, collecting and studying mainly
the parasitic fungi. A fairly representative lot of rusts were collected
from many localities about the Island. Among these were five
aecidioid and one peculiar uredinoid form which, after germination
studies, we found to be short-cycled and similar to, if not indeed
identical with, the Endophyllums.

We wish to acknowledge special obligation to Professor J. C.
Arthur, not only for determining all our rust collections after our
return, and for making many suggestions in the preparation of the
- systematic portion of this paper, but also for directing our attention,
prior to our journey, to certain unsolved problems, in particular to the
urgent need of clues in the case of the unconnected aecidia of Porto
Rico. For the preparation of the agar-water medium and for many
other courtesies we are much indebted to Plant Pathologist E. W.
Brandes and Director May of the Federal Experiment Station, as
well as to Dean Garwood, Professor C. E. Hunn and others of the
Agricultural College at Mayagiiez. For laboratory facilities and for
other freely tendered assistance we are also under great obligations
to Mr. J. A. Stevenson, plant pathologist, and to Director Tower, of
the Insular Experiment Station at Rio Piedras. After our return
from Porto Rico, most of the hosts of our fungi were determined by
Director Britton and others of the New York Botanical Garden;
the grass hosts by Professor Hitchcock and Mrs. Chase; the ferns by
Miss Slosson, to all of whom we desire to acknowledge our great in-
debtedness. We wish to express our thanks especially to Mr. Percy
Wilson of the staff of the New York Botanical Garden, who for several
days so generously placed his wide knowledge of West Indian plants
entirely at our disposal.

Arthur’s ‘“‘Uredinales of Porto Rico, based on collections by F. L.
Stevens,’’' which proved so very stimulating in our search, enumerates"
to aecidium-forms, all of which he at that time supposed to be heter-

1 Mycologia 7: 168-196, 227-255, 315-332. 1915; 8: 16-33. 1916.

44

ENDOPHYLLUM-LIKE RUSTS OF PORTO RICO 45

oecious. The discoveries of Kunkel? in the case of Caeoma nitens
(Schw.) Burrill and of Fromme,* in connection with Aecidium
tuberculatum Ellis and Kellerm., by means of which they proved
the teliosporic character of the supposed aecidiospores, also acted
as a great stimulus in our work. We tried to a limited extent
the agar medium recommended by Kunkel, but, laboring under
the rather trying tropical conditions, we came finally to use almost
exclusively the water surface method. This method proved very
efficient, as well as very simple and easy to manipulate. All our
germinations were tested successfully again and again by sowing the
spores on the surface of water drops placed on slides which were
supported up from the bottom of moist chambers. Inverted Petri
dishes, with a little water in the bottom to seal the cover, served
admirably for the latter.

In order to secure the best results, the spores must float on the
surface of the water, so that their germ-tubes may grow up into the
moist air. If; on the other hand, the spores are completely im-
mersed, the tubes then appear much like those from true aeciospores.
We found also that by chopping up bits of the host tissue with the
sorion them, and putting these so that they were not covered with water
but merely wet, much better and more abundant germination of the
telia resulted.

When once we became convinced of the short-cycled character of
one of these aecidium-like rusts, we became suspicious of all and deter-
mined to try out the spore germinations of every aecidioid rust with
which we came in contact. Our first successful find was in connection
with Aecidium Wedeliae, one of the commonest and most widely dis-
tributed of Porto Rican rusts. Professor Arthur states? that Dr.
Stevens had made the suggestion that the alternate host in this case
might be Cyperus, bearing Puccinia canaliculata (Schw.) Lagerh.
However, the very commonness of the Aecidium, occurring as it does
in all sorts of situations, all over the Island, on the host Wedelia
trilobata, combined with the comparative rarity of the Cyperus hosts
bearing Puccinia canaliculata, made us at once doubtful as to any
possible connection between the two. As stated above, our suspicions
were confirmed when trials of the germination of Aecidium Wedeliae

2 Bull. Torrey Bot. Club 40: 361-366. 1913; 43: 559-569. 1916. Amer.
Journ. Botany 1: 34-37. I9gI4.

? Bull. Torrey Bot. Club 42: 55-61. 1915.

AO hg oa Ree eh,

46 E. W. OLIVE AND HB. El WHETZEL

showed that the spores produce at once promycelia and that this form
is therefore a short-cycled Endophyllum and not a heteroecious form,
as had been thought.

In all, we germinated 13 aecidioid and uredinoid forms, in some
cases repeating the experiment several times in order to confirm our
earlier observations. In 7 of these, the spores germinated very
sparsely and very slowly; resulting at the end of 24-48 hours in a
few long, unseptated germ-tubes. We therefore became convinced
that in these 7 species (Aecidium passifloriicola P. Henn., A. tubulosum
Pat. & Gaill., A. Tournefortiae P. Henn., A. abscedens Arth., A.
Borreriae Pat., Uredo Trichiliae Arth. (ined.), and the aecial stage of
Uromyces proéminens (DC.) Pass.) we were dealing in all probability
with true aecia and therefore with long-cycled forms. We secured,
in fact, considerable evidence in two of the above cases as to possible
alternate hosts; coming to the conclusion that the first species was
probably associated with Puccinia Scleriae (Paz.) Arthur and the
second with Puccinia substriata Ellis & Barth. Mr. Stevenson, of the
Experiment Station at Rio Piedras, had also come to a similar con-
clusion in the case of the second—A. tubulosum on Solanum.

The slow and meager germination of the true aeciospores of the
above 7 forms is in marked contrast to that of the spores of the short-
cycled rusts described below. In the latter case, in an incredibly
short time, Io or 12 hours or even less, nearly all of the spores germin-
ated. When these spores are floated on the surface of water drops in
moist chambers, they push out into the free air a profuse mass of
unbranched, septate promycelia (basidia), each bearing the 4 (or in
some cases only 2) basidiospores (sporidia). It must be kept in mind,
however, in germinating these forms, that a source of error is liable to
arise if one is not extremely careful in the floating of the spores. When
entirely immersed, they always grow out into long tubes, rarely forming
sporidia, and might thus easily be mistaken for ordinary aeciospores.

The germ-tubes vary considerably in length as well as in other
characteristics in these Endophyllum-like forms. Sometimes, indeed,
even in the same lot of germinations, there is considerable variation,
due perhaps to their being grown sometimes in moist air only,
sometimes partially in water. Two of the species showed, how-
ever, a most remarkable variation, which is, in contrast to the above,
apparently not at all environmental. The spores of Endophyllum
Stachytarphetae and of E. circumscriptum, on germinating, produce

ENDOPHYLLUM-LIKE RUSTS OF PORTO RICO 47

only 2 spores to each promycelium, instead of the normal 4 basidio-
spores. We areas yet uncertain as to the constancy of this character;
neither are we yet oriented as to its probable significance.

The general characters and systematic arrangement of the short-
cycled rusts which we have found to produce promycelia are described
below, under 6 species. These are all considered in our title to be
Endophyllum-like, although it will be noted that only 4 of the 6 species
are really placed in this genus. The first one described is, in fact, not
at all aecidioid, but uredinoid in its fructifications; while the last one
of the list, while aecidioid, differs sufficiently from Endophyllum to
justify its being placed in a separate genus.

Botryorhiza Whetzel & Olive, gen. nov.

Cycle of development includes only telia.

Pycnia unknown (probably not formed).

Telia subepidermal, erumpent; teliospores thin-walled, oval, one-
celled, borne singly on long pedicels; each germinating apically on
maturity to produce a promycelium with 4 basidiospores; haustoria
botryose, or irregularly branched.

Type species, Botryorhiza Hippocrateae Whetzel & Olive, on
Hippocratea volubilis. The generic name is derived from the fact
that this form produces large, botryose haustoria, a character ap-
parently occurring also in certain smuts.°

Botryorhiza Hippocrateae Whetzel & Olive, sp. nov.

O. Pycnia wanting (probably not formed).

III. Telia mostly hypophyllous but sometimes amphigenous or
caulicolous, generally from a localized mycelium, sometimes from a
systemic invasion affecting entire young shoots; localized sori densely
crowded in more or less orbicular or irregularly shaped, somewhat
hypertrophied pulvinate areas, I mm.—I cm. or more across, the
affected areas yellowish when young, when older becoming whitish
due to the germination of the spores; in older leaves often killing
affected spots, which turn brown, the resultant rounded, swollen
dead areas then bearing a striking resemblance to certain insect galls.

Telia pulverulent, erumpent, from a definite, superficial, uredinoid

®>Lutman (Some contributions to the life history and cytology of the smuts.

Trans. Wis. Acad. Sci. 16: 1191-1244. 1910) has figured botryose haustoria in
Doassansia deformans. (See his figs. 44, 45.)

48 E. W. OLIVE AND H. H. WHETZEL

hymenium which arises just under the epidermis, without peridium;
teliospores uninucleate, borne singly at the end of pedicels which
arise from a binucleate mycelium, 13-14 by 18-24 yn, thin-walled, oval,
with a rounded apical protuberance, germinating apically at maturity
to produce each a long, cross-septate basidium (promycelium) bearing
4 basidiospores (sporidia), similar in shape to the teliospores and 8 by
[I-I2 py. | |

Vegetative mycelium composed of coarse intercellular hyphae,
made up of binucleate cells, some of which send large botryose, or
irregularly shaped, haustoria into adjacent cells.

id

On HIPPOCRATEACEAE:

Hippocratea volubilis L., Porto Rico (W. & O. No. 87, type; figs. I, 2).

It would indeed be peculiar if this conspicuous fungus had entirely
escaped description. We are, however, unable to find any published
matter pertaining to it. It is, perhaps, not so strange that it has
escaped inclusion in the rusts. In the collections at the Agricultural
Experiment Station at Rio Piedras we found it classed as an insect gall;
really quite a logical place for an old specimen, when judged alone
from its gross appearance.

As is well known, many tropical rusts are pale and inconspicuous
and otherwise quite unlike the yellowish or brownish rusts with which
we are familiar in colder climates; further, according to Professor
Arthur, “all of the so-called species of Eriosporangium and Argomyces
are white-spored, as well as the uredinia of Uredinopsis and many
others.’’ And he adds: “I see no reason why this is not a true rust,
although a very unusual one.’”®

It is, indeed, quite likely that the coarse mycelial hyphae and
the remarkable botryose haustoria will prove to be unusual features
among rusts; and that these are characters which are doubtless more
prevalent among smuts than among rusts. But, on the other hand,
the spores are cut off externally, much as in Uromyces, from the ends
of protruding hyphae; and, further, the spore-bearing hyphae are
always produced in a more or less regular, superficial hymenial layer,
which arises in hypodermal regions, generally just under the epidermis.
The latter are undoubtedly rust characteristics and not those of smuts.
It is of considerable interest, indeed, to find in this form characters
common to both smuts and rusts, thus adding emphasis to the general

6 In letter of October 6, 1916.

ENDOPHYLLUM-LIKE RUSTS OF PORTO RICO 49

belief in a common ancestry and a present near relationship for these
two:great groups. 7

Endophyllum circumscriptum (Schw.) Whetzel & Olive, comb. nov.

Aecidium circumscriptum Schw.; Berk. & Curtis, Journ. Phila. Acad.

INateoc., [lo2:°283.> 1853:

Aecidium Cissi Wint. Hedwigia 23: 168. 1884.

O. Pycnia epiphyllous, few, subepidermal, rarely breaking through
the epidermis, about 80-85 yw broad in section.

III. Telia amphigenous but mainly hypophyllous, aecidioid, nu-
merous, crowded, cup-shaped, borne in rounded, somewhat hyper-
trophied, pulvinate areas; peridium recurved, slit into a few coarse
segments; teliospores catenulate, more or less rounded-angular or
irregular from pressure, 12-13 by 15-18 yu.

ON VITACEAE:

Cissus sicyordes \.., Brazil; Costa Rica, Cuba, Dutch Guiana,

Jamaica, Porto Rico, St. Thomas (figs. 3, 4).

Endophyllum Wedeliae (Earle) Whetzel & Olive, comb. nov.

Aecidium Wedeliae Earle, Muhlenbergia 1: 16. I901.

O. Pycnia probably not formed.

III. Telia mainly hypophyllous, aecidioid, densely clustered, borne
in light yellowish areas of somewhat irregular shape; peridia scarcely
emergent, evanescent; teliospores catenulate, globoid or more or less
angular from pressure, 12-13 by 16-18 yp.

ON COMPOSITAE:

Wedelia trilobata (L.) Hitch. Porto Rico, Jamaica and other West
Indian Islands (figs. 13, 14).

This is perhaps the commonest of the Endophyllums growing in
Porto Rico. As stated above, it was this very abundance that made
us suspicious of any possible connection with Puccinia canaliculata,
as had been suggested by Stevens. .

Endophyllum decoloratum (Schw.) Whetzel & Olive, comb. nov,

Aecidium decoloratum Schw. Berk. & Curtis, Journ. Phila. Acad. Nat.
Soll 2282. 1853.

Aecidium Clibadu Syd. Ann. Myc. 1: 333. 1903.
O. Pycnia probably not formed.

50 E. W. OLIVE AND H. H. WHETZEL

III. Telia hypophyllous, aecidioid, in rounded or sometimes ir-
regular, more or less numerous areas, 2-7 mm. in diameter; peridia
evanescent, sometimes short cylindrical, with incised margin; telio-
spores catenulate, globoid or more or less angular from pressure, 12-13
by 16-13)u:

On COMPOSITAE:

Clibadium arboreum J. D. Smith, Mexico.

Clibadium Donnell-Smithu Coult., Guatemala.

Chibadium erosum (Sw.) DC., Porto Rico (figs. 11, 12).

Chbadium surinamense L. Dutch and French Guiana.

We found this Endophyllum only on the slopes of the eastern
mountains of Porto Rico, especially the foothills of El Yunque and
El Duque.

Endophyllum Stachytarphetae (Henn.) Whetzel & Olive, comb. nov.

Aecidium Stachytarphetae P. Henn. Hedwigia Beibl. 38: 71. 1899.

O. Pycnia probably not formed.

III. Telia hypophyllous, aecidioid, one to few in number to each
leaf, occurring in rounded, or somewhat irregular, rather inconspicuous,
pulvinate areas; peridia evanescent; teliospores catenulate, globoid
or more or less angular from pressure, 14-15 by 15-25 um.

ON VERBENACEAE:

Stachytarpheta cayennensis (L.C. Rich.) Vahl (Valerianodes cayen-
nensis (L. C. Rich.) Kuntze) Porto Rico, Santo Domingo, Bolivia,
Colombia (figs. 5, 6).

Stachytarpheta dichotoma Vahl, Brazil (E. Ule No. 2163.)

According to Professor Arthur, this is the first time this rust has
been reported from North America. We found it only at Rio Piedras,
in a little valley near the Experiment Station. This, also, was the
only locality in which we found the host.

Endophylloides Whetzel & Olive, gen. nov.

Cycle of development includes, so far as is known, only telia.

Pycnia unknown, (probably not formed).

Telia erumpent, the chains of spores adhering to form more or less
extended, cylindrical columns, about 2-4 times as long as broad, waxy
or horny when dry. Peridium wanting, or at least inconspicuous.

ENDOPHYLLUM-LIKE RUSTS OF PORTO RICO 51

Teliospores catenulate, one-celled, germinating at the apex of the
column.

Type species, Endophylloides portoricensis, on Mikania cordifolia.

This form differs markedly from Endophyllum in that the latter is
much more aecidium-like, with usually prominent peridium-cup and
pulverulent masses of spores. Similarly, while undoubtedly resem-
bling in some respects the type genus of Dietelia, D. verruciformis P.
Henn., yet we regard the absence of an evident péridium and the pos-
session of comparatively long, horny columns of teliospores in Endo-
phylloides, in contrast to the strongly developed peridial cells and the
globose or subglobose telia in Dietelia, as sufficiently distinctive to
warrant the formation of the new genus.

Endophylloides portoricensis Whetzel & Olive, sp. nov.

Aecidium expansum Arth. Mycol. 7: 317. I915 (not A. expansum
Diet.). |

O. Pycnia probably not formed.

III. Telia chiefly hypophyllous, sometimes petiicolous or cauli-
colous, short-cylindrical, forming more or less waxy or horny columns
about 14 mm. in diameter by 0.5-I mm. long, aecidioid, borne in irregu-
larly shaped areas, 0.5-I or more cm. in diameter; peridial cells in-
conspicuous, often collapsed, scarcely forming a continuous peridium;
teliospores rounded or oval, 12-15 by 15-20 uy, in long persistent chains,
separated from each other by prominent intercalary cells.

ON COMPOSITAE:

Mikania cordifolia (L. f.) Willd., Porto Rico (Whetzel & Olive,
No. 83, type, figs. 7-10).

Mikania odoratissuma Urban, Porto Rico.

The first host is found very commonly all over the Island; and the
fungus is also quite generally distributed. Mvzikania odoratissima, on
the other hand, is, in our experience, much rarer. Our collections of
the latter were made only on the mountain slopes of El Yunque and
El Duque, at the extreme eastern end of the Island.

BROOKLYN BOTANIC GARDEN AND CORNELL UNIVERSITY

52

E. W. OLIVE AND H. H. WHETZEI

EXPLANATION OF PLATES [-III.

All photos except Fig. 3 were taken by Mr. L. Buhle, of the Brooklyn Botanic
Garden. . Figure 3 is from a photograph taken in Porto Rico by Prof. Whetzel.

FIG. I.
FIG. 2.
FIG. 3.
FIG. 4.
Fic. 5.
Fic. 6.
FIG. 7;

Botryorhiza Hippocrateae, on leaves of Hippocratea volubilis.

The same, enlarged; X about 4.

Endophyllum circumscriptum, on leaf of Cissus sicyoides; X about 2.
The same; X about 34.

Endophyllum Stachytarphetae, on leaves of Valertanodes cayennensis.
The same, enlarged; X about 4.

Endophylloides portoricensts, on petiole of leaf of Mikania odoratissima;

enlarged; X about 4.

Fre: 8.
FIG. 9.

Fic. 10.
FaGs si
Fic, 12:
FIG.y1 3:
Fic. 14.

The same; X about 24.

The same, on leaf of Mzkania cordifolia.
The same; X about 3.
Endophyllum decoloratum on leaf of Clibadium erosum; X about 4.
The same; X about 3.
Endophyllum Wedeliae, on leaves of Wedelia trilobata; X about 4.
The same; X about 3.

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE I.

OLIVE AND WHETZEL: ENDOPHYLLUM-LIKE RUSTS.

; it i j i 2
} a } F y
if
] .
i P 3 : , ‘
5 i 2
! - Hi s
x a) ~ z= *
. — 5 4
/ a ‘
= 4 C. ' =
2) .
= =
; ae ‘ :
e / , :
ai / \ A ‘;
t <
r F
ie F
.
: ; 7 a”
.
- a
’ ‘ :
_ Fo . .
i fi é
A
7 s
Fe \ 7
' s
=. : | S
2 Z a
5
: ; 5
‘i 4
ri .
é . ‘
4
®
. 2 ‘
: iF

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE

OLIVE AND WHETZEL: ENDOPHYLLUM-LIKE RUSTS.

”
. A
i rd
,
j
ink,
2 6
; . .
. Z
. :
rian
if
= r ‘ =
: 4 ¥
i i. eee bag h
* Poe |
Fe es)
t any, ‘
uw : iY Fa Pox
iS 4 =
Bs P sess ;
é
; ak te
oS ‘ ; ‘
. %
i
\
h = fmt
3 = 5 SS sine
" ' —> a
« 4 As

-—- _

i“ 2

a thi SY

— + 4

se

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE III

OLIVE AND WHETZEL: ENDOPHYLLUM-LIKE RUSTS.

d p " a h
(TR IOVNC FSO! Re a he ee eee og

Baa : i ; Sai

Ry Busit SS Magen:
«Brechin Botanic : Carden i

iacace Seat 00 4 year. Sir
be ehtrees pa a sere COmHTARS

os

»

epi Separates should be ordered ee pr if 8 turned
yt be supplied free; ‘cover and addi onal :

oe KC?

) Remittances. should 1 be’ made’ ‘payable. tol
cents ae hae adie Ene all pone not Sie,

¥

a en ian aes Sa Ys

AMERICAN
JOURNAL OF BOTANY

=

VoL. IV FEBRUARY, I917 INOW

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA
MUTANTS!

ANNE M. Lutz

A. INTRODUCTION

The present communication is the second of a series of three, the
first having been published in a preceding issue of this JOURNAL? under
the head of ‘‘Mutants of Oenothera with diminutive chromosomes.”’
The three papers embody a portion of the results derived from a
prolonged study of the Lamarckiana group of Oenothera conducted at
the Station for Experimental Evolution’ (1907-1911), at the University
of Louvain in Professor V. Grégoire’s laboratory (1911-1912), and
later in consultation with Professor Grégoire by letter (1912-1914).

As stated in the first report of the series, the primary object of these
communications is to discuss, in the light of the Cold Spring Harbor
and Louvain studies, certain theories and conclusions which Gates
has given out from time to time and which Gates and Miss Thomas
(14) have based upon the results of their investigations.

The first paper described an interesting condition found in two
mutant types produced by 15-chromosome O. lata X 14-chromosome
O. Lamarckiana; one, a new form, O. aberrans, grown at Cold Spring
Harbor in 1908 and 1909, and the other, O. rubrinervis, grown at Am-
sterdam in 1912. The somatic cells of these three plants were shown

1 Briefly reported in a paper read before the Botanical Society of America,
December 29, 1915, and in a preliminary note published in Science (Lutz, ’16a)
entitled ‘‘ The production of 14-+-chromosome mutants by 14-chromosome Oenothera
Lamarckiana,”

* Amer. Journ. Bot. 3: 502-526. I916.

§ Maintained by the Carnegie Institution of Washington until March, I1g1t.

[The Journal for January (4: 1-52) was issued Feb. 3, 1917.]

53

54 ANNE M. LUTZ

to have fourteen chromosomes of the usual size and one small one.
The germ-cells were not examined. The significance of this 14*1-
chromosome condition in offspring of 15-chromosome mothers was
discussed in relation to the discoveries of Geerts, who showed in 1911
that seven of the twenty-one chromosomes of certain hybrids may
fragment and degenerate during reduction, and of Gates and Miss
Thomas (’14) who demonstrated that one of the fifteen chromosomes
of O. lata and certain Jlata-like forms may sometimes behave in a
similar namner.

In the paper just referred to, Gates and Miss Thomas announced
the precise somatic chromosome number of 21 plants falling under
the heads of O. lata, O. semilata and various lata-like forms. The
authors found that “‘all without exception contained 15 chromosomes”’
and have discussed many new and interesting features ‘‘in connection
with the behaviour of the extra chromosome and the phenomena of
degeneration.”’ Their researches appear to have led them to conclude
that the presence of the extra chromosome in 15-chromosome offspring
of 14-chromosome forms is invariably associated with /Jata or lata-like
characters in the soma of the mutant. Later Gates ('15a, pp. 147-148)
described a 15-chromosome mutant which he showed had a few
characters in common with O. /Jata and many others which were quite
unlike those of the latter form. It appears, however, that he re-
garded this mutant as a /ata-like form, since nowhere in this work has
he intimated that the discovery of 15 chromosomes in O. incurvata
has modified his previously expressed views concerning the relation
of /ata characters to the extra chromosome. In March of the same
year de Vries (15a, p. 187) described two types of offspring, besides a
mutant which Stomps had obtained from O. biennts semigigas pollinated
without castration by pure biennis. One of the two types, represented
by 8 individuals, had 15 chromosomes and he calls attention to the fact
that while these plants had the same number of chromosomes as
O. lata, they had none of the characters of the latter form. In December
following Gates ('150) recognized the fact that his mutant O. im-
curvata is quite different from O. lata, as is also the 15-chromosome
form which de Vries reported. He adds: ‘‘Hence we may say that
whenever a germ cell having 8 chromosomes fertilizes a normal germ
cell a new form is produced, though what its characters will be depends
upon various circumstances which need not be considered here. One
of the most important of these factors is probably the peculiar com-

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 55

bination of chromosomes received.’’ He does not state, however,
that these discoveries render untenable certain earlier conclusions of
Gates and of Gates and Miss Thomas concerning the relation of the
extra chromosome to lata and lata-like characters, but takes what
appears to be a last brave stand to save the day in the statement which
follows the words quoted above: “It is perhaps not inappropriate to
speak of all these mutants as belonging to the /ata series, or the series.
with an extra chromosome.’’ It must be conceded, however, that a
plant which de Vries clearly states has none of the characters of
O. lata, can not logically be held to belong to the Jata series of mutants.
Furthermore, it has been shown in the note which preceded this:
publication (Lutz, ’16a) that a number of distinct mutant forms, quite
unlike O. lata, have been found by the writer to have 15 chromosomes.
The chromosome number of each of the 9 unlata-like types reported
was ascertained, and their dissimilarity to O. lata fully recognized,,.
previous to the year 1913.

The primary object of this paper, therefore, is to discuss Gates’s.
and Gates and Miss Thomas’s theories and conclusions regarding the
extra chromosome at length in the light of the fact that many unlata-
like 15-chromosome mutants are now known and that many more
~ doubtless exist, in order that it may be shown that many of their
conclusions are untenable.

B. 15-CHROMOSOME MUTANTS

1. Has O. lata 14 or 15 Chromosomes? Is the Number of Somatic
Chromosomes Inconstant in this Form?

For a period of four or five years following the announcement of
the somatic chromosome number of O. Lamarckiana by Geerts in
1907, all mutant offspring of O. Lamarckiana, with the exception of
O. gigas, were supposed to have the same number of chromosomes as.
the parental form; namely, fourteen. Fifteen chromosomes had been
reported for O. albida‘ in one of the earlier notes published by the writer
(Lutz, ’08), but the discovery was not emphasized and doubtless was:
overlooked, with the result that O. gigas continued to be regarded as.
the sole mutant derivative of O. Lamarckiana with a chromosome
number differing from that of the parental form.

4Two plants. Notwithstanding the fact that they were offspring of O. lata

xX O. Lamarckiana, they were mutants, since O. albida was not employed as either
parent.

56 ANNE M. LUTZ

In the year following that in which the note was published concern-
ing O. albida, it was announced that 15 chromosomes had been counted
in two lata offspring of O. lata X O. gigas (Lutz, ’09). These plants,
of course, were not mutants, but it did not seem unreasonable to
suppose that mutant /ata would be found to have the same number of
chromosomes as the hybrids. Gates, however, had repeatedly an-
nounced 14 as the sporophyte number for O. lata and the evidence
produced seemed quite sufficient to support his claims. In a prelimi-
nary note published in 1907° he said (p. 260) : ‘‘The sporophyte number
of chromosomes in O. lata... is 14.’’ Speaking of O. lata in the
detailed report which followed,® he said (p. 92): “‘It has been deter-
mined from a number of counts in the prophase that the sporophyte
number of chromosomes . . ., is 14.’”’ (Italics not employed in the
original.) Again, later in 1907’ (p. 9), ‘‘In O. lata the count of chro-
mosomes was made in the pollen mother cells and found to be fourteen.
It has since been made in various somatic tissues of the flower, and is
found to be constantly fourteen so far as observed. There has been
no indication whatever that the number is ever higher.’’ (Italics not
employed in the original.) Again, on page 11, ‘‘Several plants of
O. lata and the pure O. Lamarckiana have been examined, all having
fourteen chromosomes.”’

As earlier stated, the primary object of the Cold Spring Harbor
studies of Oenothera, begun in 1907, was to ascertain whether or not
each particular combination of somatic characters, such as that rep-
resented by the type we know as O. Jata, for example, is associated
with a definite, fixed number of somatic chromosomes; in other words,
whether or not somatic chromosome number in Oenothera is constant.
Years of careful study, by the writer, of the vegetative characters of
plants from seedling to fruiting stage (never overlooking the importance
of taking note of the slightest deviation from the combination repre-
sented by the type) together with the precise determination of somatic
chromosome number in over 200 individuals, established the fact that
each combination of somatic characters is constantly associated with
a certain number of chromosomes; in other words, that each type of
plant has a definite, fixed number of chromosomes. It was therefore
announced in “Triploid mutants” (Lutz, ’12), and many times em-

5 Gates, ’07a.
6 Gates, ’07b.
* Gates, ’07-.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 57

phasized throughout the report (pp. 390, 416, 418, 420) that all in-
dwiduals having identical somatic characters from seedling to fruiting
stage invariably have rdentical somatic chromosome numbers, regardless of
the parentage or the origin of the plants in question. Owing to the con-
tradictory nature of the evidence produced by Gates and Lutz relating
to somatic chromosome number in O. lata, it was feared that this type
might be regarded as an exception to the rule. It was therefore stated
(p. 416) that the Cold Spring Harbor studies had established the
precise somatic chromosome number of 28 latas; that each had been
found to have 15 chromosomes, ‘“‘whether mutant, hybrid, offspring
of mutant Jata self-pollinated, or offspring of hybrid lata self-polli-
nated; whether grown at Amsterdam, Cold Spring Harbor, or the
New York Botanical Garden; and whether derived from de Vries’s
cultures, from plants descended from plants or seeds from de Vries,
or from plants of English ancestry, in no wise related to de Vries’s
cultures.”

Later in the same year in which the above announcement was made,
Gates (’12) stated that he had counted the chromosomes in one lata
plant and had found the number in this individual to be 15, and added
(p. 995): ‘‘From all the counts thus far made of the chromosomes in
O. lata it appears that the number 15 occurs at least in most individuals,
though the counts are perhaps not yet numerous enough to show that
15 is the number for all individuals, . . .”’ From a note added at the
conclusion of this report, it seems that the statement concerning
O. lata in “Triploid mutants” had not appeared in print or had not
attracted his notice at the time this communication was sent to
press. In the note he says that the preparations of O. lata described
in his first paper on the subject were re-examined, but that they had
deteriorated somewhat and he was “unable to determine with cer-
tainty whether this Jata plant contained 14 or 15 chromosomes.”’

Since Gates’s early statements were clearly based upon a number
of counts, it seems improbable that the extra chromosome, if present,

8 One may be led to inquire also whether 14+!-chromosome forms may not be
regarded as exceptions tothe rule. In considering this question it should be borne in
mind that we do not yet know whether the small member of the chromosome group
is constant or variable. Should future studies show the latter to be true, we should
then be called upon to decide whether these 14*!-chromosome forms should be re-
garded as actual exceptions to the rule, since the small body is, in*all probability,
not a chromosome, but merely a detached fragment of a whole chromosome, or a
remnant of a degenerating chromosome.

58 ANNE M. LUTZ

would have repeatedly escaped his notice. Owing to an unfortunate
error in identification during the first year of the work, Lutz (’08)
had announced 14 chromosomes for a plant supposed to be O. lata,
but later shown to be a distinct type.® It is possible, therefore, that
Gates mistook some /Jaia-like form having 14 chromosomes, for O. lata.
The number of individuals in which he counted 14 chromosomes is
not known; it is clear from the note referred to at the end of the pre-
ceding paragraph, that only one plant was mentioned in the first two
1907 papers, but his statements in the third that ‘Several plants of
O. lata and the pure O. Lamarckiana have been examined, all having
fourteen chromosomes,’’ certainly indicates that 14 had been counted
in more than one individual identified as O. lata. At any rate, Gates
appears to be convinced of error in count or identification in his early
studies of O. Jaia, since he states (’13, pp. 301-302) that Gates and
Miss Thomas’s studies of O. lata, etc. “corroborated the independent
results of Miss Lutz and Gates regarding the constancy of the fifteen
chromosomes in O. mut. laia, ”» Furthermore, Gates and Miss
Thomas (’14) not only emphasize the constancy of the 15-chromosome
condition in O. Jata without reference to the earlier count of 14, but
appear to be convinced that plants having 15 chromosomes invariably
have lata, semilata, or lata-like characters. In fact, in Gates’s recent
work (15a, pp. 167 and 296) he says, in referring to the Jata plant dis-
cussed in his first two 1907 reports, that his discoveries indicated

9 Certain forms which were studied during the first years of the writer’s work are
now known to have been erroneously classified (see ’12, p. 390, note 11, and 16), p.
514, note 7). The reappearance of the 14-chromosome form supposed to be O. lata
has shown that this mutant was not O. lata, though resembling it strongly in early
rosette characters (to be demonstrated in a later publication). The 16th chromo-
some of one figure of a second /Jata has since been demonstrated to me by Professor
Grégoire to be merely a deceptive anastomosis between two parallel chromosomes
although Gates (’12) has since reported two 16-chromosome cells in a 15-chromosome
lata and one 16-chromosome cell has also been found in a C. S. H. /ata. The 15-
chromosome form called O. nanella was a dwarf, not O. nanella. Likewise, it has
since been demonstrated that the 14- and 15-chromosome forms designated as ob-
longa did not duplicate each other and that the first type is certainly not oblonga,
though oblonga-like, and the second (type 5509) may be oblonga but is believed to be
a modified form of de Vries’s mutant. The latter type has continued to appear
in Lamarckiana and other cultures since 1908. These errors of identification were,
the result of premature publication of inexperienced work and are most regretable,
as they serve only to mislead others. By withholding later publications untilfidenti-

fications and results of investigations could be verified, it is BeOS that similar errors
have been avoided. ;

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 59

‘“‘about’’ 14 chromosomes, but that the number for O. /ata has since
been shown by Lutz, Gates, and Gates and Miss Thomas, to be 15.

These facts are outlined at length in order that the reader may
understand that mistaken identification of one plant by Lutz and
possible mistaken identification or error in count by Gates are
responsible for the early statement that O. lata had 14 chromosomes.
It may be assumed with safety now that the number of somatic chro-
mosomes present in the type of O. lata produced by O. Lamarckiana
is invariably 15.

2. O. lata and the ‘‘Extra’’ Chromosome

In ‘Recent papers on Oenothera mutations” Gates (’13, pp. 30I-
302), as stated above, mentions the then unpublished results of in-
vestigations conducted by Gates and Miss Thomas (’14) which had
disclosed 15 chromosomes in 21 plants classified by them as O. mut.
lata, O. mut. semtlata, O. lata to semilata, O. mut. lata rubricalyx,
O. biennis mut. lata and as lata-like forms. Referring to O. lata
rubricalyx which appeared among the F» offspring of two 14-chromo-
some forms crossed (O. rubricalyx X O. grandiflora), he says: ‘‘The
possession of fifteen chromosomes by this plant also shows that
whenever a meiotic irregularity leads to the formation of an individual
having an extra chromosome, such a plant will have the leaves and
habit of Jata or semilata.”’ Although he adds in a footnote that “It
is possible that one or two other mutants also have an extra chromo-
some,” he does not state or intimate that such forms are not Jata-like;
furthermore, Gates and Miss Thomas say in the later report (pp. 551—
552), ‘Certain other mutants indicate by their hereditary behaviour
that they may also have aberrant chromosome numbers, but this has
not yet been proved, except in gzgas.”’

Gates was the first to show that one of the heterotypic chromosomes
of a form may pass ‘‘into the same daughter-nucleus as its mate, instead
of into the opposite nucleus.’’ He first demonstrated this significant
irregularity in 14-chromosome Lamarckiana in 1907, but has since
observed the same peculiarity in many other forms. With reference
to this occasional 6-8 distribution of heterotypic chromosomes in 14-
chromosome forms, Gates and Miss Thomas say (p. 550): ‘‘ Whenever
this irregular meoitic division occurs in a pollen mother-cell, such a
cell will, at least in many cases, give rise to two Jata-producing pollen
grains in addition to two having only 6 chromosomes. The latter

60 ANNE M. LUTZ

apparently always degenerate. Similarly, when such an irregularity
occurs in the megaspore meiosis, if the 8-chromosome megaspore
functions it will, after fertilisation by a 7-chromosome pollen grain,
give rise to a Jata-like mutant. . . . Moreover, in lata or semilata
when crossed with their 14-chromosome parents or when self-pollinated,
the percentage in which the mutant reappears will depend upon the
relative number of their 8-chromosome and 7-chromosome germ-
cells which function.”

The authors then state that the frequency of this unequal division
appears, from the observations of Gates to be “‘of the order of I per
cent.” This, they say, would give about two 8-chromosome pollen
grains in 400, or 0.5 percent, and that “If the frequency of this ir-
regularity in the megaspore mother-cells is the same, about 1 per cent.
of Jata mutations should be anticipated.”’

Gates’s claim that ‘whenever a meiotic irregularity leads to the
formation of an individual having an extra chromosome, such a plant
will have the leaves and habit of O. lata or O. semilata”’ leaves no
loophole for escape from the conclusion that all 15-chromosome off-
spring of 14-chromosome forms—or at least all which are derived from
the fertilization of an 8-chromosome egg by a 7-chromosome sperm
—‘‘have the leaves and habit of O. lata or O. semilata,’’ while Gates
and Miss Thomas’s estimates of the percentages of offspring of O. lata,
selfed, and of O. lata X O. Lamarckiana which may be expected: to
reproduce the characters of O. Jata, lead the reader to conclude that
progeny resulting from 8 + 7 unions invariably have the characters
of O. lata, O. semilata, or some lata-like form.

All of the 15-chromosome mutants which Gates and Miss Thomas
mentioned in this report and which Gates has discussed in earlier
publications, were classified as O. lata,! O. semulata, lata to semilata
or Jata-like forms, and it appears that these were the only 15-chro-
mosome mutants whose somatic chromosome numbers had_ been
ascertained by them at that time; if such be the case, this chance
occurrence is probably responsible for their conclusions. That many
of the 15-chromosome mutant offspring produced by 14-chromosome
forms have J/ata or Jata-like characters cannot be questioned, but it is
equally certain that a far greater number do not. It seems to the

10 [Including O. lata rubricalyx and O. biennis lata.

11 As we have seen, Gates (’15a) has since reported a 15-chromosome mutant,
O. incurvata, which still later (’150) he says is quite different from O. lata.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 61

writer that the authors’ announcements were a bit premature in view
of the fact that the somatic chromosome numbers of so many mutant
offspring of 14-chromosome O. Lamarckiana had not been reported.
For example, the somatic chromosome numbers of O. scintillans,
O. spathulata, O. elliptica, O. sublinearis, O. leptocarpa, O. subovata,
etc., all Lamarckiana mutants, had not been announced at that time.
The same is true of O. nanella lata,” O. nanella oblonga and O. nanella
elluptica—compound types appearing in 14-chromosome O. nanella
cultures and elsewhere. Furthermore, as previously stated, it has
been shown (1908) that O. albida, one of the very common Lamarckiana
mutants, is a 15-chromosome form, yet this plant cannot be listed
either as O. lata or as a Jata-like form.

I have studied the somatic chromosomes of 305 plants of the
Lamarckiana group and have determined the precise number of 234
individuals. Exclusive of the two 14*!-chromosome mutant types
mentioned in the preceding report, of the offspring of O. lata X O. gigas,
O. Lamarckiana X O. gigas, O. nanella X O. gigas and of O. gigas,
selfed, 26 distinct mutant types were found among these 234 indwiduals
(17 among the 305) having more than 14 chromosomes; 11 of the 16 were
15-chromosome forms, 3 were 16- and 2 were triploid. The 17th type
having 14+ chromosomes was also a triploid form. In addition to the
above, one type having a number of characters in common with one
of the eleven 15-chromosome types mentioned, was also found to
have 15 chromosomes. Also, 15 chromosomes were repeatedly counted
in root-tips from a mutant grown at Amsterdam in 1912 and identified
by Professor de Vries as O. oblonga, but a certain irregularity found in
several tips of the plant (to be described later) indicated a possible
abnormal condition and made it seem inadvisable to accept the count
in this individual as typical of the species until verified by counts in
other oblongas.'4 Besides the II or 12 types which were ascertained,

12 Hunger (’13) found plants which he identified as O. lata nanella and O. oblon-
ga nanella in cultures of O. Lamarckiana.

13 | have since verified this count in 13 additional albida mutants.

14 Professor de Vries has kindly aided me in every way to determine the precise
number of chromosomes in the Amsterdam type. In the early summer of 1915 he
sent me a number of young oblonga rosettes from his gardens, but, unfortunately, all
perished before reaching their destination. He also sent me a generous supply of
seeds from one of his best plants, but very few of these germinated and only one
seedling survived. Root-tip fixations were prepared from this plant, but no satis-
factory counts have been obtained from them thus far.

62 ANNE M. LUTZ

beyond doubt, to have 15 chromosomes, 2 quite distinct types had
15(?) chromosomes (number not determined precisely).

We may now consider, briefly, the evidence furnished by these 15-
and 15(?)-chromosome forms.

(a) Distinct Types Having 15 Chromosomes.—Four of the 11 distinct
types are very common Lamarckiana mutants, though found in other
cultures, as well: (1) O. lata, (2) O. albida, (3) O. bipartita (C.S.H.)
and (4) type 5509 (C.S.H.), supposed to be a modified form of de
Vries’s oblonga. Among the less common forms are (5) O. nanella
lata,” obtained from de Vries’s culture of O. lata * O. Lamarckiana
(1912), but found also in cultures of O. Lamarckiana, O. nanella, etc.
(6) O. subovata, obtained from O. lata X O. Lamarckiana, but also
produced by O. Lamarckiana. (7) A dwarf mutant, type 2256, found
in a culture of O. nanella, (8) type 4499, produced by O. lata, selfed,
and O. lata X O. Lamarckiana; and three mutants which have been
observed in Jata cultures only, thus far: (9) O. exilis, (10) O. exundans
and (11) type 5365.

O. bipartita is a remarkably beautiful and interesting form. The
peculiarities of the young plant not yet come to flower are shown in
Figure 1. The leaves, particularly those of the young plants, are
thin and papery feeling; those of the adult form being more crinkled
and more finely crinkled, being somewhat broader in proportion to
their length, than Lamarckiana leaves. Although bipartita attains
the height of the tallest Lamarckiana, it is more dainty in appearance
than the parental form (Fig. 2). Like the latter, it produces a circlet
of basal branches which are somewhat shorter and less decumbent than
the rosette branches of Lamarckiana (Fig. 3). Not only are the
branches more slender, but the buds, which are regular and tapering,
are shorter and the flowers smaller than in the case of Lamarckiana.

O. bipartita is distinguished by the large number of flowers pro-
duced having more than 4 regular, tapering, stigmatic rays. Flowers
with 4+-rayed stigmas are common to most forms, yet the number of
flowers produced daily by bipartita having 4+-rayed stigmas forms a
higher percentage of the total than has been found to be true of any
other one of the freely blooming plants. The percentage of flowers
having 4+-rayed stigmas varies greatly among the individuals
of a given type. Daily records were made during the greater part of a

15 Professor de Vries states that he uses the term ‘QO. lata nanella”’ and ‘O.
nanella lata”’ interchangeably.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 63

Fic. 1. O. bipartita, plant No. 5561. C.S.H., 1910. Mutant offspring of
O. Lamarckiana X O. Lamarckiana. Not yet come to flower.

64 ANNE M. LUTZ

Fic. 2. O. bipartita, plant No. 5561, at height of flowering period. One of the
two uppermost flowers shows cleft petal.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 65

flowering season of the flowers produced by two biennial Lamarcki-
anas, and it was found that a larger percentage of the early, than
of the late, flowers had 4+-rayed stigmas.

Fic. 3. O. Lamarckiana, plant No. 5958, C.S.H., 1909. Offspring of O. La-
marckiana, selfed. | Photographed late in the season to show manner of branching.

This seemed to be due to a tendency on the part of the first flowers
of the stem and vigorous rosette and stem branches to have 4 +-, rather
than 4-rayed stigmas, though flowers with 4-rayed stigmas were com-
mon among the first, and flowers with 4+- were frequently found near
the terminal portions of these parts. After the plants had bloomed
a short time, it was found that the number of flowers having 4-rayed

66 ANNE M. LUTZ

stigmas exceeded the number of those having 4+-, and towards the
close of the flowering season it was seen that the number of the for-
mer greatly exceeded that of the latter.

On a certain day, at the height of the flowering season of 1910, 62
flowers unfolded on one bipartita mutant, and 52 of these had 4 +-rayed
stigmas. On August 31 of the same year, 80 percent of the 214 flowers

Fic. 4. Fe: O. lata X O. gigas, plant No. 4930, C.S.H., 1909. Flower

showing normal arrangement of petals.

produced by 9 ditpartita mutants had 4+-rayed stigmas, while less
than 1 percent of the 312 produced by 9 Lamarckianas selected at
random on the same day, were distinguished in this manner. All
had been in flower about the same length of time.

When a bud is held with the apex of the cone upward and the ele.
are then stripped backward, it will be found that the petals are rolled,

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 67

in the majority of cases, from left to right (viewed from the sepal side)
with the right lateral margin free and the left overlapped by the right
lateral of the preceding petal (Fig. 4). Occasionally, when a bud opens,
it is found that the relative positions of the right and left margins of two
neighboring petals are reversed, the left of one overlapping the right
of the other. A few buds have been found in which a complete re-

Fic. 5. F, O. Lamarckiana X O. gigas, plant No. 3750, C.S.H., 1908. Flower
showing reversed petal arrangement at two points, the left lateral margin of petal
I (J. m.) overlapping the right lateral of 4 (7. m.) and the left of 3 overlapping the
right of 2, leaving petals 2 and 4 in, and 1 and 3, out.

versal of relative positions had occurred at all four points; commonly
this takes place at but one. Complete reversal produces no disturb-
ance, but partial reversal frequently, though by no means invariably,
causes interference in the growing bud. For example, when the left
lateral margin of a petal, which we may designate as 2, becomes dis-

68 ANNE M. LUTZ

placed and overlaps the right lateral margin of petal 1, both lateral
margins of petal 1 are left in (Fig. 6), and both of petal 2, out.
Therefore the left of 2 and right of 4 are both out, and in the growing
bud sometimes forms a sort of an X contact along the left distal
margin of 2 and right distal of 4.° The right distal of 4 may grow on

ya \
Laima. ndim.
Fic. 6. F, O. gigas X O. Lamarckiana (de Vries), plant No. 133(1), Lafayette,
Indiana, 1913. Detached petals of flowers showing petal cleavage resulting from re-
versed petal arrangement at one point, the left lateral margin of petal 2 overlapping
the right lateral of 1, leaving both margins of 1 in, and both of 2, out, in open flower.
Most common form of irregular arrangement and petal cleavage. a,b = right distal
lobes; a’, b’ = left distal lobes; 7.m. = right lateral.margin; /.m. = left lateral mar-

gin; /.d.m. and r.d.m. = left and right distal margins.

16 Since Figs. 4, 5 and 6 show flowers photographed from the stigma, instead of
from the sepal, surface, the margins which are referred to as right and left in the
descriptions of the bud appear in reversed positions when viewed from the inside of
the open flowers. The margins are labeled, however, as seen from the sepal surface
of the bud.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 69

both sides of the left distal of 2, the latter, as it grows, being crumpled
into the slit thus produced by 4. Stripping back the sepals of a bud
about to flower, one will find the two petals locked in this manner,
the a lobe of 4 overlapping the a’ of 2, and the 0’ lobe of 2 overlapping
the } of 4.17 When the flower unfolds, petal 4 may be found with a
diagonal slit extending from about the middle of the right half of the
distal margin towards the median line of the petal. Petal 2 may or
may not have a slit extending from about the middle of the left distal
margin towards the median line. It is usually present, and shorter
than the slit in petal 4. In the case of the flower shown in Fig. 6,
petal 2 has the longer incision, indicating that lobe a’ of petal 2 over-
lapped lobe a of petal 4, and that lobe b of 4 overlapped 0’ of 2.
Flowers are found with one, two, three, or all four, of the petals cleft.
Sometimes a petai has two slits, one on each side of the median line.
In an irregularity such as that first described above, the right margin
of 1, as wellas the left of 3, is in, consequently these two someiimes
interfere, Causing an incision either in the left distal margin of 3, or the
right distal of 1, or both. Sometimes petal 1, both margins of which
are in, wraps around the filaments and anthers. The right and left
distal margins may then come in contact in such a way as to cause the
two incisions sometimes found in a petal. Not infrequently an ir-
regularity in petal arrangement causes no cleavage. Although cleft
petals are sometimes produced by overhanging anthers, in the majority
of cases they result from interferences caused by a partial reversal
of the direction in which the petals are normally rolled within the bud.

O. bipartita is distinguished by the large number of flowers produced
with cleft petals. Flowers with cleft petals are found occasionally on
individuals of almost any type, including Lamarckiana, but they are
more common to bipartita than any other mutant type observed.!8

17 The petal whose base is attached at the point where the filaments separate is
designated as petal 1.

18 The first flowers of vigorous rosette and stem branches (probably also of the
stem) appear to be more subject to this irregularity than those produced by the same
parts near the extremity. It is possible also that the first flowers of the short, weak
secondary branches produced late in the flowering season are less subject to this
irregularity than the first flowers of more vigorous parts, but the facts have not been
ascertained regarding this point. The buds of two biennial Lamarckianas were
examined daily (with occasional exceptions) throughout the greater part of the flower-
ingseason. Plant 4 came to flower June 23 and Babout thesametime. The former
was examined for cleft-petaled flowers for the first time on June 29, and the latter
on June 28. The results for these days and the 11 and 12 following, were recorded
as follows:

70 ANNE M. LUTZ

Scarcely a day but one or two cleft-petaled flowers were found on each
plant, and usually many more. Ona certain day, 22 of the 45 flowers
produced by one of these mutants had cleft petals. The records for
the 9 mutants employed for 4+-rayed stigma counts (see page 66)
on the day previously mentioned, August 31, will serve as a typical
illustration: 74 of the 214 flowers produced, almost 35 percent, had
cleft petals. |

The pollen of O. bipartita consists of 3-lobed grains. In the buds
of the mutants observed it was found to be entirely absent, produced in
small quantities, or present in moderate amounts; these conditions were
found in the various buds of each plant. A large percentage of the
grains produced are bad, and it is exceeding!ly difficult to obtain seeds
from these forms, selfed.

Type 5509, presumably a modified form of de Vries’s oblonga,
seems to bear about the same relation to the Amsterdam mutant as
type 3514 (see Lutz, ’16b) bears to de Vries’s rubrinervis; yet it is
possible that the two are identical forms.

(b) A Related Type, Having 15 Chromosomes.—Type 2806, a form
having many points in common with type 5509. Also found in
cultures of O. Lamarckiana.

(c) Distinct Types Having 15(?) Chromosomes.—These are (1) a
plant from de Vries’s 1912 culture of O. lata X O. Lamarckiana, said

A B

Number with Number with

Total Num- | Number with |Irregular Petal} Total Num- | Number with | Irregular Petal

ber of Irregular Petal} Arrangement ber of Irregular Petal) Arrangement

Flowers Arrangement | and Cleft Flowers Arrangement and Cleft
Petals Petals
June 2s... = a a= 30 13 8
‘ |

H 29. 54 3 | 3 52 4 2
30.60 67 5 5 70 4 4
Jaly “1. 63 e) 0) 82 2 2
Sina? BEDS ast = = aa 100 IO o)
e Beek 120 4 I 74 O oO
ae oe Te — — 90 — —
Chasm 132 I I 80 O O
ee One: 99 I I 59 oO oO
bp Bane 59 0) e) AS 0) fe)
x Sn an oa — — =
% Oa — — — 0) 0)
LOva. 22 Oo Oo O O

From July 10 until the close of the flowering season cleft-petaled flowers con-
tinued to appear occasionally, but much less frequently than during the early
flowering period of the plants.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 71

to have combined the characters of O. lata with the smooth, shining,
leaves of O. Leata. (2) O. elliptica, a Lamarckiana mutant.

Of the 11 distinct types known to have 15 chromosomes, (1) O. lata,
(2) O. albida, (5) O. nanella lata and (6) O. subovata are well-known
forms, originally described by de Vries; (4) type 5509 was mentioned
in an early note (Lutz, ’08); (7) type 2256, as previously pointed out,
was erroneously identified in 1908 as O. nanella. All of the remaining
five forms are new. The somatic chromosome numbers of O. laia,
O. albida, types 5509 and 2256 were reported in early notes by Lutz;
those of O. bipartita, O. nanella lata, O. exilis, O. subovata, O. exundans,
types 4499 and 5365 were communicated in the preliminary note to
this paper (Lutz, ’16a). 3

A mutant of the Lamarckiana group is distinguished from the
parental form and other mutants, not so much by some particular
character—for few characters are peculiar to any one type alone—as by
‘the combination of characters which is peculiar to itself. Thus,
O. lata has broad, heavily crinkled leaves, irregularly shaped buds
(particularly true of the early buds), light yellow flowers with crumpled
petals, barren anthers, etc. No one of these characters is peculiar to
O. lata alone. The cleft petals and large percentage of flowers having
4-+-rayed stigmas are striking characters of O. bipartita, yet neither
is peculiar to this form alone; it is the combination of characters pre-
viously enumerated which distinguishes it from all other forms. A
very striking illustration of this point may be found in the previously
mentioned 15-chromosome mutant reported by Gates ('I5a, pp.
147-148), namely, O. incurvata. His illustrations and descriptions of
this form clearly show that it is not entitled to be regarded as a lata-
like form, for, although he states that it agrees with O. lata ‘‘in the
obtuse tips and deep crinkling of the leaves,’ he also says that it
differs from O. lata ‘‘(1) in the much narrower leaves with long
petioles, (2) in having one edge of the leaf characteristically folded
over, (3) in being as tal! as Lamarckiana with long internodes, (4)
in having more squarish buds which produce pollen.’ If, in con-
nection with these statements, one compares his photographs of
O. incurvata (Figs. 56 and 57) with that of O. lata (Fig. 37), all in the
rosette stage, one will see that O. lata and O. incurvata are about as
unlike as any two mutants which may be mentioned. It is quite
clear that the ‘‘obtuse tips and deep crinkling of the leaves’’ do not
entitle this form to be regarded as J/ata-like, since it is wholly unlike

TP ANNE M. LUTZ

O. lata in the majority of its characters. The tendency of the margin
of the leaves of ¢ncurvata to roll towards the upper surface of the midrib
is one of the most striking characteristics of the full-grown rosette
leaves of O. albida (compare Fig. 7 with Gates’s Fig. 56).

Fic. 7. O. albida, plant No. 3472, C.S.H., 1908. Offspring of O. Lamarckiana
x O. Lamarckiana. Mutant in late rosette stage showing margins of leaf blade
rolling towards the midrib; a typical albida character.

In view of the above facts, it is not surprising to find that a few of
the twelve 15-chromosome types had one or two characters suggestive
of O. lata, just as others had one or two suggestive of O. Lamarckiana,
of O. rubrinervis or some other form; yet, since the majority of the
characters were wholly unlike those of O. Jata in the first case and
wholly unlike those of O. Lamarckiana and O. rubrinervis in the latter
instances, the first could not be called Jata-like nor the latter La-
marckiana- or rubrinervis-like. In fact, only 2 of the 12 types were
lata-like; namely, O. lata and O. nanella lata. On the other hand,
2 plants (2 types) were found in Cold Spring Harbor cultures of O. La-
marckiana, which were conspicuously lata-like in appearance, though
differing from O. lata sufficiently to be regarded as distinct forms, and each
had 16, and not 15, chromosomes.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS TS

De Vries (’09, Vol. I., p. 224) gives a table showing the origin of
new species from the type, O. Lamarckiana. From a first generation
of 9 Lamarckianas he records 7 generations of descendants through
O. Lamarckiana (1886-87 to 1899, inclusive) consisting of a total of
54,334 individuals, of which 834 were mutants, as follows:

TABLE [
O. La- : eee,
: marck- |O.nanella| O-7¥67t-| ©, lata |O. albida \O: Scintil-|¢, oblonga| O. gigas
Species Srp 14 neruis Is 15 lans 14 OF 15 28
14 14 or 14t] Prob. 15
oe fi ng eS a —_— —
Number of in-
dividuals. ..] 53,500 158 B2 229 56 8 350 I

Although the somatic chromosome number of O. scintillans is
probably 15, the actual number is unknown, hence this type will be
excluded from our calculations. If oblonga has 14 chromosomes, we
see that approximately 20 percent of the total number of 15-chromo-
some forms were neither O. /ata nor lata-like individuals. If oblonga
has 15, then we see that almost 64 percent were neither O. Jata nor
lata-like forms. But de Vries states that the list is incomplete, as
only the more important mutants were recorded; furthermore, since
the records date from 1886, it is probable that even the common
types were occasionally overlooked in the early years of the work;
in fact he says (p. 229) that albida was passed by as a diseased
form in 1888 and 1890. For these reasons the records of his 1895.
Lamarckiana cultures are perhaps more significant (pp. 262-263).
Of the total of 10,614 offspring of O. Lamarckiana he states that
614, or about 6 percent were mutants, “of which O. albida made
Hp. 251-7, 0. lata 1.7--%, O. nanella 1.1 %, O. oblonga 0.7 %
and the rest altogether 0.1 %.’’ If we include oblonga and the un-
named mutants among the 15-chromosome forms, we see that the
total number of albida plants alone equalled the total number of all
other 15-chromosome mutants listed, including O. lata. We will
assume that some of the unnamed types had 15 chromosomes and
others 14 and that oblonga also had 14. If such were the case, the
number of albidas not only exceeded the number of latas, but exceeded
the combined number of 15-chromosome forms not listed as O. albida.

It is well known that a large percentage of Oenothera seeds fail to
germinate in the short time commonly allowed them when sown in

74 ANNE M. LUTZ

seed pans in January. De Vries (15a) and Davis (’15a) have em-
phasized this fact recently and have suggested means of greatly
increasing the percentage of germinations. ‘‘. . . we can not feel
confident,’’ Davis states (’150), “that the records of any cultures of
Oenothera so far reported are complete for their possible progeny.
The percentages calculated for ‘mutants’ and the ratios of classes in
breeding experiments can not be accepted as final in exact genetical
work. Weare not ina position even to guess what may be the changes
of front when exact data become available.’’ It is clear that future
records of cultures, to be of value, must show that they are complete
for their possible progeny.

Asarule, seeds are obtained 1n greater abundance from 14-chromosome
forms selfed, or pollinated by other 14-chromosome forms of the same, or
different species, than from 14+ -chromosome forms selfed, or pollinated
‘by other 14-+-chromosome plants of the same, or different species,—par-
ticularly tf the 14 +-chromosome individuals have more than 14, but fewer
than 28, chromosomes.

Beginning with the summer of 1908, I adopted the practice of
counting all seeds sown; of planting seeds at spaced intervals in seed
pans, and of recording the germinations. Only ina few instances have
these precautions been neglected. The results have clearly shown
that when seeds not more than one year old are sown in pans of stert-
lized soil in January and kept under ordinary greenhouse conditions,
usually larger percentages of germinations are secured within the first
four or five months from seeds of 14-chromosome plants selfed, or polli-
nated by other 14-chromosome plants of the same, or different species,
than from 14+-chromosome plants selfed, or pollinated by other 14+-
chromosome plants of the same, or different species—particularly tf the
14-+-chromosome individuals have more than 14, but fewer than 28,
chromosomes. Hence it appears that the number of seeds produced by
a form and the ability of the seeds to germinate, at least within the time
limits specified, are factors which are associated with the chromosome
number of the plant, or numbers of the plants, producing them. The
ability of a seed to germinate appears to depend, not wholly, but to a certain
extent, upon the number of chromosomes which it: bears, and, possibly,
an accordance with Gates’s suggestion (15a, p. 194), upon the compait-
bility, or incompatibility of the chromosomal combination which the
number represents. It also appears that the ability of a seed to ger-
minate 1s directly associated with its own chromosome number and only

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 75

indtrectly with that of its parents, for the results derived from the Cold
Spring Harbor and Louvain studies indicate that 14-chromosome off-
spring of 14+-chromosome forms may germinate as readily as 14-
chromosome seeds of 14-chromosome forms. |

In the discussions of this report and others to follow, due allowance
will be made for the facts above mentioned. With reference to the
relative percentages of the various types of mutants appearing in de
Vries’s cultures, as quoted, these may not represent the actual per-
centages as they would have appeared had all the viable seeds sown
germinated, but they offer sufficiently satisfactory evidence to prove
that many 15-chromosome offspring of 14-chromosome Lamarckiana,
alone, have neither /ata, semilata, nox lata-like, characters; furthermore,
it is probable that the majority of these 15-chromosome mutants,
whether /aia-like or not, are products of 7 + 8 unions. However, if
we believe that each 15-chromosome mutant is represented by a par-
ticular chromosomal combination, then we will agree that an occa-
sional 6 +9 union might produce the same combination as 7 + 8
and that the same type of mutant might result from the former as from
the latter. This possibility may be illustrated very simply.

Throughout this paper and in future reports, when we wish to
indicate difference in sex without designating the particular sex of
either gamete, we shall employ small capitals in italics to represent
the chromosomes of a gamete of one sex and the same, marked ’,
to indicate the chromosomes of a gamete of the opposite sex; thus,
(Comoro) Arcprre + (of or 9) A’ Bc’ pz FG = Aa’ Ba! cc!
pp’ EE’ FF’ GG’, When it shall be necessary to indicate the sex of
gametes, the chromosomes of the female will be designated by lower
case letters in italics, and those of the male by the same, marked ’;
tins 9: abcdefs + o a’ b' cd’ ef! g’ = aa’ bb’ cc’ dd’ ee’ fF ge’.
Now, assuming that the regular female gamete of O. Lamarckiana
containsa bc def gchromosomes and the regular malea’ b’ c’ d’ e’ f’ g’;
that the somatic cells of this form contain aa’ bb’ cc’ dd’ ee’ ff’ gg’
chromosomes; then aabcdefg-+a'b’c'd'e'f'g’ might produce
O. lata having aaa’ bb’ cc’ dd’ ee’ ff’ gg’ chromosomes. So also might
aa bbcdefg+a-c'd'eé' f’ g’ produce O. lata having aaa’ bb cc’ dd’
ee’ ff’ ge’ chromosomes. While it is possible that a 6-chromosome cell
is incapable of functioning in union with one having 7 chromosomes,
or fewer, but is capable of functioning in union with one having 8 or
8-+ chromosomes, thereby producing a 14- or 14+-chromosome con-

76 ANNE M. LUTZ

dition (Lutz, ’12, p. 424), it cannot be assumed with safety that these
common 15-chromosome Lamarckiana mutants result from the fusion’
of 6- and 9-chromosome gametes, except, possibly, in rare instances,
for 5-9 distributions of heterotypic chromosomes doubtless occur still
more rarely than 6-8, and a 9-chromosome cell would be expected to
unite with a 7- far more frequently than with a 6- and to produce a
16-chromosome mutant; yet 16-chromosome offspring of O. Lamarcki-
ana X O. Lamarckiana appear to be comparatively rare.

Gates (09a, pp. 4-5) has pointed out that, owing to irregularities in
chromosomal distribution, a germ cell might be formed containing two
chromosomes of one pair and lacking both representatives of another
pair. The number of chromosomes would therefore remain constant,
he states, but such germ-cells would be entirely deficient in a
particular kind of chromosome. He has further shown (15a,
p. 298) that if both members of one pair of chromosomes
may pass to one pole of the heterotypic spindle, resulting in a 6-8
distribution of chromosomes, it is conceivable that both members of
another pair might, on rare occasions, pass to the opposite poles at the
same time. This would equip each daughter nucleus with 7 chromo-
somes, but not with the usual combination, 4BCDEFG. Let us
assume that this has occurred during male reduction and that two
pollen grains bearing a’a’—c’ d’ e’ e’ f’ 9’, and two bearing — bb’ c’d’e’
f’g’ chromosomes have been formed. Then should one of these male
gametes, say of the first type, unite with a regular 7-chromosome female
gamete, we should expect the 14-chromosome plant resulting to have
aa'a’ b— cc’ dd’ ee’ ff’ ge’ chromosomes instead of the usual aa’ bb’ cc’
dd’ ee’ ff’ gg’ combination. Gates and Miss Thomas suggested that
“the variability of the lata-semtlata series may depend upon the fact
that the extra chromosome belongs to a different pair in different cases,”’
and add: “‘since there are seven pairs of chromosomes, we should then
expect seven more or less distinct Jata-like types,’’ but conclude that
“there is at present no evidence that the plants having 15 chromosomes
can be divided in this way.”

If both members of any one of the seven pairs of chromosomes
were capable of passing to one pole, while both members of any one
of the remaining six were capable of passing to the opposite pole
during male, as well as during female, reduction; if regular and ir-
regular 7-chromosome male and female gametes were formed capable
of uniting with each other and producing viable seeds, a large number

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS Tape

of 14-chromosome combinations would be possible. If the common
types of 15-chromosome mutant offspring of O. Lamarckiana could be
limited to seven, one might assume that these are “half mutants”’
(borrowing de Vries’s term, but applying it differently) resulting in
each case from the union of a regular 7- with an 8-chromosome gamete,
each of the latter entering into the union differing from every one of
the remaining six with respect to the particular extra chromosome
which it possesses. The union of any one of the many irregular 7-
chromosome gametes with an 8- might produce one of the rarer types
of 15-chromosome mutants and such a form might be regarded as a
““whole mutant” (de Vries). The objections to these suggestions are
obvious: irregular 7-chromosome distributions would be expected to
occur more rarely than 6-8, and germ-cells resulting from the former
would be expected to unite with regular 7-, producing 14-chromosome
half mutants in the vast majority of cases, and to unite with cells having
8 chromosomes only in extremely rare instances. We would be obliged
to conclude that the common 14-chromosome mutant offspring of
O. Lamarckiana result from @ or o& regular (Lamarckiana) 7 + @ or
o irregular 7. If such were the case, one of these mutants, such as
O. nanella, for example, could reproduce itself only by means of the
union of dissimilar gametes,!® probably of the same types as those which
entered into the original combination. We have designated the
Lamarckiana combinationasabcdefg +a’'b’c' de’ f' 2’ = aa’ bb’ cc’
dd’ ee’ ff’ gg’; then if the mutant nanella resulted from abcde/ g
+a’ b’b’ — de’ f' g’ = aa’ bb'b’ c— dd’ ee’ ff’ ge’, and if the male
and female gametes produced by the mutant were each represented
by the two types of gametes entering into the original combination,
O. Lamarckiana X O. nanella should produce two types of offspring:
abcdefg-+a’ b’b’—d'e'f's' =a’ bb'b’ c— dd’ ee’ ff’ gg’, O. nanella,
aucded D¢.d.¢f ¢ + ab’ cd’ e' f’ 2’ = aa’ bb’ cc’ dd’ ee’ ff’ 22’, O. La-
marckiana. ‘The same results should be secured from the reciprocal
cross. As a matter of fact, de Vries (13, p. 207) has shown that
these are the results obtained from the two crosses; but how shall we ex-
plain the behavior of O. nanella, selfed? It is well known that this
mutant breeds true, while on the basis of our previous assumptions,
we should expect it to produce three types of offspring: (1)a bb—d ef g
+ a’ b’b’—d’ e' f’ g’ = aa’ bbb’b’—dd’ ee’ ff’ ge’, unlike both parents;
tea bode ¢ +a’ b'b’—d' ef’ 2’ = aa’ bb’b’ c— dd’ ee’ fF’ ge’, O.

19 Unless apogamous development were possible.

78 ANNE M. LUTZ

nanella. ‘The reverse combination should also reproduce the parental
type. (3) abcdefg:+a’b' cd’ é f' 2’ = aa’ bb’ cc’ ce’ fF" a5 0. La-
marckiana. ‘The first combination might be excused on the pretext
of incompatibility, but this would hardly be sufficient to account for
the absence of O. Lamarckiana from among the offspring of selfed
nanella. Our difficulties are not lessened, as a little figuring will
show, by assuming that two types of gametes are produced by one
sex, and only one by the other, suchas 9 abb—defgandabcdefg
+ oa’ b'b’—d' e' fs’ or Sa’ b’c' d’e'f' g'; ot by assuming than
all of the female gametes are of one type and all of the male of another
type. Let us then consider the problem from another viewpoint.
We may assume that these irregular 7-chromosome mutant gametes
of O. Lamarckiana, notwithstanding their numerous opportunities to
unite with regular (Lamarckiana) 7-, are incapable of doing so,
because of incompatibility, and that a gamete of this type can unite
only with another of its kind: a bb —defg +a’ b’b’—d'e’'f' g’ = aa’
bbb’b' — dd’ ee’ ff’ gg’, O. nanella; also that the mutant produces
male and female gametes of the same, single type. Our difficulties
would still be with us, for O. nanella X O. Lamarckiana, and the
reciprocal, would result in a new type (the same in both instances)
quite unlike either parent. Furthermore, if the original irregular
7-chromosome mutant gamete produced by O. Lamarckiana were
incapable of uniting with a regular 7-Lamarckiana gamete in O.
Lamarckiana selfed, we would expect the two to be unable to unite
in the crosses, yet we know that seed and offspring are readily secured
from both. Even should we assume that the nanella group aa’ bbb’b’
—dd' ee’ ff’ gg’, resulting from the union of identical gametes, produces
female gametes of one type and male of another, such as a bbb—d—f g
and a’ b’ — d’ e’e’ f’ g’, our difficulties would not disappear. It seems
impossible, on a chromosomal basis, to find an explanation for the
fact that nanella and Lamarckiana, when selfed, produce only nanella
in the first case, and only Lamarckiana in the second (barring rare
exceptions), but that O. nanella * O. Lamarckiana and the reciprocal,
produce both parental types in each case. Truly he who attempts
to explain mutation on a strictly chromosomal basis finds his pathway
beset with many obstacles.

We do not know whether 15-chromosome mutant offspring of 14-
chromosome O. Lamarckiana result from unions of 97 with <8, or
~ 98 with o7, or from both combinations; since there is considerable

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 79

evidence to indicate that functional 8-chromosome cells of one sex
only are produced by certain 15-chromosome individuals, and that
these, in many forms, are female, it is possible that the functional 8-,
perhaps even all functional 7+ -chromosome germ-cells produced by
O. Lamarckiana and certain 14-chromosome mutant derivatives, are
female.” More attention has been directed to the study of male
than female reduction in various forms. Gates (’10, reported in 1907)
and Davis (11) each recorded 6-8 distributions of male heterotypic |
chromosomes in O. Lamarckiana, but as Davis remarks (p. 952) ‘‘we
do not know whether or not fertile pollen-grains may be formed with
chromosomes in a greater or less number than the normal.’’?! Gates
and Miss Thomas (’14) found the expected 7-8 distributions of he-
terotypic chromosomes in the pollen mother cells of O. lata; we know
that 8-chromosome female gametes are produced by this form, yet
we shall see that it appears that very few, if any, 8-chromosome male
gametes, capable of functioning, are formed; nor is the presence of the
two 16-chromosome plants in the C.S.H. culture of O. Lamarckiana
positive proof of the production of both male and female 8-chromosome
germ-cells, since we do not know whether these individuals arose from
8 + 8, or9 +7 unions. It is quite certain, however, that 7 +-chro-
mosome cells of one sex or the other, if not of both, are formed occa-
sionally, since 15-chromosome mutants are quite common.
We may summarize our conclusions, therefore, as follows:

(a) 15-chromosome mutant offspring of 14-chromosome forms are
not invariably distinguished by the somatic characters of QO. lata, O.
semilata or lata-like forms.

(b) Lata-like forms, and those combining certain lata characters
with others not distinctive of O. lata, are not invariably characterized by
I5 chromosomes. :

Thus far we have considered only (a) whether when a meiotic
irregularity in a 14-chromosome form results in the production of a

20 This statement merely expresses a possibility and not the writer’s established
convictions. If 7-+--chromosome male gametes, capable of functioning, are never
produced by O. Lamarckiana, then we must concede that O. gigas de Vries arose in
some one of the various ways suggested by Gates (and recent evidence tends to
strengthen, rather than weaken, Gates’s arguments in support of this conclusion)
and that it was not the product of the union of two 14-chromosome gametes, as
maintained by Stomps and Lutz.

21 By O. Lamarckiana.

80 ANNE M. LUTZ

15-chromosome offspring, such an individual will have the leaves and
habit of O. lata or O. semilaia invariably, or even in the majority of
cases (Gates, ’13) and (b) whether the frequency of the occurrence of
an irregular distribution of the chromosomes of 14-chromosome plants
into 6-8 groups may determine the frequency with which Jata-like
mutants will appear (Gates and Miss Thomas, ’14). We have yet to
consider (c) whether when O. lata is crossed with its 14-chromosome
parent, or is selfed, the percentage in which O. Jata appears among the
offspring is indicative of the number of 9- and 7-chromosome germ-
cells which function (Gates and Miss Thomas). This question will
be treated under the following head.

3. Are 15-chromosome Forms Inconstant?

Of the twenty-one 15-chromosome mutants which Gates and Miss
Thomas reported, three were identified as O. semilata, one as lata to
semilata and two as semilata to lata. Referring to de Vries’s cultures
of O. lata and O. semilata they say (p. 527): ‘‘Oe. lata was classed by
him as an inconstant species, but semtlata was incorrectly classed as
constant. They are both obviously inconstant, however, and the
presence of the odd chromosome shows why this must be so.” (Italics not
employed in the original.) Gates (’15a, pp. 111-112) has since found
that the mutant which he described as semizlata is not the same as de
Vries’s mutant of this name, but has decided to retain the name for
the form reported by Gates and Miss Thomas, since the Amsterdam
type is extinct. Therefore, when it becomes necessary to distinguish
between these two types, we shall designate them as O. semuilata de
Vries and O. semilata Gates, respectively.

Gates and Miss Thomas’s statement raises the question, Does the
presence of the odd or extra chromosome necessarily render a form
inconstant? Are 15-chromosome forms never constant?

Since 15-chromosome forms produce, as a rule, no pollen, very
little, or a moderate amount containing a high percentage of bad
grains; since seeds are obtained from selfed forms with difficulty,
and when secured, usually a much lower percentage of these than of
the seeds derived from 14-chromosome forms succeed in germinating
in the short time commonly allowed them, their constancy has not
been tested on an extensive scale. Inasmuch as we know that in

2 Plants having more than 14, but fewer than 28, chromosomes are much more
inclined to be male- than female-sterile. Just why this is so, is not yet clear.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 8I

certain cases but one small culture of offspring from these plants, selfed,
has been grown, and have reason to believe that only one or two, con-
taining but very few offspring, have been grown from others, and since
it is probable that the offspring derived from these forms, as recorded
by various workers, do not represent the whole of the possible progeny
in any case, 2. e., do not represent all that would have been obtained
had means been employed to secure the germination of every seed
capable of germinating—it is clear that the evidence upon which we
would like to base our conclusions is not wholly reliable. This fact
should be borne in mind throughout the discussions which follow.
Nevertheless, the evidence as it now stands is not devoid of signif-
icance.

O. lata (15) produces, as we shall see, O. Lamarckiana, O. lata® and
a certain percentage of mutants,—the number of Lamarckianas
greatly exceeding the number of latas in the cultures grown by Mac-
Dougal and de Vries. The behavior of sublinearts, if a 15-chromosome
form, appears to be similar to that of O. lata, since the 31 offspring
which de Vries obtained from sublinearis, selfed (09, Vol. I., p. 401)
were classified by him as follows: 19 Lamarckiana, 3 sublinearis,
I lata, 1 nanella, 1 albida, 3 subovata, 2 oblonga and 1 gigas.24 Here
again we see that the number of Lamarckianas greatly exceeded the
number of forms which reproduced the characters of the mutant
parent. O. bipartita (15), selfed, produced O. Lamarckiana, O. bi-
partita, a few forms resembling the parent in most ways but having
fewer flowers with cleft petals and extra-lobed stigmas than is common,
and a few mutants. Here, also, a higher percentage of Lamarckianas
than of bipartitas was obtained in the time allowed for the germination
of the seeds, probably about 4 months.

As previously stated, the number of chromosomes present in O.
scintillans is unknown, but it is probable that itis15. This form, when
selfed, according to de Vries ('13, p. 257), produces a variable number
of scinitllans; sometimes 35-40 percent or less and again as high as

*3 Bartlett (’15a, p. 103) calls attention to the similarity in the behavior of O. lata
and O. stenomeres mut. lasitopetala. From the latter form, selfed, he obtained 60
percent stenomeres and 40 percent laszopetala. Mr. Arzberger has counted 14
chromosomes in O. stenomeres; the chromosome number of O. lasiopetala has not yet
been announced, but it is probably 15.

tas De Vries has since concluded (’12, p. 34) that this plant and the other identi-
fied as O. gigas in 1899 (’09, Vol. I., p. 327) were probably triploid, and not tetraploid,
forms.

82 ANNE MJ LULZ

70-80 percent of the total number of offspring reproduce the char-
acters of the mutant parent. The remainder are, for the most part,
O. Lamarckiana, but with a considerable number of O. oblonga (“oft
bis 20%’’) and a few other mutants.

O. semilata Gates (15), is an inconstant form, as Gates and Miss
Thomas (’14, p. 532) and Gates (’I5a, pp. 114-115) have shown, pro-
ducing O. Lamarckiana, O. semilata Gates, a few O. lata which may
be classed as mutants and (p. 114) others “forming a continuous
series running to Lamarckiana.”

O. elliptica, having 15(?), chromosomes, reverts almost entirely to
Lamarckiana, according to de Vries (09, Vol. I., pp. 397-398). From
one 1895 mutant, selfed, he obtained ‘‘some hundred of seedlings,”’
all of which proved to be ordinary Lamarckiana. From a second
mutant of the same year 500 offspring were secured, I of which was
elluptica, and the remainder Lamarckiana. A third 1895 mutant
“gave rise to 27 seedlings not one of which was an elliptica.”” From
an 1896 mutant he obtained 32 offspring, 5 of which were elliptica and
the remainder Lamarckiana; from an 1899 mutant he secured about
100 offspring, al! of which were O. Lamarckiana.

O. lata rubricalyx, in which Gates and Miss Thomas counted
15 chromosomes, when selfed, according to Gates (15a, p.: 288),
produced a nearly uniform lot of offspring (44 plants), ‘‘ all having the
red pigmentation of rubricalyx, but were intermediate between rubri-
calyx and grandiflora in foliage and buds. ... The plants which
were examined had 14 chromosomes, as was doubtless the case with
all of them.” No lata rubricalyx plants were found among the off-
spring.

While all of the above forms are clearly inconstant, de Vries’s
researches indicate that a 15-chromosome form may breed perfectly
true. He selected 5 biennial albida plants (’o09, Vol. I, p. 229) in
1897 and grew a second generation consisting of 86 individuals in 1898
and a third, consisting of 36, in 1899. ‘‘Both generations,’ he adds,
‘““were absolutely constant and exhibited no signs of reversion.’’

If O. oblonga be a 15-chromosome form, it indicates even more
strongly (because of the larger number of offspring obtained) that a
15-chromosome form may be constant. During a period of over 13
years, de Vries (pp. 346-348; also, ’13, p. 315) selfed a number of
oblonga mutants and obtained a total of 2,919 offspring, all of which,
with the exception of 11 mutants (7 rubrinervis, 3 albida, 1 elliptica)
were oblonga.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 83

Certain somatic characters of many mutant offspring of O. La-
marckiana X O, Lamarckiana and of O. Lamarckiana, selfed, indicate
that a very large percentage of the mutant offspring of O. Lamarckiana
have 15 chromosomes and that a larger number of 15- than of 14-chromo-
some mutant offspring are produced by this form. Not only does there
appear to be a larger number of distinct types of 15- than of 14-chromosome
mutants, but a higher percentage of 15- than of 14-chromosome mutant
indiiduals produced by O. Lamarckiana. Many of the mutant off-
spring of O. Lamarckiana never have been brought to flower; further-
more, new forms are appearing each year. It will be necessary to
determine the somatic chromosome numbers of a large percentage of
the mutant types produced by O. Lamarckiana, to bring the 15-chro-
mosome forms to flower, to self flowers on all parts of the plants, to
adopt methods which will secure the germination of all viable seeds,
and to grow large numbers of offspring,—in order to ascertain whether
15-chromosome forms are more commonly inconstant than constant.
The majority of the 15-chromosome forms whose constancy we have
considered have produced very few offspring, yet we may safely assert
that the evidence available at present indicates that most 14-chro-
mosome forms are constant and most 15-chromosome forms inconstant.
Furthermore, although our present knowledge of the behavior of
14+-chromosome forms is very limited, largely owing to the infre-
quency with which good pollen is produced by such forms, it may be
stated that the evidence available at present indicates that tnconstancy
1s commonly associated with the 14+ -chromosome condition. It seems,
however, that forms having twice 14 chromosomes are more likely to
be constant—in the same sense that O. gigas de Vries is constant—
than those having more than 14, but fewer than 28, chromosomes.

4. Factors Determining the Constancy or Inconstancy of 15-Chromosome
Forms.

De Vries obtained the same results from O. scintillans selfed, as
from O. scintillans X O. Lamarckiana (pp. 257-262); also the same re-
sults from selfed hybrid lata, descended through O. lata X O. La-
marckiana from O. lata & O. semilata, as from O. lata X O. Lamarckiana
(og, Vol. I., pp. 240, 360; '13, pp. 244-257). This led him to conclude
that female gametes of O. scintillans and this hybrid lata do not bear
the same hereditary characters as the male gametes of these forms;
that the characters of the mutant, in the first case, and of the hybrid

84 ANNE M. LUTZ

(which are the same as those of mutant O. Lamarckiana lata), in the
second, are transferred to the offspring through the egg cells, and
not through the pollen; that the pollen, in each case, behaves precisely
as the pollen of pure Lamarckiana (pp. 257, 258, 262, 272, 273, 323).

At the time of the publication of “‘Gruppenweise Artbildung’’, O. lata
was popularly supposed to be the only 15-chromosome mutant pro-
duced by O. Lamarckiana or other forms, hence de Vries has discussed
these very important results without reference to the chromosome
numbers of the plants in question. However, since a number of forms
are now known to have 15 chromosomes, their behavior may be
further considered in the light of this fact.

Bartlett (‘I5a, p. 103), discussing the behavior of selfed O. lata
and O. stenomeres mut. lasiopetala, concludes, in agreement with de
Vries, that “it appears that the good pollen grains of Oe. lata are genet-
ically the same as those of Oe. Lamarckiana, and do not carry the
lata-characters.”’ ‘‘Thus,’’ he states, “‘it appears that there is a class
of mutations of which the eggs are of two kinds; one kind carries the
characters of the parent species, the other kind the characters of the
mutation. The pollen grains, however, appear to be of one kind only,
and to carry the characters of the parent species.’’ He says ‘“‘we
must assume that the male 8-gametes are eliminated’’ and asks if
it is not possible ‘that the male gametes which carry the characters
of the mutation are eliminated because of some physiological defect?”’
‘Oe. lata,’’ he states, ‘produces two classes of gametes, with 8 and
7 chromosomes, respectively. If two 7-gametes fuse, we have Oe.
Lamarckiana; if a 7-gamete (presumably male) fuses with an 8-gamete
(presumably always female) we have Oe. lata.’ We shall see that
the evidence indicates such are the usual, though not the invariable,
results.

MacDougal (’07) obtained 94 offspring from selfed O. lata which
were identified as follows: 10 O. lata, 80 O. Lamarckiana, 1 O. albida
and 3 O. oblonga. Albida has 15 chromosomes and oblonga 14 or 15.
De Vries (13, p. 256) obtained 442 offspring from a hybrid Jata,
selfed, 33 percent of which were O. lata and 4 percent mutants. The
remainder were, doubtless, O. Lamarckiana. Itis probable that several
types of 15-, and one or more 14-chromosome forms were included
among these 17 or 18 mutants.

In 1908 3 mutant lata offspring of O. Lamarckiana were selfed at
Cold Spring Harbor. A total of 360 seeds were obtained and these

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 85

were planted at spaced intervals in pans of sterilized soil, December
II of the same year. 129 germinations resulted; 2 plants died uniden-
tified as seedlings and a third, identified as O. Lamarckiana, died later.
The remaining 126 were transferred to the garden May 12, 1909.
One /ata from each of the three 1909 cultures was then selfed and the
259 seeds obtained were sown in the same manner as those of the pre-
ceding season, March 7, 1910. 99 germinations resulted; all of the
seedlings survived and all of the young plants were transferred to the
experimental garden May 16, following.

Of these 226 plants, 109 (approximately 50 percent) were clearly
identified as O. lata, 8 as O. lata (?), 57 as O. Lamarckiana (approxi-
mately 25 percent) and 4 as O. Lamarckiana (?). The chromosome
numbers of the plants in the second and fourth groups are unknown.
In addition to the foregoing there were 7 distinct types of 15-chromo-
some mutants (23 individuals) which could not be classified either as
O. lata or as laia-like forms. Still other mutant types, whose chro-
mosome numbers were unknown, were believed to be 15-chromosome
forms. In addition to the 57 Lamarckianas there were 3 types (4
individuals) of 14-chromosome mutants, quite unlike O. Lamarckiana.
These are believed to represent approximately the total number of
14-chromosome forms produced. 196 of the 226 plants grown rep-
resented types whose chromosome numbers are now known, and but
one of the 196 had 16 chromosomes.” De Vries’s mutants were not
classified, but it is quite clear that no one of the 94 offspring which
MacDougal obtained from selfed Jata had 16 chromosomes. We do
not know how many of MacDougal’s and de Vries’s seeds failed to
germinate, but we have seen that 63 percent of the Cold Spring Harbor
seeds sown failed to germinate in the few months allowed them, hence
we do not know what would have been the relative percentages of
I4-, I5- and 16-chromosome forms, had all the viable seeds sown
germinated.

In connection with these studies of selfed latas, the results obtained
from crossing one of these 1908 mutants with O. Lamarckiana will be
of interest. In 1908 I pollinated O. lata, mutant No. 3500, with
O. Lamarckiana, No. 3814, and covered the stigmas of the latter plant

25 In addition to the 14-, 15- and 16-chromosome offspring referred to, one 2I-
and one 22-chromosome mutant were produced, as previously reported (Lutz, ’12).
The 226 offspring of these six selfed latas will be carefully tabulated and fully de-
scribed in a later report.

86 ANNE M. LUTZ

with small quantities of pollen obtained from the former. 15 chro-
mosomes were counted in the somatic cells of No. 3500 and 14 in those
of No. 3814. 320 seeds from O. lata * O. Lamarckiana were planted
at spaced intervals in seed pans, December 12, 1908; 49 of this number
germinated previous to the time of transplanting in May. Four of
the young plants died as seedlings and the remainder were classified

as follows:
TABLE II

O. lata No. 3500 X O. Lamarckiana No. 3814

. Lam- O. aber- ;
pee ote: ae O. lata O. albida | Type 5432 Totals
14 chromosomes...... 15 ney awed Sent A Py 15
1A be cutee dla Sout rees I Cae ipa a 2 I
15 i ee acne es oe By. I eee 28
: Gite aaa cree pices eee oe Ar =| I I

On December 11, 1908, the same number of seeds from the second
cross (O. Lamarckiana X O. lata) were planted in the same manner
as the above. Only 18 germinated; therefore, on February I, 1909,
119 seeds from the same capsules as the preceding were planted and
58 seedlings obtained previous to the middle of May. The 76 plants
derived from this cross were classified as follows:

TABLE III
O. Lamarckiana No. 3814 X O. lata No. 3500

OL es Obie

. Lam- modifie . bipar-

arckiana | QO. nanella| ,ybyiner- O. lata tita (7) 26 Totals
vis?)

14 chromosomes...... 63 I 9 spina 2 ae 73

15 s IS ee aoe pee i I 2 2

From these tables we see that in the time allowed for germination,
almost twice as many I5- as 14-chromosome offspring were derived
from O. lata X O. Lamarckiana, while only 3, possibly only 1, of the
76 plants derived from O. Lamarckiana X O. lata had 15 chromosomes.
Since one or more 15-chromosome mutants usually are found in
Lamarckiana cultures of this size, it is probable that the 15-chromosome
offspring of O. Lamarckiana X O. lata resulted from 98 + 07 and

*6 The identification of these supposed bipartitas was based upon the characters
of the greenhouse rosettes, as the plants were not transferred to the garden.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 87

possible (providing 8-chromosome gametes are formed by O. Lamarc-
kiana) that one or two of those derived from the reciprocal cross were
products of 27 + o'8 unions. Notwithstanding the fact that 85
percent of the /ata X Lamarckiana, and 83 percent of the Lamarckiana
x lata, seeds failed to germinate in the time allowed them—approxi-
mately the same percentage of failures in both cases—15-chromosome
forms appeared in considerable numbers among the offspring of the
first cross, but were almost entirely absent from the second culture.
However, we do not know what the percentages of 14- and 15-chro-
mosome offspring would have been had all the seeds of each cross
germinated.

Bearing in mind that we do not know ‘‘what may be the changes
of front when exact data become available,’ it may be said that the
C.S.H. cultures of mutant O. Lamarckiana lata selfed and crossed both
ways with O. Lamarckiana confirm, in the main, the earlier statements
by de Vries and Bartlett and point to the following conclusions re-
garding this mutant: (a) Lata characters are transmitted through a
portion of the egg cells, and not, except possibly in rare instances,
through the pollen of thisform. (0) 8- and 7-chromosome female gam-
etes, capable of functioning, but as a rule, only 7-chromosome male
gametes, capable of functioning, are produced by O. lata.27 The
majority of the 8-chromosome female gametes (probably not all) are
bearers of Jata characters, while the majority of the male and female
7-chromosome gametes (probably not all) are bearers of Lamarckiana
characters. (c) It now appears that when offspring result from 8 + 7
and 7 +7 unions, the majority of the former have Jata, or lata-like
characters, and the majority of the latter Lamarckiana characters,
but, as in the case of O. Lamarckiana, selfed, it is not safe to assume
that such are the invariable results, since it is probable that at least
a portion of the 15- and 14-chromosome offspring derived from O. lata
selfed, O. Lamarckiana, selfed, and O. lata X O. Lamarckiana which
cannot be classified as O. lata and O. Lamarckiana, are products of
8 +7 and 7 +7 unions, respectively. As in the case of O. La-
marckiana, O. albida is one of the common mutant offspring of O. lata,
selfed, and of O. lata X O. Lamarckiana and it seems quite probable
that this form results from 8 + 7 unions.

*7 Fourteen- and fifteen-chromosome mutants, particularly the latter, are found
in practically all fair-sized cultures of O. lata X O. Lamarckiana. No statement can

be made concerning the appearance of these forms in cultures of the reciprocal cross,
since only one has been reported thus far.

88 ANNE M. LUTZ

While, as we have seen, de Vries has shown that O. scintillans, when
selfed or pollinated by O. Lamarckiana, behaves in much the same way
as O. lata under similar conditions, statements concerning the chro-
mosomal combinations resulting from these operations must be
wholly speculative since the numbers of chromosomes present in
QO. scintillans and its oblonga offspring are unknown. However, if
scintillans has 15, as is probable, there is much evidence to indicate
that 7- and 8-chromosome female gametes, capable of functioning,
and, as a rule, only 7-chromosome male gametes, capable of function-
ing, are produced by this form,?® whether oblonga has 14 or 15 chro-
mosomes. ‘That only 7-chromosome male gametes are produced which
behave in every way like the 7-chromosome male gametes of O. La-
marckiana is clearly indicated by the fact: that while O. scintillans,
selfed, and O. scintillans X O. Lamarckiana produce O. scintillans,
O. Lamarckiana and O. oblonga, O. Lamarckiana X O. scintillans
yields 100 percent O. Lamarckiana (de Vries, 13). A noteworthy
peculiarity in the behavior of O. scintillans is the relatively large
number of offspring of one type derived from the mutant, selfed, and
pollinated by O. Lamarckiana, which display neither the characters
of O. scintillans nor of O. Lamarckiana, but of the mutant, O. oblonga.
This indicates either that a relatively large percentage of the offspring
resulting from 8 + 7 unions fail to reproduce the characters of the
mutant parent, or that a relatively large percentage of those derived
from 7 + 7 unions fail to display the characters of O. Lamarckiana.

The behavior of O. bipartita and of O. sublinearis (if the latter has
15 chromosomes, and it is probable that it has) indicates that all, or
nearly all, of the gametes of one sex which are capable of functioning,
contain 7 chromosomes, while a portion of those of the other sex
contain 7 and the remainder 8. The same may be said of O. semilata
Gates if the offspring which Gates refers to as “‘forming a continuous

28 In the case of O. Jafa and other forms to be discussed in this report, it will be
understood that the writer does not exclude the possibility of other gametes being
formed occasionally in addition to those enumerated—gametes having fewer than
7, or more than 7 or 8chromosomes. For instance, as earlier stated, it is conceivable
that 6-chromosome gametes may function in union with 8- or 8+-. It is possible
that 9-chromosome gametes, capable of functioning, may be produced occasionally
by 14- and 15-chromosome forms, particularly the latter, and we know that there is
much evidence to show that 14-chromosome gametes are sometimes produced by 14-
chromosome forms (possibly also by 15-) and 15-chromosome gametes by 15-chro-
mosome forms.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 89

series running to Lamarckiana”’ have either 14 or 15 chromo-
somes.”

O. stenomeres mut. lasiopetala, as Bartlett (’15a) has pointed out,
resembles O. Lamarckiana mut. lata in its behavior, with, of course,
this exception: the 14-chromosome offspring of the former bear the
characters of O. stenomeres, and not of O. Lamarckiana. In each case,
however, the offspring bear the characters of the 14-chromosome form
which produced their mutant parent. Furthermore, as in the case of
O. lata, a portion of the offspring (presumably having 15 chromosomes)
reproduce the characters of the mutant parent.

The behavior of O. elliptica, if a 15-chromosome form, indicates
that all of the gametes of one sex, capable of functioning, have 7 chro-
mosomes, while the majority of those of the other sex which are
capable of functioning have 7 chromosomes, and only a very few, 8.

The behavior of O. nanella lata indicates that only 7-chromosome
female gametes are produced by this form since de Vries (’09, Vol. f.,
p. 374) found that O. nanella lata X O. nanella “gave rise to ordinary
nanella only.”’ From O. lata * O. Lamarckiana de Vries ('13, p. 257)
obtained two dwarfs through mutation. ‘‘Dereine hatte nebenbei die
Merkmale der Lamarckiana, der andere diejenigen der Lata. Beide
hatten Pollen, und wurden damit rein befruchtet. . . . Die letztere
gab zwar auch nur Zwerge . . . , spaltete sich aber in bezug auf die
Lata-Merkmale in 9 Lata-Zwerge idl 18 gewohnliche Zwerge, a
The latter, upon self-fertilization, proved constant, but the We
dwarfs behaved in the same manner as the parent, when selfed.®® If
this lata dwarf is the same form as de Vries’s O. nanella lata, then the
available evidence indicates that all of the female cells of this mutant
which are capable of functioning contain 7 chromosomes, while the
majority of the male contain 7 and a smaller number 8.*!_ Little or no
consideration should be given to this evidence, however, since we do
not know that the two mutants combining nanella-lata characters
were duplicate types; furthermore, we have no assurance that a suf-
ficient number of offspring of O. nanella lata X O. nanella were grown

29 This condition was earlier indicated by Bartlett (15a, p. 103) in the statement
that ‘‘ Oe. scintillans acts like Oe. lata in every way.”’

30 QC. rubrinervis lata, which appeared in Schouten’s 1906 culture of O. rubrinervis
(Schouten, ’08), suggests a 15-chromosome condition, as Gates (’15) has stated, but
since the chromosome number of this plant is unknown and that of the parental
type unestablished, the behavior of this mutant will not be discussed at present.

31 The behavior of O. oblonga, if a 15-chromosome form, is somewhat contra-
dictory and will not be discussed here.

gO ANNE M. LUTZ

to prove that O. nanella only, invariably results from this cross; and
finally, since only 27 offspring were derived from the later mutant,
selfed, it is clear that we are not justified in formulating definite
conclusions concerning its behavior.

The behavior of O. lata rubricalyx is somewhat unique. If all of
the offspring of the mutant, selfed, had 14 chromosomes, as Gates
thinks probable, this fact would indicate that all of the male and
female cells produced which were capable of functioning, had 7 chro-
mosomes. This is further indicated by the fact that although Gates
crossed the mutant both ways with several other forms, ‘the offspring
(few in number) which developed proved to be all of 14-chromosome
types.’ He does not state whether the several forms employed in
these crosses were 14- or 15-chromosome plants, but the results in-
dicate that only 97 + o’7 unions occurred in every case, or at least
that only seeds resulting from these unions germinated in the time
allowed them. Of especial interest is the fact that the 14-chromosome
offspring of this plant were intermediate between rubricalyx and grandt-
flora (the grandparents). Would the mutant behave differently if
produced by rubricalyx, selfed? Is this precise type ever produced
by rubricalyx?

In the case of O. albida, all of the cells of one sex appear to contain
7 chromosomes and all of the other 8, since de Vries obtained albida
offspring only from this form, selfed.

Thus, in addition to O. lata, the records show that a certain number
—commonly more than half—of the offspring derived from selfed
O. bipartita, O. scintillans, O. sublinearis and almost all of those ob-
tained from selfed elliptica (the evidence is not clear in regard to O.
semilata Gates) were 14-chromosome plants and that all of them, or
all but a few mutants (providing the oblonga offspring of O. scintillans
had 15 chromosomes), were O. Lamarckiana.* Hence the evidence
as it now stands indicates that all but relatively few of the 7-chromo-
some male and female gametes of these plants (providing scintillans
and sublinearis are 15-chromosome forms) are bearers of Lamarckiana
characters® and that the mutant characters of the 15-chromosome

8? It is probable that the list of forms which produce large numbers of O. Lam-
arckiana when selfed, might be extended to include many other direct and indirect
15-chromosome mutant derivatives of O. Lamarckiana, but not all; O. nanella lata
and O. rubrinervis lata may be suggested, among others, as probable exceptions.

88 The evidence indicates that all of the 7-chromosome gametes of one sex pro-
duced by O. elliptica and all but comparatively few of the opposite sex, which are

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS OI

parent in each case are transmitted to the offspring through the 8-
chromosome gametes. In cases with which we are familiar, the 8-
chromosome cells appear to be produced wholly, or almost wholly,
by one of the two sexes; this is clearly female in certain cases, but we
may find that these cells are male in others.34 Without considering
the question of whether these forms, when given off as Lamarckiana
mutants are ‘“‘half mutants”’ (de Vries), it is clear that a Lamarckiana
offspring is produced by selfed 15-chromosome mutant derivatives of
O. Lamarckiana when two gametes, each bearing Lamarckiana char-
acters, and, possibly, the chromosomal combination peculiar to these
gametes, unite and produce a seed capable of germinating. Bartlett
(15a), after stating that certain mutants appear to produce two kinds
of eggs, one carrying the characters of the mutant producing them and
the other the characters of the parent species, while the pollen grains
appear to be of one kind only and to carry the characters of the parent
species, adds: “If so, Oe. lata might be supposed to originate by the
union of a Jata-egg (itself constituting the true mutation) with a
Lamarckiana-sperm.” Accepting Bartlett’s suggestion, it may be
said of all 15-chromosome mutant derivatives of O. Lamarckiana
which, when selfed, produce two types of offspring, one duplicating
the characters of the mutant parent and the other those of O. Lamarcki-
ana, that an offspring of the first type is obtained from these forms,
selfed, when an 8-chromosome gamete bearing the characters of the
mutant parent, and, possibly, the chromosomal combination peculiar
to these gametes, unites with a 7-chromosome gamete bearing La-
marckiana characters, and possibly, the chromosomal combination
peculiar to Lamarckiana gametes, and producesa seed capable of ger-
minating. If we consider this matter on a strictly chromosomal
basis, we will concede, of course, that 9 + 6, or any other union which
produces the same chromosomal combination as the ordinary 8 + 7,
might result in an offspring which would duplicate the vegetative char-
acters of the 15-chromosome parent. However, since 14-chromosome
offspring of 15-chromosome plants and 15-chromosome offspring of I4-
capable of functioning, are bearers of Lamarckiana characters. It is probable that
7-chromosome Lamarckiana gametes (of one sex) are produced by O. albida, but ex-
perimental evidence has not indicated the facts in the case.

34 Doubtless these assertions, and the conclusions regarding O. Jata under the
heads of (a), (0) and (c) on page 87 would be equally applicable to many other

15-chromosome mutant derivatives of O. Lamarckiana, if the names of these mutants»
in the latter case, were substituted for that of O. lata.

92 ANNE M. LUTZ

chromosome plants never duplicate the vegetative characters of their
parents, it is clear that 8 + 8 (lata § + lata 8, for example) could not
produce an individual having the same vegetative characters as the 15-
chromosome parent.® :

In view of the fact that when certain 15-chromosome forms are
selfed, the parental mutant type appears to be reproduced by 8 + 7,
and never by 8 + 8, unions, one may ask whether we are not justified
in asserting that the 7-chromosome cells are as truly transmitters of
the mutant characters as the 8-; it seems that they are not, for the
22-chromosome offspring of O. lata X O. gigas, which presumably
result from @8 + oI4 unions, suggest lata-gigas characters, while
21-chromosome hybrids, which doubtless result from 97 + G14
unions, suggest Lamarckiana-gigas combinations. Since the latter
bear no trace of Jata characters, it is clear that these are transmitted
through the 8-, and not through the 7-chromosome gametes.

If 7- and 8-chromosome male and female gametes were produced
in equal numbers, all capable of functioning in union with 7- and 8-
chromosome cells of the opposite sex, we would expect a selfed 15-
chromosome form to produce 14-, I15- and 16-chromosome offspring in
the ratio of 1:2:1. How, then, shall we explain the fact that the
number of 14-chromosome offspring produced usually (not invariably,
as we have seen) exceeds the number of I5-, while 16-chromosome
forms are almost unknown in such cultures? How shall we explain
the fact that F; cultures derived from 15-chromosome forms pollinated
by 14- usually contain many more 14- than 15-chromosome plants ?*

1. In the first place, it will be recalled that no evidence has been
brought forward to show that 8-chromosome male gametes, capable

35 It is quite probable, however, that Jata-like mutants may result from /ata 8+
lata 8.

36 After the manuscript for this report had left the writer’s hands, an important
contribution from de Vries appeared, entitled ‘‘ New dimorphic mutants of the
Oenotheras’’ (Bot. Gaz. 62: 249-280, Oct., 1916). In this report de Vries has shown
that O. cana, which is probably a 15-chromosome form, produces O. cana and O.
Lamarckiana when selfed, but that a larger percentage of the offspring duplicate the
characters of the mutant parent when a biennial, than when an annual, plant is em-
ployed. This fact demonstrates, as he states, that the behavior of O. cana is largely
dependent upon the vigor of the individual employed. De Vries believes that this
is true of other dimorphic mutants, since he had earlier demonstrated this difference
in the behavior of annual and biennal scintillans. The bearing of these important
facts upon the statements which are included under the heads of 1, 2 and 3 above,
will be discussed in a later publication.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 93

of functioning, are produced by 15-chromosome forms. So far as has
been observed, the male gametes appear to be, with rare exceptions,
exclusively 7-. It is possible that the embryo-sac is more frequently
differentiated from a 7- than from an 8-chromosome group, and that
more 7- than 8-chromosome eggs, capable of functioning, are produced.*?

2. Since many 14+ -chromosome forms produce both 7- and 7+-
chromosome female gametes (?/ata 7 and 8), but are usually male-
sterile, hence produce, ordinarily, neither 7- nor 7 +-chromosome male
gametes, our knowledge of controlling factors is still too limited to
warrant the suggestion that 7+-chromosome groups are more likely
than 7- to fail to produce male gametes capable of functioning.®® It is
probable, however, that 7-chromosome gametes unite with 7- and
produce seeds capable of germinating more readily than any other
combination. ‘There is also considerable evidence to indicate that,
as a rule, gametes unite and produce viable seeds more readily (a)
when one gamete contains 7, and the other a multiple of 7, chromo-
somes®® (7 +14; 21-chromosome Lamarckiana-gigas offspring of
O. lata X O. gigas, O. Lamarckiana X O. gigas, etc.); (b) when both
gametes contain a multiple of 7 chromosomes (14 + 14; 28-chromo-
some offspring of O. gigas, selfed, than (c) when each contains more than
7, but fewer than 14, chromosomes.” Plants having more than 14,

37 Gates and Miss Thomas (’14) have stated that 8-chromosome megaspores
of O. lata evidently have fewer prospects of functioning than 7-chromosome mega-
spores, since the percentage of latas derived from /ataXLamarckiana usually falls
below 20 percent, and sometimes to 4 percent.

38 We should not overlook the fact that in absolutely male-sterile 15-chromosome
forms, 7-chromosome male gametes are eliminated as completely as the 7+-.

39 There is also some evidence to indicate that 8- unites with 14- and produces
seed capable of germinating more readily than two gametes, both of which contain
more than 7, but fewer than 14, chromosomes (22-chromosome offspring of O. lata
XO. gigas). We have no evidence whatever on which to base conclusions regarding
7 +13. Whether 8 + 14 combinations are less likely to occur and produce viable
seed than 7 + 14, cannot be stated. The culture of 133 offspring which de Vries
(13, p. 186) obtained from O, lata X O. gigas in 1907 consisted of 65 Lamarckiana-
gigas (presumably 7 + 14) and 68 Jata-gigas (presumably 8 + 14), offspring.

40 The writer’s experience with cultures of 14- and 15-chromosome forms pol-
linated by 28- has not been sufficiently extensive to justify the assertion that seed
and offspring are less readily secured from these than from 28- pollinated by 28-,
but numerous attempts were made to pollinate 28-chromosome O. gigas de Vries
with 14-and 15-chromosome forms, invariably with the same result; only flat seeds,
or seed-like structures were secured, and these, of course, were utterly incapable
of germinating. De Vries and Davis, however, have each grown offspring of O. gigas
x O. Lamarckiana, though Davis’s culture (’10) was quite small,—1r2 plants.

94 ANNE M. LUTZ

but fewer than 28, chromosomes commonly produce no pollen, very
little, or at most, only a moderate amount; furthermore, one ordi-
narily finds that very few of the grains produced are normal appearing,
hence the majority are probably incapable of functioning. On the
other hand, 14- and 28-chromosome plants not only produce larger
quantities of pollen, as a rule, but usually a much larger percentage of
the grains produced are normal appearing. In the case of 14-chro-
mosome forms, usually about 60-70 good appearing grains per 100
(sometimes fewer, sometimes more) are found in the early and mid-
season buds. These factors are undoubtedly primarily responsible for
the difficulty commonly experienced in obtaining seeds from forms
having (14-+-28)*! chromosomes, as compared with the relative ease
with which they are ordinarily secured from 14-chromosome forms,
selfed, or even from selfed 28-chromosome plants; but it does not tell
us why seeds of 21-chromosome forms, when secured, germinate
much less readily than seeds of 14-chromosome plants and less readily
than those of 28-chromosome O. gigas de Vries.

Gates (15a, p. 194) says: “It is clear that triploidy leads to the
production of many new chromosome-numbers, through the irregu-
larities it introduces into the meiotic phenomena. . . . It is at present
unknown whether the number alone determines the viability, or
whether particular chromosome combinations will, owing to incom-
patibility, fail to produce an,embryo after fertilization.’’ Elsewhere
(p. 234) he speaks of the difficulty experienced in making “giant
crosses”’ (doubtless referring to crosses of 28- with 14- and 15-chromo-
some forms) and says: ‘‘ This is undoubtedly a result of the unbalanced
chromosome numbers and the meiotic irregularities to which they
lead: 2. Sigs

The writer has long thought it probable that the incompatibility
of certain combinations, particularly such as those brought about by
selfing triploid forms, is partially responsible, not only for the small
number of seeds produced, but for the relatively small number of seeds
which germinate. Doubtless many of the heterotypic distributions
in triploid forms (20- and 22-, as well as 21-chromosome individuals)
are irregular, resulting in the production of daughter groups with
fewer than 7, more than 7, regular 7, or, possibly, an irregular assort-
ment of 7 chromosomes. Even such combinations as 10 + 10,10 + II
and If +11 in selfed 21-chromosome forms may be less compatible

41 More than 14, but fewer than 28.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 95

than is apparent, for two gametes having the same number of chro-
mosomes, or nearly so, may be represented by entirely different chro-
mosomal combinations. Yet Gates’s assertion that “‘it is at present
unknown whether the number alone determines the viability, or
whether particular chromosome combinations wil, owing to incom-
patibility, fail to produce an embry» afte: fertilizarion’’ voices a
quesiion which is still unanswered. !f we believe that each of the
14-chromosome mutant types produced directly by O. Lamarckiana
is represented by a chromosomal combination differing from that of
the parent, then in crossing these various types one would expect to
find some of the 7 + 7 combinations resulting less compatible than
that of O. Lamarckiana, selfed. All 14-chromosome forms which have
been tested by the writer have been found to produce an. abundance
of seed when selfed or crossed with other 14-chromosome forms; further-
more, relatively larger percentages of these seeds (compared with those
derived from forms having more than 14, but fewer than 28, chromo-
somes) were found to be capable of germinating. No evidences of
incompatibility have been observed thus far; indeed fertility appears
to be associated with the 14-chromosome condition wherever found.

Ordinarily only a small percentage of the few seeds derived from
selfed 21-chromosome forms are capable of germinating in the few
months commonly allowed them in seed pans. De Vries (’09, Vol. I,
p. 261) has shown that Lamarckiana seeds may lie in the ground for
two or more years before germinating, and the writer has recently
verified these results. Ifso many 14-chromosome seeds may germinate
late one may be led to inquire whether certain 14-+-chromosome
seeds may not be even more inclined to germinate tardily. In the
opinion of the writer, the lower percentage of germination commonly
exhibited by seeds of most 14 +-chromosome forms is due to the total
inability of many 14-++-chromosome seeds to germinate, rather than
to a greatly increased tendency on the part of these seeds to germinate
late. It is furthermore probable that, ordinarily, a larger percentage
of the products of selfed 14 +-, than of selfed 14-chromosome forms are
merely seed-like structures, and therefore incapable of germinating.
Particularly is this probably true of the products of selfed triploid
individuals, for one frequently finds that a large percentage of the few
seeds obtained are small, flat and unpromising appearing.

Regarding the relative percentages of 14-, 15- and 16-chromosome
offspring derived from selfed 15-chromosome forms, it is possible that

96 ANNE M. LUTZ

gametic incompatibility is partly responsible for the usual production
of fewer 15- than 14-chromosome offspring by 15-chromosome forms
pollinated by 14-, and by selfed 15-chromosome plants. Lata 8+
Lamarckiana 7 is plainly an unbalanced combination, but lata 8 +lata
8, assuming that the two gametes combined duplicate chromosomes
(aabcdefg +a’a'b'c'd' e’ f' g’), could be designated as a balanced
combination; would these two gametes be compatible? More so
than lata 8 + Lamarckiana 7 (aabcdefg +a’'b’c'd'e' f' 9’) and
as much so as Lamarckiana 7 + Lamarckiana 7 (abcdefg+a'b’
gs a gs i ge :
Whatever the facts regarding these questions it is clear that if
8-chromosome gametes, capable of functioning, are produced by one
sex only (barring rare exceptions), as appears to be the case in many
instances at least, this alone is sufficient to explain the almost complete
absence of 16-chromosome mutants in cultures of 15-chromosome forms,

selfed.
3. The elimination of the extra chromosome by means of one or
more of the various processes observed by Gates and Miss Thomas *?

“ Gates (15a, p. 288) commenting upon the fact that O. lata rubricalyx produced,
presumably, only 14-chromosome offspring when selfed and crossed both ways
with several other forms, said: ‘‘Since there was an abundance of pollen, it would
appear probable that many of the grains must have received the extra chromosome
and that the latter was frequently lost during the divisions in the pollen tube.”
Gates’s suggestion is well worth considering, but we should not overlook other
possibilities in the case. While Gates and Miss Thomas (’14, p. 545) tell us that
lata rubricalyx “produced a good amount of viable pollen’”’ and that it ‘developed
long stout capsules”’ (p. 533), thereby indicating that seeds were produced in abun-
dance, Gates (’15a) further states that few offspring were obtained from crosses of
this mutant both ways with several other forms. Since 15-chromosome forms com-
monly produce very little pollen capable of functioning, or none at all, it is probable
that these plants were 14-chromosome forms; however that may be, let us assume
that only 7-chromosome eggs, capable of functioning, were produced by lata rubri-
calyx. Even though well-filled fruits were developed, if only relatively few of the
large number of seeds produced succeeded in germinating, perhaps those resulting
from 97 + o’8 unions were incapable of germinating, or failed to germinate in the
time allowed them. Perhaps one of the 15 male chromosomes was eliminated oc-
casionally during reduction by one or more of the numerous irregularities observed
by Gates and Miss Thomas in this form, such as failure to be included within the
heterotypic daughter nucleus and subsequent degeneration; degeneration on the
homotypic spindle, etc. In this way many more 7- than 8-chromosome pollen
grains may have been formed. We have no assurance that every seemingly good grain
is capable of functioning; neither can it be said, because 15-chromosome offspring have
not been found in cultures of O. Lamarckiana pollinated by certain 15-chromosome

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS Q7

during reduction in 15-chromosome forms may result in the production
of a greater number of 7- than of 8-chromosome female gametes and
may partially account for the greater number of 14- than of I5-
chromosome forms among the offspring of these plants selfed, and
among those of 15-chromosome plants pollinated by 14-.

Gates (’090) found all of the 21 chromosomes of the triploid forms
which he studied were distributed to the two poles of the heterotypic
spindle in groups of Io and 11, ordinarily, and 9 and 12, occasionally,
No evidences of degeneration were recorded. Geerts (’11), on the other
hand, found that only 14 of the 21 chromosomes of the triploid
hybrids which he examined were regularly distributed in groups of
7 each, the remaining 7 fragmenting and degenerating. The observa-
tions of these two workers being so unlike, the following statement was
made by Lutz (’12, pp. 404-405): “‘The evidence does not indicate
that we shall find one type of reduction exclusively in 21-chromosome
mutant A, for example, and another in a sister mutant B. It is
possible . . . that the type of reduction present in the male and
female germ cells of a flower depends upon its position on the plant.
. . . For instance, the reduction division in both the male and female
germ cells of the first flowers of a triploid plant might be represented ‘
by the Gates type almost exclusively, while that of the late flowers on
the same branch (or stem) might exhibit chiefly the Geerts type of
reduction, or vice versa. Perhaps also, the first flowers of a weak
lateral or sub-lateral branch may differ from the first flowers of the
stem or a strong basal branch.’’ An interview with Dr. Geerts later
revealed the fact that his fixations had been prepared in September and
October and that they had been taken from seed-plants, therefore
from individuals which had produced their first flowers much earlier
in the season; hence it was stated (p. 405) that “This indicates that
Geerts’s type of reduction appears in the later flowers, and Gates’s
probably in the earlier ones.’ Gates (’15a, p. 188) has since supported
this assumption by the statement that the material from which his
studies were made had been collected at the height of the flowering
season.

Gates and Miss Thomas (’14) have shown that one of the extra
chromosomes of O. /ata and various Jata-like forms sometimes degener-

forms, that 8-chromosome male gametes are not produced by the latter. Perhaps
all of the 8-chromosome poilen grains of O. lata rubricalyx and certain other 15-chro-
mosome forms, whether seemingly good or not, are incapable of functioning.

98 ANNE M. LUTZ

ate. While they did not tell us whether this occurs more frequently in
termina! than in basal buds, it is probable, judging from the evidence
produced by the 21-chromosome hybrids mentioned, that the extra
chromosomes of 14-+-chromosome forms degenerate more frequently
in the buds produced near the end of the stem or branch than in earlier
ones; or more frequently in the buds of a short, weak lateral or sub-
lateral produced near the close of the flowering period of the plant,
than in the buds of a vigorous branch.

Since 15-chromosome forms commonly produce no pollen or very
little seemingly good pollen, one can exercise but little choice in the
selection of pollen-flowers; yet if it were possible to self one of the early
flowers of the stem or a vigorous branch (not necessarily one of the
first), he might secure a higher percentage of 15-chromosome offspring
than is common, and even some 16-chromosome forms. ‘Terminal
flowers are avoided for obvious reasons; they are commonly regarded
as less vigorous than earlier onés, and those of annual plants are usually
produced too late to ripen seeds; furthermore, even those of 14-chro-
mosome forms frequently produce less pollen than earlier ones and
such pollen as is produced often contains a low percentage of seemingly
good grains. Terminal flowers of 15-chromosome forms or other
(14.+-28)-chromosome individuals producing but little pollen are
usually entirely male-sterile; yet by covering the stigmas of early and
late flowers of O. lata with Lamarckiana pollen and employing some
method which will secure the germination of all viable seeds, one should
be able to ascertain whether the early flowers produce more 8-chro-
mosome eggs, capable of functioning, than the late.*

48Tn ‘‘ New dimorphic mutants of the Oenotheras”’ referred to in note 36, de
Vries has shown conclusively since the above was written that in the case of the indi-
vidual O. cana which he employed, at least, as high a percentage of offspring dupli-
cating the characters of the mutant parent was produced by seeds derived from
selfed terminal buds of the stem and side branch as from those of selfed basal buds
of the same parts. If O. cana has 15 chromosomes, as is probable, he has shown that
the relative number of 8-chromosome female gametes was not less in the terminal,
than in the basal, buds of the stem and a side branch of the plant employed. We
may find this to be true of all 15—, or even of 14-++--chromorome forms in general, but
we must not overlook the fact that the germ cells of O. cana have not been studied
as yet, and it may be that degeneration occurs less frequently in certain 14-++-chrom-
osome types, or individuals of a given type, than in others. It may be, further-
more, that chromosome degeneration is less likely to occur in plants having 15 than
in those having 21 or 22 chromosomes. De Vries having found that a higher per-
centage of the offspring of selfed biennial, than of selfed annual, O. cana and O.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 99

Bartlett (’150, p. 141) says: ‘‘ Recent discoveries are making it very
clear that mutative changes in the chromosome number occur fre-
quently, and that such changes are always associated with a modi-
fication in the morphological characters of the plant. In other words,
certain mutations are probably dependent upon, or, at any rate, closely
associated with, visible changes in the nuclear mechanism. We have
every reason to believe, therefore, that the different chromosome
numbers of different species were acquired simultaneously with the
acquisition of other specific characters.”’

Gates ('150), as earlier quoted, states that whenever a germ-cell
having 8 chromosomes fertilizes a normal germ-cell (and he would
doubtless concede the reverse as well), a new form is produced and that
one of the most important factors determining the nature of the char-
acters of the new form is probably the peculiar combination of chro-
mosomes received.

It is now quite certain that whenever an offspring is derived from a
14-chromosome form by means of the union of an 8- with a regular or
irregular 7-chromosome cell, the offspring will invariably differ from
the parent in somatic character as well as in somatic chromosome
number. Likewise, it has been shown that whenever an offspring is
derived from a 15-chromosome plant by means of the union of two
7-chromosome cells, it will invariably differ from the parent in somatic
character as well as in somatic chromosome number. We have seen,
however, that unsurmountable difficulties are soon encountered when
one attempts to explain mutation on a strictly chromosomal basis.

We may now return to the question of the probable number of
chromosomes which were present in O. semzlata de Vries.

Having found that O. semilata de Vries and O. semilata Gates are
distinct types, Gates (15a, p. 111) concludes regarding the former that
“since it bred true it probably had 16 chromosomes,” and adds that it
will therefore be understood that de Vries’s form ‘“‘is another mutant
which probably had 16 chromosomes.”’

Whether or not O. semilata de Vries had 16 chromosomes is in itself
a matter of small importance, since the strain has died out and this
particular mutant type may never reappear; but the questions which
Gates’s statements raise are of considerable interest.

scintillans reproduce the characters of the mutant parent, indicates, in the opinion
of the writer, that the extra female chromosome degenerates less frequently in strong
bienntals than in the less vigorous aunuals and that 8+7 unions occur more frequently
than 7+7, when biennials are selfed, for the simple reason that relatively fewer female
7-chromosome gametes are produced.

100 ANNE M. LUTZ

It may be asked, Why should we assume that O. semilata de Vries
had 16 chromosomes rather than 14, or even 15? 7

We know that most 14-chromosome forms breed true and there is
evidence to indicate that a 15-chromosome form may breed true. The
semilata de Vries whose constancy de Vries tested, appeared among
the offspring of O. lata (15) X O. Lamarckiana (14).44 As is well
known, this cross produces among others, 14-chromosome O. nanella
and O. Lamarckiana, 15-chromosome O. albida and 14- or 15-chromo-
some QO. oblonga, all of which are said to breed true. The three 16-
chromosome mutants described in this communication are the only
16-chromosome forms which have been reported thus far and these
three plants exposed no pollen whatever, hence no offspring have been
obtained from them and we have no evidence to indicate that 16-chro-
mosome forms are more likely than 15- to breed true. While tt 1s
probable we shall find that most forms having an odd number of chromo-
somes are tnconstant, whether the number be 15 or 15+, 1t does not
follow that forms having an even number of chromosomes may be expected
to breed true when the number 1s in excess of 14, excebt, perhaps, when
this number 1s twice 14. It is probable that a Lamarckiana offspring or
descendant having a double set of the original, parental 14 chromo-
somes, as is possible in the case of O. gigas de Vries, might be more
stable than a 16-chromosome offspring, yet it is quite certain that the
28-chromosome mutant offspring of 14-chromosome forms with which
we are familiar do not breed true in the same sense as O. Lamarckiana.
The progeny of O. gigas de Vries do not all duplicate the parental
individual in every character, as in the case of O. Lamarckiana. As
is well known, a number of types are found among the offspring, yet
since they have many characters in common and since one type, when
selfed, seems to produce the same lot of offspring as any other, we
speak of this form as constant. Gates’s 28-chromosome Palermo
strain of O. gigas also ‘‘showed a considerable range of variation,
though not so great as in the Amsterdam race” (p. 121). Heribert-
Nilsson’s race (Heribert-Nilsson, ’12), representatives of which Gates
(15a, p. 124) has recently shown to have 28 chromosomes, also pro-
duces a variable lot of offspring. Furthermore, Bartlett (15c) has
announced that the 28-chromosome mutant offspring of O. pratincola
which appeared in his cultures, namely O. pratincola mut. gigas gave

44 De Vries recognized three semilata mutants in his cultures. All had lata
mothers.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS IOI

a diverse progeny of dwarfs, not a single individual of which resembled
the mother. While it appears that we are entitled to regard this form
as aan Professor Bartlett, in discussing this question by letter
savs, ‘‘The fact must not he overlooked, however, that this particular
individual of Oe. fratincola mut. gigas belonged to a mass mutating
strain.”

Gates (15a, pp. 189-190) described the meiotic distributions of
the chromosomes of a 22-chromosome offspring of O. gigas < O. lata
rubricalyx, presumably resulting from a 9 14+ 8 union. Here he
found that ° ‘the arrangement in the heterotypic telophase is distinctly
not into two equal groups of 11 each, but usually (and apparently with
much regularity) inte 10 and 12.’”’ In addition to these, 9-, 11-, and
13-chromosome groups were observed and certain other irregulavities
of distribution, but these were apparently uncommon. ‘Hence,’ he
says, ‘“‘we conclude that a considerable number of the pollen grains will
contain only nine chromosomes, although the majority will probably
COMA LO, L1, Or-12.”

The behavior of O. lata has shown us that although male reduction
ina plant may form daughter groups containing different numbers
of chromosomes, it does not necessarily follow that more than one type
of pollen grain, capable of functioning, will be produced. Thus, Gates
and Miss Thomas (14) have shown that male reduction in O. lata
usually results in 7-8 distributions of chromosomes, yet there is much
to indicate, as we have seen, that only 7-chromosome pollen grains,
capable of functioning, are produced by this form. Hence the ap-
pearance of 9-, 10-, 11- and 12-chromosome groups at various stages
of the male reduction process in the 22-chromosome hybrid does not
assure us that more than one type, or that any type of pollen grain,
capable of functioning, was produced by this form. Gates (15a,
p. 213) studied a sample of pollen from this hybrid containing 281
grains and found that 11.4 percent of the grains were “good.” It is
quite possible that all but one type of grain were eliminated and that
only one type of female gamete, capable of functioning in union with

20 The 22-chromosome offspring of O. lata X O. gigas sometimes produce small
quantities of pollen containing about the same percentage of seemingly good grains
as the pollen from Gates’s hybrid. I have repeatedly attempted to self these plants,
but in every instance have failed to secure a single seed. These results may have
been brought about by incompatibility of fertilization combinations, or it may be
that the “seemingly good’’ grains were just as incapable of functioning as the
shriveled and distorted ones.

TO2 ANNE M. LUTZ

the male gamete, was produced. If the fertilization combination
(such as 910 + o’12, or vice versa) resulted in 22-chromosome seeds
capable of germinating, or if, regardless of the chromosomal contents
of the gametes of both sexes, of the chromosomal combinations which
resulted from fertilization, only 22-chromosome seeds were capable of
germinating, it is clear that a 22-chromosome hybrid might breed true.
On the other hand, if functional male gametes of two or more types
(such as 10- and 12-chromosome cells) and a female gamete of a single
type (say 10-chromosome cells), or vice versa, were produced, and if a
single type of cell of one sex were capable of uniting with two or more
types of cells of the opposite sex and of producing seeds capable of
germinating, it is quite clear that the 22-chromosome hybrid would
not breed true.

If enough good pollen were produced by 16-chromosome mutant
offspring of selfed O. lata, O. Lamarckiana, or of O. lata X O. Lamarc-
kiana, to self them, one might expect them to prove less stable than
14-chromosome forms resulting from 7 + 7 unions, since they would
contain two extra chromosomes, whether derived from 8 + 8, or
9 +7 unions. It is quite possible, of course, that a germ-cell com-
bination would be formed which would enable the mutant to breed
true, such as 9 and o& 8, 9 7 and G6 9oor 2 o and o 7, but no evi
dence has been produced to assure us that such would occur. It would
not be at all surprising to find that 16-chromosome forms derived
from 8 + 8 unions produce only 8-, or 7- and 9-chromosome female
gametes, and, in case viable pollen is ever formed, only 7- male; and
that those derived from 9 + 7 unions produce 7- and 9- female, and
only 7- male gametes.

Returning to the case of O. semilata de Vries, de Vries states (’09,
Vol. I, p. 359) that he selfed one of these mutants and obtained 276
offspring. Of this number, 3 were O. nanella and 4 O. lata. ‘The
remaining plants clearly exhibited the characters of semilata and
justify the establishment of this form as a constant species.’ He also
pollinated O. Jata with this semilata plant and obtained 105 seedlings,
39 of which were O. lata, 1 O. albida, 61 O. Lamarckiana, 2 O. nanella
and 2 O. oblonga (the first two types having 15, the second two 14, and
the fifth, 14 or 15 chromosomes). ‘‘These forms,’’ he adds, ‘‘and the
proportions in which they occur are the same as those which O. lata
produces when crossed with other species’”’ (meaning, probably, when
QO. lata is pollinated with 14-chromosome species, since sufficient pollen

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 103

for fertilization purposes is rarely obtained from 15-chromosome forms;
however, this statement does not necessarily exclude the latter, since
no evidence has been produced to show positwely that 7 +-chromosome
male gametes, capable of functioning, are produced by any 15-chromosome
form). The results are practically the same as those derived from
15-chromosome O. lata X 14-chromosome QO. Lamarckiana. This
certainly indicates that the buds of O. semilata de Vries which were
employed in this cross, produced only 7-chromosome male gametes
capable of uniting with the 7- and 8-chromosome eggs of O. lata
and producing viable seeds. If this mutant had 16 chromosomes and
if only 7-chromosome male gametes were produced by all buds, it is
obvious that when this form was selfed, the only eggs which united
with the 7-chromosome male gametes and produced seeds capable of
germinating in the time which de Vries allowed them, had 9 chromo-
somes,

All facts and possibilities considered, it seems quite as probable
that O. semilata de Vries had 14, as 16, chromosomes. The production
of pollen by this form and the evidence of constancy, when selfed, do
not preclude the possibility of its having had 15 chromosomes.

We may briefly outline our conclusions regarding the factors which
determine the constancy or inconstancy of a plant as follows:

It has been shown that somatic chromosome number in Oenothera |
is constant; therefore, unless 15-chromosome offspring are produced
apogamously or unless the chromosomes in excess of 15 are eliminated
after fertilization takes place, it is evident that a 15-chromosome form
can breed true, 7. e., produce offspring having the somatic characters
of the parent in every case, only when two gametes having dissimilar
chromosome numbers, one odd and the other even, unite and produce
viable seed. While not all offspring resulting from such combinations
reproduce the parental characters, it is certain that, with the possible
exceptions noted, the parental type can be duplicated in no other
way. However, as we have seen, the constancy or inconstancy of a
plant 1s not determined solely by the presence of an even number of chro-
mosomes 1n the first case, and of an odd, in the second. All depends upon
the types of male and female, germ-cells produced and the fertilization
combinations which result in the production of seeds capable of germinat-
ing. Thus, mutant A, having 15 chromosomes, may produce only
8-chromosome gametes, type a, of one sex, and only 7-chromosome
gametes, type 0, of the other sex, or, although others are formed by

104 ANNE M. LUTZ

either or both sexes, these may be the only two that are capable of
uniting and producing viable seeds. If ithe 8 + 7 combinations unite
gametes which, together, reproduce the parental characters, the
plant will, of course, breed true. If they unite other types the plant
will prove inconstant, notwithstanding the fact that the offspring, like
the parent, will have 15 chromosomes.

C. 16-CHROMOSOME MUTANTS*%6

1. Lata-like Forms

The first 16-chromosome mutant recognized at Cold Spring Harbor
or elsewhere, was found in a 1908 culture of O. Lamarckiana X O. La-
marckiana, and the second in a I910 culture of the same form. Since
the somatic chromosome number of the 1908 mutant was ascertained
in the winter of 1908-1909 and that of the 1910 plant in the spring of
I9I1I, they were not known to be mutants of particular interest at
the time of their growth and were not photographed.

While the two were in no sense identical forms, both have been
properly characterized as Jata-like plants. In common with O. lata
Nos. 5343 (1908) and 3474 (1910) had crinkled leaves, yellow-green
foliage, irregularly shaped buds, and were male-sterile. The leaves
of No. 5343, in all stages of development, were conspicuous because
of their relatively short and broad leaf-blades and long petioles, but
the leaves of No. 3474 were very much narrower and more sharply
pointed than those of O. Lamarckiana lata. In both cases these dii-
ferences were very conspicuous in the full-grown rosettes. The true
lata mutant produced by Lamarckiana is usually much shorter
than Lamarckiana, but No. 3474 was almost as tall as the parental
form when full grown, its height being correlated, undoubtedly, with
the great distance between nodes—one of the conspicuous characters
of the plant. In proportion to the length of the stem, the branches

46 The discovery of 16-chromosome mutants in Oenothera was announced with the
following statement in 1912 (Lutz, ‘‘Triploid mutants in Oenothera,” p. 433): ‘‘I
may anticipate a future report sufficiently to state that I have found many quite
distinct types of mutants with 15 chromosomes, and some even with 16.”’ No
further information concerning these mutants was given out at that time and the
plants were not mentioned again by the writer until referred to in a paper read
before the Botanical Society of America in December, 1915, and in the note which
followed (Lutz, ’16a). Gates stated in 1915 (‘‘The mutation factor in evolution,”
p. 167) that 16-chromosome forms had been described, but since there appear to be

no recorded descriptions of such forms antedating the note just mentioned and the
paper in hand, it is probable that he referred to the mutants reported in 1912.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 105

of O. lata are rather long, but those of the two mutants were relatively
short. Lamarckiana and lata are almost invariably annual in this
climate (when seeds are sown in January and rosettes transplanted
in May) and flower quite early, the latter frequently earlier than the
former. No. 3474 bloomed quite late, about the middle of August.
The buds, seed-capsules, stem and branches of this plant were covered
with long hair. The petals of the open flower did not have the ordinary
crumpled appearance characteristic of O. Jaia, but were creased longi-
tudinally, much as if the flower had been drawn through a very small
finger ring. Many flowers had five or more petals. The stigmas
were very irregularly branched, much more so than in O. Jata, and an
anther occasionally bore a rudimentary petal. Somatic metaphase
groups from Nos. 3474 and 5414 are shown in Figs. 8a, 8) and 8c.

A
Xe
VAC
Rs c

b

Fic. 8. aand 0, unidentified Jata-like mutant, plant No. 3474, C.S.H., 1908
Offspring of O. Lamarckiana X O. Lamarckiana. Polar view of metaphase figures
from transverse sections of root-tips, showing 16 chromosomes. cc, unidentified
mutant, plant No. 5414, C.S.H., 1910. Offspring of O. lata, selfed. Polar view
of metaphase figure from transverse section of root-tip, showing 16 chromosomes.

2. The Dwarf Form Produced by O. lata, Selfed

This plant, No. 5414 (1910), was abnormal in appearance in all
stages of development. It is shown as a greenhouse rosette in Fig. 9a,
and as a full-grown garden rosette in 0, the diameter of the latter not
exceeding one fourth of the diameter of a full-grown lata rosette. The
plant came to flower late in the season on a very short stem and it is
impossible to state whether this was due to the character of dwarfness
(suggested in the rosette) or to a depauperate, abnormal condition.

106 ANNE M. LUTZ

No. 5414 produced no pollen whatever. A somatic metaphase group
from this plant (Figs. 9a, 9b) is shown in Fig. 8c.

a

Fic. 9. a, plant No. 5414 (see 8 c) in greenhouse rosette stage. 6, same plant
in late garden rosette stage. About 1.5 dm. in diameter; growth completed.
Note its abnormal appearance.

3. Origin of the 16-chromosome Condition in Offspring of 14- and 15-
| chromosome Forms.

As has been pointed out on preceding pages of this report, the 16-
chromosome condition in the three mutants may have arisen through
9 +70r8 + 8 unions; it is difficult to state which is the more probable
in either case. If the plant produced by O. lata was the product of the
first combination, it is probable that it resulted from a 994+ 07,
rather than a 9 7 + o’9Q, union.

The 1908 and 1910 Lamarckiana mutants were far more suggestive
of O. lata than the 1910 offspring of selfed Jata. The latter could not
be designated as a Jata-like plant. It might be suggested that the
two former may each have arisen from the union cf two 8-chromosome
gametes bearing /ata characters and the latter from a 29+ 7
union, but we should then be.required to explain why the Lamarckiana
mutants did not have duplicate vegetative characters.

The 1908 plant was grown during the first season in which the
writer studied the vegetative characters of various plants of the
Lamarckiana group with particular care, hence minor differences
between No. 5343 and ordinary lata may have been overlooked. The
records refer merely to the distinguishing characters of the leaf,

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS 107

branching habits, etc. If this mutant bore signs of abnormality or
irregularity such as were observed in the other two forms, the fact was
not observed; yet the indications of abnormality in Nos. 3474 and
5414 lead us to conclude that the gametic combination or combinations
which produced these two plants may have been less ‘‘compatible”’
than the combination which produces O. Jaia.

SUMMARY

1. The primary object of this series of three reports, of which the
one in hand is the second, is to discuss, in the light of the Cold Spring
Harbor and Louvain studies of Oenothera, certain theories and con-
clusions which Gates has advanced from. time to time and which Gates
and Miss Thomas have based upon the results of their investigations.

2. O. Lamarckiana mut. lata, long believed to have 14 chromosomes,
is now known to have, invariably, 15. The researches of Gates and
Miss Thomas appear to have led them to conclude, further, that the
presence of the extra chromosome in the somatic cells of 15-chromo-
some offspring of 14-chromosome forms is invariably associated with
the presence of Jata or lata-like vegetative characters. Later, Gates
recognized the fact that his 15-chromosome mutant O. incurvata is
quite different from O. /aia, as is also a 15-chromosome form which
de Vries reported. That he is loth to concede that these discoveries
render untenable certain earlier statements of Gates and Miss Thomas
is indicated by the statement that “It is perhaps not inappropriate
to speak of all these mutants as belonging to the /aéa series, or series
with an extra chromosome.” The primary object of this paper,
therefore, is not only to emphasize the fact that these two mutants
cannot be regarded as lata-like forms, but to show that many other
15-chromosome mutant offspring of 14- and 15-chromosome forms,
are not /ata-like.

3. The distinct types of mutants which the Cold Spring Harbor
and Louvain studies have shown to have 15 chromosomes, are: (1)
O. lata, (2) O. albida, (3) O. bipartita, (4) type 5509 (supposed to be
modified O. oblonga)—all Lamarckiana mutants. (5) O. nanella lata,
produced by O. Lamarckiana, O. nanella, O. lata X O. Lamarckiana,
etc. (6) O. subovata, found in cultures of O. Lamarckiana and O. lata
x O. Lamarckianae. (7) A dwarf type, 2256, produced by O. nanella,
selfed. (8) Type 4499, found in cultures of O. lata < O. Lamarckiana
and O. lata, selfed. (9) O. exilis, (10) O. exundans and (11) type

108 ANNE M. LUTZ

5365, all found in cultures of O. lata, selfed. In addition to the fore-
going, type 2806, having many points in common with type 5509,
also has 15 chromosomes.

4. Of the above 12 types (11 of which were quite distinct) now
known to have 15 chromosomes, only two are Jaia-like; namely,
O. laia and O. nanella lata.

5. Certain somatic characters of many mutant offspring of O. La-
marckiana X O. Lamarckiana and O. Lamarckiana, selfed,: indicate
that a very large percentage of the mutant offspring of O. Lamarckiana
have 15 chromosomes and that a larger number of 15- than of 14-chro-
mosome mutant offspring are produced by this form. Not only does
there appear to be a larger number of distinct types of 15- than of 14-
chromosome mutants, but a higher percentage of 15- than of 14-chro-
mosome individuals produced by O. Lamarckiana. Only a small
percentage of the former may be classed as Jata-like, or as semilata-
like, forms.

6. While /Jata-like forms are commonly characterized by 15 chro-
mosomes, three distinct types have been found in Cold Spring Harbor
cultures with 16 chromosomes. ‘Two appeared in separate cultures of
O. Lamarckiana X O. Lamarckiana (1908 and 1910) and one among
the offspring of O. lata, selfed (1910).

7. Owing to the fact that 15-chromosome forms are very often
male-sterile, or produce but little pollen capable of functioning, their
constancy has not been tested upon an extensive scale. 15-chromo-
some mutants O. lata, O. semilata Gates, O. lata rubricalyx, O. bipartita
and 15(?)-chromosome O. elliptica, are known to be inconstant, while
de Vries’s researches indicate that 15-chromosome O. albida and 14-
or 15-chromosome O. oblonga are constant.

8. The evidence available at present indicates that most 14-chro-
mosome forms are constant and most 15-chromosome forms incon-
stant. Furthermore, the available evidence indicates that inconstancy
is commonly associated with the 14+-chromosome condition. It
seems, however, that forms having twice 14 chromosomes are more
likely to be constant in the same sense that de Vries’s O. gigas is
constant, than those having more than 14, but fewer than 28, chromo-
somes.

9g. While it is probable that we shall find that most forms having
an odd number of chromosomes are inconstant, whether the number
be 15 or 15+, it does not follow that forms having an even number

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS I09

of chromosomes may be expected to breed true when the number is in
excess of 14, except, perhaps, when this number is twice 14. Hence
the fact that O. semilata de Vries bred true, scarcely warrants the
conclusion that this form probably had 16 chromosomes. We now
know that 14-chromosome forms usually breed true and the evidence
indicates that an occasional 15-chromosome form is also perfectly
constant, while there are no records to show that offspring have been
obtained as yet from forms known to have 16 chromosomes. It there-
fore seems quite as probable that semilata de Vries had 14, or even I5,
as 16, chromosomes.

10. Since somatic chromosome number has been shown to be
constant in Oenothera, it is clear that unless 15-chromosome offspring
are produced apogamously, or unless the chromosomes in excess of 15
are eliminated after fertilization takes place, 15-chromosome forms can
breed true, 7. e., produce offspring having the somatic characters of
the parent in every case, only when two gametes having dissimilar
chromosome numbers, one odd and the other even, unite and produce
viable seeds. While not all offspring resulting from such combinations
reproduce the parental characters, it is certain that, with the excep-
tions noted, they can be duplicated in no other way.

11. The constancy or inconstancy of a plant is not determined
solely by the presence of an even number of chromosomes in the first
case and of an odd in the second. All depends upon the types of
male and female germ-cells produced and the fertilization combinations
which result in the production of seeds capable of germinating.

12. As a rule, larger quantities of seeds are obtained from 14-
chromosome forms selfed, or pollinated by other 14-chromosome
forms of the same, or different species, than from 14-+-chromosome
forms selfed, or pollinated by other 14+-chromosome plants of the
same, or different species, particularly if the 14-+-chromosome
individuals have more than 14, but fewer than 28, chromosomes;
furthermore, higher percentages of germination are usually secured
from the former than from the latter when seeds not more than one
year old are sown in pans of sterilized soil in January and are kept
under ordinary greenhouse conditions.

13. The number of seeds produced by a form and the ability of the
seeds to germinate, at least within the time limits specified, are factors
which appear to be associated with the chromosome number of the
plant, or numbers of the plant, producing them.

IIO ANNE M. LUTZ

14. The ability of a seed to germinate seems to depend, not wholly,
but to a certain extent, upon the number of chromosomes which it
bears, and, possibly, in accordance with Gates’s suggestion, upon the
compatibility or incompatibility of the chromosomal combination
which the number represents.

15. The ability of a seed to germinate appears to be directly asso-
ciated with its own chromosome number and only indirectly with that
of its parent, for the results derived from the Cold Spring Harbor and
Louvain studies indicate that 14-chromosome seeds of 14 +-chromo-
some forms germinate quite as readily as 14-chromosome seeds of 14-
chromosome forms.

16. Plants having more than 14, but fewer than 28, chromosomes
are more inclined to be male- than female-sterile. Just why this is
so, 1s not yet clear.

LAFAYETTE, INDIANA

BIBLIOGRAPHY

Bartlett, H. H. ('15a).. The mutations of Oenothera stenomeres. Amer. Journ. Bot.
2: 100-109. Feb., 1915:

(15). The experimental study of genetic relationships. Amer. Journ. Bot.
2: 132-155. March, 1915.

(15c). Mass mutation in Oenothera pratincola. Bot. Gaz. 60: 425-456. Dec.,
IQI5.

Davis, B. M. (10). Cytological studies on Oenothera. I. Notes on the behavior of
certain hybrids of Oenothera in the first generation. Amer. Nat. 44: 108-115.
Feb., I91I0.

(11). Cytological studies on Oenothera. IIT. A comparison of the reduction
divisions of Oenothera Lamarckiana and O. gigas. Annals of Botany 25:
941-974. Oct., IQII.

(15a). A method of obtaining complete germinations of seeds in Oenothera and
of recording the residue of sterile seed-like structures. Proc. Nat. Acad. Sci.
I: 360-363. June; 1015.

(150). Review of Gates’s ‘‘ The mutation factor in evolution with particular
reference to Oenothera.”’ Science, n. ser., 42: 648-651. Nov., 1915.

Gates, R. R. (’07a). Preliminary note on pollen development in Oenothera lata, de
Vries, and its hybrids. Science, n. ser., 25: 259-260. Feb., 1907.

(076). Pollen development in hybrids of Oenothera lata X Oe. Lamarckiana,
and its relation to mutation. Bot. Gaz. 43: 81-115. Feb., 1907.

(’o7c). Hybridization and germ cells of Oenothera mutants. Bot. Gaz. 44:
I-21. July, 1907.

(‘oga). Studies of inheritance in the evening primrose. Chicago Med. Re-
corder, repaged separate, pp. 6. Feb., 1909.

(090). The behavior of chromosomes in Oenothera lata X O. gigas. Bot. Gaz.
48: 179-198. Sept., 1909.

FIFTEEN- AND SIXTEEN-CHROMOSOME OENOTHERA MUTANTS III

(10). The chromosomes of Oenothera mutants and hybrids. Proc. Seventh
Internat. Zool. Congress, Boston, Aug., 1907.

(12). Somatic mitoses in Oenothera. Annals of Botany 26: 993-1010.

(13). Recent papers on Oenothera mutations. New Phytol. 12: 290-302.
Oct., 1913:

(15a). The mutation factor in evolution with particular reference to Oenothera.
Macmillan and Co., London, pp. 353. 1915.

(150). Heredity and mutation as cell phenomena. Amer. Journ. Bot. 2:
519-528. I0915.

Gates, R. R. and Nesta Thomas (’14). A cytological study of Oenothera mut. lata
and Oe. mut. semilata in relation to mutation. Quart. Journ. Micro. Sci. 59:
523-571. I914.

Geerts, J. M. (11). Cytologische Untersuchungen einiger Bastarde von Oenothera
gigas. Bericht. Deutsch. Bot. Ges. 29: 160-166. March, I1ort.

Heribert-Nilsson, N. (’12). Die Variabilitat der Oenothera Lamarckiana und das
Problem der Mutation. Zeitschr. f. ind. Abst. u. Vererb. 8: 89-231. 1912.

Hunger, F. W. T. (’13). Recherches experimentales sur la mutation chez Oenothera
Lamarckiana, executées sous les Tropiques. Ann. Jard. Buitenzorg, II. 12:
92-113. I913.

Lutz, Anne M. (’08). Chromosomes of the somatic cells of the Oenotheras. Science,
n. ser. 27: 335. Feb., 1908.

(09). Notes on the first generation hybrids of Oenothera lata 9 X O. gigas of.
Science, n. ser. 29: 263-267. Feb., 1909.

(12). Triploid mutants in Oenothera. Biol. Centralbl. 32: 385-435. Aug.,
1912.

(16a). The production of 14-+-chromosome mutants by 14-chromosome
Oenothera Lamarckiana. Science, n. ser. 43: 291-292. Feb., 1916.

(160). Oenothera mutants with diminutive chromosomes. Amer. Journ. Bot. 3:
502-526. Nov., 1916.

MacDougal, D. T., A. M. Vail, G. H. Shull, and J. K. Small (05). Mutants and
hybrids of the Oenotheras. Carnegie Inst. Pub. No. 24: pp. 57. 1905.

MacDougal, D. T., A. M. Vail, and G. H. Shull ('07). Mutations, variations, and
relationships of the Oenotheras. Carnegie Inst. Publ. No. 81: pp. 92. 1907.

Schouten, A. R. (’08). Mutabilitat en Variabilitat. Dissertation, Groningen, pp.
196. 1908.

de Vries, H. H. (09). The mutation theory. Vol. I. Transiated by Farmer and
Darbishire. The Open Court Pub. Co., Chicago, pp. 575. 1909.

(12). Die Mutationen in der Erblichkeitslehre. Gebriider Borntraeger,
Berlin. Vortrag, Rice Institute, pp. 42. I9gI2.

(13). Gruppenweise Artbildung, unter spezieller Beriicksichtigung der Gattung
Oenothera. Gebriider Borntraeger, Berlin, pp. 365. 1913.

(15a). The coefficient of mutation in Oenothera biennis L. Bot. Gaz. 50:
169-196. March, 1915.

(150) Oe0nothera gigas nanella, a Mendelian mutant. Bot. Gaz. 60: 337-345.
Nov., I915.

THE INFLUENCE OF TEMPERATURE ON THE GROWTH
OF ENDOTHIA PARASITICA

NEIL E. STEVENS

In an earlier paper (4) the writer discussed the influence of certain
climatic factors on rate of vegetative growth and production of
ascospores in Endothia parasitica (Mur.) And. and And. From the
data then available it was concluded that the rate of lateral growth of
cankers on Castanea dentata (Marsh) Borkh. was directly dependent
on the amount and duration of temperatures favorable for growth and
apparently unaffected by the amount or frequency of rainfall. Asco-
spore production on the other hand seemed to be dependent chiefly
on the presence of abundant moisture. The data on which these
conclusions were based were obtained from observations made at a
series of stations extending from Concord, N. H., to Charlottesville,
Va., during the summers of 1914 and 1915.

Although it has been necessary to abandon several of the stations
because of the increasing abundance of the chestnut blight, observa-
tions have been continued in six localities. The results seem to
warrant a brief statement. As the methods employed have been
fully discussed in the earlier paper they need not be considered here.

RATE OF LATERAL GROWTH

The abundant rainfall of the summer of 1915 resulted in the
production of ascospores on practically all the inoculations at every
station, consequently no further data on this point could be obtained.

TABLE [|

Lateral Growth of Cankers of Endothta parasitica in Various Localities

Locality laa nese anata: Centimeters
Concord. NSE ste ees ers notin ees 350 May 18 14
Williamstown? iMacs< er ge 711 (900) May 22 15
Amherst, Mass. (two stations)........... 222 May 17 17
Woodstock: sNie Nan ne mae mevera 0. cae, Gee 1,000 May 24 15
Washington SC aie. wae. ee ene II12 (400) May 4 21
Charlottesville {Vary ee ee ees re eee 854 opribns: a 23

GROWTH OF ENDOTHIA PARASITICA I13

The lateral growth of the cankers at the various stations is given in
Table I. The amount given is, as in the earlier paper, an average of
all the normal appearing cankers from ten inoculations.

TABLE II
Total Precipitation (in Inches)

Concord, Williams- Amherst, Mohonk Washing- | Charlottes-

New ELs town, Mass. Mass. Lake, N. Y.| ton, D.C. | ville, Va.
JS Tic) AOC | oa eee — oa == = == 0.49
Rey nO iso 2ce 2 0.99 1.46 1.20 2.54 2.18 2.44
WONC, TOTS a... 00% 2. - 1.39 173 3.00 2.65 6.58 5.32
Wietive: MOMS. 6 8s. is be es 10.29 9.37 9.13 8.24 3.21 a7
PMUSUS TOTS... 5... . 6.26 4.47 8.28 7.94 7.00 798
September, I915..... 1.21 3.44 1.37 2.87 1.39 2.38
October, 1915....... 2:02 Zao 2.89 2.50 2572 4.39
November, I915..... ZO7, 2.03 2.20 T.22 0.93 1.92
December, I915...... 3.41 5.03 5.86 8.90 2.80 3.54
Jamiary, 19OT6..:.... 1.22 2.05 2.56 2.64! 1.57 1.34
February, 1916.>.... 4.18 1.53 5.27 5.54 225% 4.10
Maren, 1Or6)........ 20% 3.51 3.97 5.76 2.80 4.23
POPS OO. os es 2.96 2.48 3.69 4.05 2.96 2.35
Ry TOMO 5. 3.95 3.52 Bu2i 2.93 2.30
Total for year ending.) 5-31-16 | 5-31-16 | 5-31-16 | 5-31-16 | 5-31-16 | 3-31-16
nis 43.83 41.87 51.43 55:24 38.13 41.69

TABLE III
Number of Days with Precipitation .or Inch or More
Concord, Williams- | Amherst, Mohonk Washing- | Charlottes-

N. H. town, Mass. Mass. Lake, N. Y.| ton, D.C ville, Va.
J 181 010) re — — — —- — 4
MTA AOE Sn es eas 9 8 1k 9 II 10
UME RSEOTS <5. sw o's oie 10 II 8 Gf 14 IO
ily 1615. ok 16 20 14 18 13 12
PUSS TOTS i. 15 15 14 10 18 15
September, I915..... 5 6 77 6 7 7
October, 19O15....... 9 10 7 4 13 14
November, I1915..... 16 13 7. 2 8 4
December, I19I5...... 12 16 19) 6 9 6
january; 1916....:.. 8 12 9? 9 13 10
Bebruiary, 1916...... 13 14 14 6 10 9
March, 1916. ... <5... 10 13 We 8 10 12
Pepril, TLS: so. . 15 10g} 13 17 13 9
ER ee 0, i II 10 12 II II
Total for year ending .| 5-31-16 | 5-31-16 | 5-31-16 | 5-31-16 | 5-31-16 | 3-31-16

140 153 | 129 105 139 113

1 Data taken from West Point.
2 Data taken from West Point.

Report from Mohonk Lake missing.
Report from Mohonk Lake missing.

Il4 NEIL E. STEVENS

Comparison of the amount of growth at the various stations for
the year ending in the spring of 1916 with that in the same localities
for the years ending in May and in August, 1915, shows a general
agreement, although the growth at Charlottesville was only 23 centi-
meters for the year ending in April, 1916, as against 25 centimeters
for the year ending in April, 1915.

RELATION OF RAINFALL TO GROWTH

In considering the influence of rainfall on vegetation both the
total amount of precipitation and its frequency must be taken into
account. Tables II and III give the monthly totals and number of
days with over .o1 inch of rain for each month during the period under
consideration, together with the totals of the twelve calendar months
most nearly coinciding with the period for which growth was actually
measured. From these it is apparent that no causal relation exists
between the amount or the frequency of rainfall and the rate of growth.
For example, the total rainfall for the year was very nearly the same
at Williamstown as at Charlottesville but the growth was fifty percent
greater at the latter point. Even more significant is the fact that
although the rainfall at Concord, Williamstown, and Mohonk Lake
was much greater for the year ending in May, 1916, than for the year
ending in May, 1915, a difference of about twenty inches at Mohonk
Lake, there was no perceptible difference in the rate of lateral growth.
A comparison of the number of days with rain and of the rainfall
for the warmer months at the various stations also fails to show any
relation between rainfall and rate of growth.

METHODS OF COMPUTING TEMPERATURE EFFICIENCY

No method of interpreting climatological temperature data with
reference to the influence of temperature on plant growth has yet been
devised. The monthly and annual mean temperatures given in the
climatological reports are obviously of little use for this purpose.
Length of frostless season is of course important for many plants but
has little or no significance for a fungus like Endothia parasitica, whose
growth is by no means confined to the frostlessseason. In order that the
temperature data given in meteorological reports may be really useful
in plant climatology, it is necessary to obtain some kind of temperature
indices which will express the effect of temperatures on plant growth.

GROWTH OF ENDOTHIA PARASITICA II5

Such temperature indices must take into consideration both the daily
temperature means and the frequency with which those means occur
during the period under consideration.

Among the methods suggested for attaining this desired end the
one most widely used has recently been designated by Livingston (3)
as a summation of remainder indices. This method consists in sub-
tracting a certain assumed minimum from each daily mean tempera-
ture and summing the remainders. A second* method was suggested ©
a few years ago by the Livingstons (2). It is based on the supposition
that plant growth follows the chemical principle of van’t Hoff and
Arrhenius, which states that the velocity of many chemical reactions
approximately doubles with each increase in temperature of 10° C.
On this basis these authors have computed efficiency indices for the
various temperatures, using 40° F. as unity.

The two methods just described are open to the theoretical objec-
tion that they fail to take into account the fact that the highest tem-
peratures experienced in nature do not permit as rapid growth as
somewhat lower temperatures.

In an attempt to overcome this defect Livingston (3) has recently
published a series of temperature efficiency indices based on actual
physiological experiment. Using the data obtained by Lehenbauer
(1) for the average hourly rates of elongation of shoots of seedling
maize plants when exposed for periods of twelve hours to temperatures
of 12 to 43° C., he has derived a series of indices which express the
average hourly growth rate for each degree C. or F. in terms of the

growth rate for 4.5° C. (38° F.) considered as unity.

This series differs from the two described above in that the indices
gradually increase up to a certain point (89° F.) and then decrease at
higher temperatures. The optimum temperature thus indicated is of
course that of the maize seedling under the conditions of Lehenbauer’s
experiment and is higher than any daily mean reached during this
investigation. Moreover, the rate of increase in index value between
the minimum and optimum for growth is much more rapid in the
physiological series than in either of the other series. So far as the
present study is concerned this constitutes the chief difference between
this system and the other two.

’ These methods of interpreting temperature data are rather fully discussed by
Livingston (2 and 3) and their application to the study of Endothia parasitica by
the writer (4).

116 NEIL E. STEVENS

RELATION OF TEMPERATURE TO THE GROWTH OF FEndothia parasitica

As in my earlier paper, the temperatures as given by the U. S.
Weather Bureau reports for the various localities under consideration
were computed according to the methods of summing remainder
indices and summing exponential indices. The results are given in

Aer Oy cpl (25S ance
x 8 qe os
Se er en
ee ee oe ieee
S N SX
R WYN Q = NS
SD ga Somes aaa Bet
ee S N

EXPLANATION:

een AL LATERAL GROWTH

OF CANKERS
790
— EMIPERATURE COMPUTED BY

REMAINDER SUMMATION INACES
°

oone= TEMPERATURE COMPUTED BY |

EXPONENTIAL SUMMATION
780 INOICES

mee KEM PERATURE COMPUTED BY |
PHYSIOLOGICAL SUPIMATION
INOICES

100 ©

Fic. 1. Lateral growth of cankers of Endothia parasitica and temperature
computed in various ways for the year ending in May, 1916. All data expressed in
percentage of that at Concord.

Table IV, and Figure 1. The graph expresses the rate of growth of
the fungus and the temperatures at the various localities in percentage
of that at Concord, growth and temperature at Concord being con-
sidered 100 percent. The results of computing temperature by these
two methods are closely comparable, the curves being nearly parallel

GROWTH OF ENDOTHIA PARASITICA II7

throughout their length. These two curves are in turn very similar
to the curve of growth, although they rise somewhat more rapidly
in the more southern localities. This is in general the same relation
which was found to hold for the years ending in May and in August,
1915, at a still larger number of localities. Taken together these
results furnish a considerable body of evidence that either of these
methods of calculation expresses satisfactorily the relation between
air temperature and the growth of Endothia parasitica within this area.

In computing the physiological temperature efficiency the daily
mean temperatures were calculated by the formula mean = %
(maximum + minimum). For this mean temperature the equivalent
index from Livingston’s (3) Table II, p. 406, was substituted and the
sums of these daily indices considered the index for the year.

TABLE IV

Lateral Growth of Cankers of Endothia parasitica and Temperatures Computed in
Various Ways for the Year Ending in May, 1916

Remain- Expo- Physio-
Growth der nential logical
in Centi-) Percent} Sum- | Percent Sum- Percent Sum- Percent
meters. mation mation mation
Indices Indices Indices
Concord, N. H....... EA i LOO} | 2.067, -Loo 366 | 100 5,514 | 100
Williamstown, Mass...... Wn LO 3,038)) L024 16.373 7) LOLs 254576) TOLD
Amherst, Mass.......... L7 cs L204 3,360: 0113.0 | 431) 117.8|- 6,673-) 127.0
Woodstock, N. Y..°\..... £5, 1,107.3 | 3,100 |°104.6:|\- 386. | 505.5'| 5,632 |. 502.2
NWwashington): D.C... 0.2.) 28 |. 150.0 |'4;976 | 167.7 |. 603 | 165 -| 11,080 | 201.1
Charlottesville, Va....... 22 74 104.3 15,300 | 180,7.| 636) 4/. 174" 11,620::|- 271

The results are given in Table IV. It will be observed that the
physiological temperature indices rise considerably faster from north
southward than do the summation indices, and that accordingly
they correspond rather less well with the rate of growth of the fungus.

It is of course not surprising that the results obtained from the
use of the physiological temperature indices given by Livingston should
not more nearly approximate the growth of Endothia parasitica, since,
as Livingston correctly points out (p. 407), the indices are based upon
tests of only a single plant species, maize, and from the growth of
seedlings, and it is entirely probable that they are not even approxi-
mately true for plants of some other species. On the other hand,
when the necessarily approximate nature of many of the data are
considered, the agreement between the curve of growth of cankers of

1i8 NEIL E. STEVENS

Endothia parasitica and those showing the temperature of the various
localities is remarkable. This together with the evident lack of agree-
ment between the rate of growth and the amount of rainfall, strongly
suggests that the rate of growth of this fungus while growing as a
parasite on Castanea deniata, is influenced chiefly by temperature.

The data presented in this and the preceding paper indicate clearly
that the growth of the chestnut-blight fungus is more rapid in the
southern portion of its present range than in the region farther north.
Unless some unforeseen factor checks its development, the disease
may reasonably be expected to spread still more rapidly as it advances
southward.

SUMMARY

The lateral growth of cankers of Endothia parasitica on Castanea
dentata in various localities was about the same for the year ending in
May, 1916, as for the year ending in May, I915.

Neither amount nor frequency of rainfall seems to have any
influence on rate of lateral growth. Wide differences in the rainfall
for the two years produced no change in rate of growth. te

The temperature for the period under investigation was computed
according to the systems of ‘‘remainder summation indices,” “‘expo-
nential summation indices,’ and ‘physiological indices.’’ Of these
the last seems to agree least well with the rate of growth of E. parasitica.

The first two systems give practically identical results.

The agreement between the curves of temperature and of growth
is so Close as to indicate that temperature is the chief climatic influence
in determining the rate of growth of Endothia parasitica.

INVESTIGATIONS IN FOREST PATHOLOGY,

BUREAU OF PLANT INDUSTRY,
WASHINGTON, D. C,

LITERATURE CITED

1. Lehenbauer, P. A. Growth of Maize Seedlings in Relation to Temperature.
Physiol. Res. 1: 247-288. f. 4. 1914.

2. Livingston, B. E., and Livingston, Grace J. Temperature Coefficients in Plant
Geography and Climatology. Bot. Gaz. 56: 349-375, f. 3. 1913.

3. Livingston, B. E. Physiological Temperature Indices for the Study of Plant
Growth in Relation to Climatic Conditions. Physiol. Res. 1: 399-420. 1916.

4. Stevens, Neil E. The Influence of Certain Climatic Factors on the Development
of Endothia parasitica. Amer. Journ. Bot. 4: I-32. 1917.

“Tnvine W. eee Oe
@ ae. dha

pt The Gals Sader eee except ae “Au us and: September. 9
. Subscription price, $5: 00a year. Single copies ( 60. cents, plus postage. “Postage -w ee =

PRE ANY

vs Cereed to. all Hoes Aqdaee Saree Apes Nae rto Rit >; Panama, o's 7p

i (i
antics in’ re

eyes
A

Towne of ‘Botany, ee

; a amount (Gopoxtnately aio oe tabular. after 5 15,¢¢ r
for oe” se t 2 ibs |

cee J Renistinces Goula be cade payable 0 ‘Americ

ee

s must be Bee all checks not ee a

ee ‘Journa .
‘OF cs

at

te of mailing. The publishers vail: “supply missin

os ae Coen ‘concerning editorial pater should
C. New ombe, ) Geddes Heights, Ana Arbor MCR yo

| Business. ‘correspondence, including embe of chang
cerming. 5 ‘prints, ‘should: be. addressed to. Americ: in
Botanic. eee i, Brooklyn, ag Mer or 41: “Ove

tf)

AMERICAN
JOURNAL OF BOTANY

NOL, LV. MARCH; 1917 No. 3

MATROCLINIC INHERITANCE IN MUTATION CROSSES OF
OENOTHERA REYNOLDSIE

GArEs DD VvA RUE. AND di;:H. BARTLETT
INTRODUCTION

This paper is concerned primarily with the peculiar type of in-
heritance exemplified among the mutations of Oenothera Reynoldsi11,
a species elsewhere described as showing the phenomenon of “‘mutation
en masse,’ or mass mutation. It has been found that the mutations
characteristic of mass mutation in this species, when crossed among
one another, or with the parent form, give crosses which in general
conform exactly to the type of the pistillate parent, quite regardless
of which way the cross may have been made.

De Vries? has shown that in Oenothera Lamarckiana, the most
thoroughly studied of the evening-primroses, the total number of
mutations lies in the neighborhood of 2.2 percent. Certain mutations
from Oe. Larmackiana are themselves more mutable than their parent.
Thus Oe. lata produces twice, and Oe. scintillans three times as many
mutations as Oe. Lamarckiana itself. Before the discovery of mass
mutation in Oe. Reynolds and Oe. pratincola, a form was considered
highly mutable if its progeny contained as many as five or six percent
of mutations. Aside from Oe. Lamarckiana, however, only one species,
Oe. biennis, had been extensively grown for the detection of mutability,

1 Prior to 1915 the work upon which this paper is based was carried on by the Office
of Physiological and Fermentation Investigations, Bureau of Plant Industry, U. S.
Department of Agriculture, and since then by the University of Michigan. Pub-
lished by permission of the Secretary of Agriculture.

Papers from the Department of Botany of the University of Michigan, no. 155.

2 De Vries, Gruppenweise Artbildung, p. 329 et seg.

[The Journal for February (4: 53-118) was issued Feb. 17, 1917.]
119

120 CARL D. LA RUE AND H. H. BARTLETT

and this species had been shown by Stomps’ and De Vries‘ to be less
mutable than Oe. Lamarckiana. In the recently discovered mass-
mutating species the number of mutations may rise to almost 100
percent of the progenies.

The elementary species that have thus far shown mass mutability
are both segregates from the collective species that passes in our floras
as Oenothera biennis. ‘True Oe. biennis seems to be found in America,
but the records in regard to its occurrence have not yet been published.
It is therefore not incorrect to state that the species (in the narrow
sense) is definitely known only in Europe, where it occurs as an in-
troduced weed. The name Oe. biennis has been applied correctly by
De Vries and Stomps, but very loosely indeed by American geneticists,
with the result that the literature is considerably confused. Oe.
Reynolds and Oe. pratincola are two, among a number of segregates
from the collective species of the floras, that have been described and
named? for the purpose of keeping a clear record of the genetical ex-
periments that are being carried out with them. They are not recog-
nized in current systematic works.

The first paper dealing with Oenothera Reynoldsu* was written
before any mutation crosses had been made. It was therefore only
natural to suggest that the whole series of mutations to which it was
giving rise were probably Mendelian recessives. Work on the closely
related segregate Oe. pratincola shortly afterward disclosed the fact
that the mutations characteristic of mass mutation were not Mendelian
recessives, but showed matroclinic inheritance in crosses with their
parent form.’ It has now been determined that the first suggestion
in regard to the mutations of Oe. Reynoldsii was entirely erroneous,
since they likewise show matroclinic inheritance. Although the
special purpose of this paper is to present the data in regard to matro-
clinic inheritance, there is one other striking discovery which it is
possible to announce at this time, namely, that in Oe. Reynolds, as

®Stomps, Theo. J., Mutation bei Oenothera biennis L., Biol. Centralbl. 32:
521-535. 1912; Parallele Mutationen bei Oenothera biennis L., Ber. Deutsch. Bot.

Ges. 32: 179-188. I9I4.
4De Vries, H., The Coefficient of Mutation in Oenothera biennis L., Bot. Gaz.

59: 169-196. I9QI5.
> Bartlett, H. H., Twelve Elementary Species of Onagra, Cybele Columbiana 1:

37-56.. 1915.
6 Bartlett, H. H., Mutation en masse, Amer. Nat., 49: 129-139. I915.
7 Bartlett, H. H., Mass mutation in Oenothera pratincola, Bots Gaz., 60: 425-456.

1915.

MATROCLINIC INHERITANCE 125

well as in Oe. pratincola, the occurrence of mass mutation is associated
with a remarkable increase in seed sterility. This very significant
fact is being made the subject of further study. The degree of seed
sterility in mass-mutant Oe. Reynoldsii is much greater than in
Oe. pratincola, and is so marked that otherwise indistinguishable in-
dividuals, the one stable, the other mass-mutant, can easily be dis-
tinguished by an examination of the seeds.

Without going into detailed repetition of data published in the
former paper on Oe. Reynoldsi1, it may be recalled that the wild form
of the species, f. typica, has given rise to the derivatives mut. semzalia,
mut. debilis, and mut. bilonga. The f. typica is remarkable in that it
exists in two morphologically identical phases, one of which is relatively
stable, whereas the other is mass-mutant, giving rise to polymorphic
progenies containing all of the mutations enumerated, as well as others
which have not yet been carefully examined. Mut. semzalta was so
named because the plants of the early cultures, grown in Maryland,
were about half as high as f. typica. ‘The cultures of the season of
1916, grown in Michigan under other environmental conditions, did
not show so great a disparity in height, but in other respects the forms
were no less distinct than before. The shape of f. typica is depressed-
conical, because of the long, widely spreading lower branches, whereas
mut. semzalta has relatively erect lower branches and is therefore some-
what cylindrical rather than conical in shape. Mut. debilis is a weak
dwarf with much reduced foliage. Mut. bilonga was so named because
its fruits are twice as long as those of mut. semialta, which it closely
resembles in form and stature. In other respects, however, it will
be shown that mut. bilonga more closely resembles mut. debilis, from
which it springs, than mut. semialta.

All the mutations come true from seed, except that mut. semialta
is capable of giving rise to mut. debilis, and that the latter may in
turn give rise to mut. bilonga. Mut. semzalta has once thrown a
mutation which will be known as mut. rigida. It came entirely true
in a large progeny grown in 1916, and will receive a larger share of
attention in a future paper. A few other types have appeared in the
cultures, but it has not been possible to obtain seeds from them.

SUMMARY OF THE CULTURES

Figure I is a chart giving the pedigree of all the cultures of Oe.
Reynolds and its mutations that have thus far been grown from self-

122 CARL D. LA “RUEVAND Ens BARTLETT

pollinated seeds. Each progeny represented in the chart has a key
number which serves to identify it with the detailed analysis of the
same progeny in Table I. Several of the earlier progenies were not as

. @,

n®

"3 00 |

» OOONODGBE)
. ODO00 €

Fe PEO

_ Fic. 1. Pedigree of Oenothera Reynoldsi and its mutations. Each numbered
progeny is represented by a circle. T=typica. S=semialta. D=debilis. B=b1-
longa. R=rigida. T,=uniform culture of typica. T,=polymorphic culture con-
taining typica. Letters on the lines leading to circles indicate the parentage of the

cultures. A star (*) indicates a plant used as a parent for the crosses referred to in
Table II.

large as seemed desirable, on which account supplementary cultures
were in several cases grown a year or two later from any seeds that
had been left over. Such division of progenies between two seasons
has provided a very desirable check on the classification of the plants,

CO
N
a

MATROCLINIC INHERITANCE

0} sWIO0d YJOg

“SULIOMOP JO oUIT}
ye squeyd ysnqos as0ul sonpod asojyasay} pue ‘198u0] a8e}s 9930S01 DY} UI ALIS S[eNPIAIpuT [erUUATG dq} INq ‘WoOsvas oUO UT AyIN}eUL
‘jeruuaIq 10 JenuUe Jay} ‘YIMOIS Jo WqGey UT A[UO SuTIoyIp ‘sassejo OM} OZ! 9prArp syenprArpur vI2daq oy, 4
‘Zz1 ‘d 90S ‘sjueyd users AjUO spnjour JUso1ed puke JaquINU UOT}CUIUNIO) ,

OSI OSI ,%0:Sc | voi oSZ D8U011q zz—€S$ C68 Te a gz
I 6g1 ' os ‘ OOF 4) Ve 14908 S1g D8U0]1q ¢S—C—6g Vy a oz
a I cV : a a4 ae a “a syiqap $z—C—68 Vy as | vz
- GZE gs g a C2e 1,0, 0Ve one. MV QOsT DEUOTILG Cg—1z—£-68 Vy ayo Ie
5 I vg os Sg a Je oe supigap 1z7—t—68 Ty Wy ZZ
re v : £Q oe 19 : : ee DqqDIWMas bz—¢—69 ly ry IZ
= S$ Iv Be ov 3 a e DyyDIMas z~z—£-68 Vy as 0z
oe z 09 < es z9 Sez, rae) zlQ sujiqap be I—-L 1-68 Vy oq 61
; ; ve 00z o 00z OzL.| ogf'1 | 169'T pists | Ly—V—-C1-C—6g Ty J | QI
He OOI oOoI %29 |} L6S z96 nyyvimas | QiI-v—C1—-t—6g oy a | Li
Sj I LV : oS %L°38| gS 9S9 DyvIMas b-C1—€—68 ly oy gI
I ¢ He C1 Se Oz %eV | Oc clv paigky | 6L-S1-C1-€--6g fs 97 SI
z Re 7s G 16 CI cre Ve I Z19 pargky | oS-S1-C1-€—-6g é 7 VI
v I CI oS Zo 06 %9'V z6 000‘z DI1g K4 Ci1-C 1-€—-6g oq or
z II ce el 19 iS v9 p91gX4 V1i—-C1—-L—6g ; as | ZI
ee ° O °
ee ee vod £1-£-65 tcs| on
i ; : a OOI | OOI a p91gX4 b—C—68 ‘ a | OI
I QI Te 6z 6L BY : pga} €—698 : a 6
; : a eA OOI | OOI 7,65 | 2ge Dg X4 ob—Sg—68 as | 8
: : ae OOI | OOI OV eet Daud X4 C968 se a Z,
I ; 22 1486 66 OFT .80S Dg 4 1-€S—68 a | 9
on Se OOL | OOI %z2g| 617 Doug 4 ¢S—6g a ¢
I ; a vor | Sor a ue pag K4 68 . oy v
ae : : Olio OT Fee : pag ky o=16 ey ¢
I : : s €z vz oP: oy p14 4 16 i ey z
a OI OI a] as DIG K4 a oe Ty I
Aqiny % “ON uoneNL | VeIsIpag (13
enon g D210 BIN 0} poeiuelg quoi1Vd Jusied jo 90ue jo uol} I°stq
en [/ DGUO7I G 2.62 20) eB -Ie9 5 )
a 722 \S2719aCT | Vplsiy -21aS 22g [, Ee eee spseg | jo AyQuepy jo ON aanqng Se aes ene
sjuR[g | RS -elouer) | -e1aues)

SS

GIOL fo uospay ay] surpnjrUT puDv o7 qn ‘sassory Buipnpyxy ‘suoyninyy si] puy uspjoukay vaayjouag fo aaasipag ay} fo Kapmuns

I A1dV I

T24 CARL D. LA RUE AND dH. H. BARTLETT

but has brought it about that the different lines composing the cultures
have not all been carried the same number of generations since the
foundation of the pedigree in 1910. ‘The oldest lines have now been
carried through six generations by self-pollination.

At the close of the season of 1914 the cultures of Oe. Reynoldsui had
been maintained for four generations, and the pedigree, as summarized
in the former paper,® showed clearly that the individuals of f. typica,
externally alike, were of two kinds, giving rise, respectively, to uniform
and polymorphic progenies. It had been shown, also, that both kinds
of f. ¢ypica occurred in polymorphic progenies. One point, however,
was still obscure. The original mass-mutant individual of f. typica
had been the only plant of its generation self-pollinated for continuing
the line, and it was therefore uncertain whether the sister plants of
the same culture would have resembled it in giving rise to polymorphic
progenies or would have given uniform progenies. In other words,
was the original mass-mutant individual of f. typica itself of the nature
of a rare physiological mutation? In order to answer this question it
was necessary to go back to old seeds and retrace two generations.
This has been done, with the result that two sister plants of the original
mass-mutant individual have given two generations of uniform progeny.
(It should be noted that here, as well as in the explanation of the
pedigree, Figure I, a progeny is for convenience termed uniform, as
opposed to polymorphic, even if it contains a few mutations, provided
the mutability was unaccompanied by unusual seed sterility. The
sporadic mutations that have appeared in so-called uniform progenies
have in no case been those characteristic of mass mutability. No
confusion can possibly arise from this terminology for the reason that
Table I gives a detailed analysis of every culture concerned in the
experiments.) It is therefore possible to conclude that mutating and
non-mutating individuals of f. typica may occur together in either
uniform or polymorphic progenies. In the former case the mutating
individual must itself be regarded as a physiological mutation, or
perhaps even as a premutation in the sense of De Vries.®

Premutation, according to De Vries, is the process of preparation
for mutation. In forms showing ordinary mutability the various
mutational types occur in every sufficiently large progeny in every
generation, and the process of premutation must therefore be assumed
to have taken place far back, and to have brought about a hereditary

8 De Vries, Gruppenweise Artbildung, pp. 9 and I0, 335, 346.

MATROCLINIC INHERITANCE 125

change by virtue of which all individuals of the line became mutable.
The phenomena are not quite comparable in the case of mass-mutant
Oe. Reynoldsii, and lead one to wonder if the change in the genetic
physiology of the original individual of mass-mutant Oe. Reynoldsiu
may not have been a premutation accidentally detected at the actual
time of origin. Speculation on such a point, however, will hardly be
worth while until the investigations shall have been pushed much
further than they have been as yet.

To those who may desire to explain the mutations of Oe. Reynoldsu
on a Mendelian basis the facts are very refractory. The lines have
been grown from guarded seeds since 1911, and have probably been
self-pollinated much longer, for the species is one of the smaller-flowered
self-pollinating types, producing abundant pollen that is liberated on
the receptive stigma a day, or even two days, before the flowers open.
If the wild parent plant had been an Fy hybrid, or a heterozygote of a
later generation, the first generation in the garden should have shown
segregation, whereas the first polymorphic progeny was obtained two
generations later. An explanation based on the multiple factor hy-
pothesis is blocked by the fact that the mutations do not act as Men-
delian recessives, but show strict matroclinic inheritance when crossed
with the parent type.

SEED STERILITY OF THE MaAss—MutTAnT INDIVIDUALS

Returning to the problem presented by the two types of individuals
of f. tybica, we see from Table I that there is at least one character
by which they may be distinguished. Ali those plants giving rise to
uniform progenies have reasonably good seeds, relatively many of
which (58 to 84 percent) readily germinate. Those giving rise to
polymorphic progenies, on the contrary, have very poor seeds, few
of which (2 to 5 percent) are capable of germinating.

When the seeds for two of the polymorphic progenies (Nos. 11)
and 12 in Table I) were counted off it was found that only about 5
percent of the seed-like structures were actually perfect seeds with a
good embryo. The remainder were either empty shells, or else con-
tained a small amount of yellowish disintegrated tissue. Many were
examined. It is therefore certain that the low germinability of the
seeds that yield polymorphic cultures is not to be attributed to delayed
germination. If a cytological study now in progress throws any
light on the reason for the seed sterility, we may be well on the way

126 CARL D. LA RUE AND H.-H. BARTLETT

to an understanding of mass mutation. Certainly there is likely to be
some causal relationship between such closely associated phenomena.

One must of course take into consideration the possibility that the
defective seeds represent zygotes of f. typica that failed to develop.
Reckoned from the total number of seed-like structures sown, rather
than from the number of plants obtained from them, the proportion
of mutations in the polymorphic progenies would not be at all unusual.
We are not inclined to believe, however, that any such explanation is
the right one. Why should the typica zygotes in one case develop
into uniformly strong and viable embryos, but in another case, en-
vironmental conditions remaining the same, fail to produce even
mature embryos? Moreover, if there were no essential difference
between uniform and polymorphic progenies other than the failure
of typica zygotes to develop, why should the mutations found in the
polymorphic progenies be characteristic of the latter? It may be
urged that the evidence is not sufficiently clear that the non-mass-
mutant individuals might not throw mutations semialta, debilis and
bilonga if grown in sufficiently large cultures. For the present it
must suffice to say that they have not done so, although we are keenly
aware of the fact that the cultures have not been as large as one would
wish for convincing evidence on this point. Very much larger cultures
to test this question are planned for next year. It should be remarked
that the mutations of Oe. Reynoldsii are not sufficiently characteristic
in youth to admit of accurate classification, and that consequently
every plant of each culture must be carried to maturity if it is to be
certainly identified. With most of the other mutable species it is
possible to discard many of the typical individuals which make up the
bulk of the cultures without giving them garden space, since the
young plants are as easily distinguished as the mature ones.

In marked contrast with Oe. Reynoldsii, all individuals of f. typica
in the mass-mutant strain of Oe. pratincola seem capable of throwing
the mutations characteristic of mass mutation in that species, and such
individuals differ among themselves as widely as possible in degree of
mutability. Moreover, in Oe. pratincola the number of abortive seeds
seems to vary in approximately inverse proportion to the number of
typica individuals obtained from the seeds. ‘This fact might beadduced
as an argument for considering the bad seeds as resulting from the
abortion of typica zygotes. Wedo not wish to minimize this possibility
but prefer for the present the hypothesis that the zygotes which fail
to develop represent mutational types of excessively weak constitution.

MATROCLINIC INHERITANCE 27

It appears at present that mass mutation in Oe. Reynoldsw differs
considerably from the similar process in Oe. pratincola, the chief dif-
ference being that in the former species the characteristic mutations
are produced only by certain individuals of f. typica in which there is
great seed sterility, whereas in the latter species any individual of
f. typica belonging to the mass mutant strain may give rise to the
characteristic mutations, the mutable individuals differing widely
among themselves in mutability and seed sterility. The process is
alike in both species in that the characteristic mutations occur only
in strains some members of which are excessively mutable (7. e.,
mass-mutant) and in that the characteristic mutations in both cases
show matroclinic inheritance.

Before turning to the evidence in regard to matroclinic inheritance
there is a further feature of seed abortion to which attention should
be called. The germination data in Table I show clearly that muta-
tions arising trom highly infertile mass-mutant f. typica are not them-
selves excessively infertile. The degree of seed abortion is not nearly
as great in the mutations as in the parent plant that produced them.
Seeds of mass-mutant f. typica have given germinations varying from
2.3 to 5.1 percent. In striking contrast to this low viability, seeds
of mut. semzalta have given germinations of 88.7 percent and 62.0
percent; seeds of mut. rigida, 72.0 percent. The germinations recorded
for mut. bilonga are much lower than the true value, because only
green plants that survived were counted. This mutation has the
curious characteristic of giving rise to progenies consisting of a mixture
of green and yellow plants. The latter lack the capacity for chloro-
phyll production, and die shortly after the cotyledons unfold. The
relative numbers of green and yellow plants have not yet been deter-
mined. Leaving yellow plants out of consideration, mut. bilonga has
given progenies numbering 34.6 percent, 24.0 percent, and 25.9 percent
of the number of seeds sown—well in excess of the viability of mass-
mutant f. typica. Complete records have been kept for only one
progeny of mut. debilis. This form is a weak dwarf, of which the
seeds are much less viable than those of the other mutations. Only
7-5 percent of germinations were obtained. It must be remembered,
however, that mut. debilis, the most sterile of the mutations, gives
rise to mut. bilonga, a form showing a distinct increase in fertility over
its parent. Wecan not doubt that in the case of the mutations seed
sterility is in a large measure inversely proportional to the vegetative
vigor of the parent plant.

128 CARL D. LA RUE AND H. H. BARTLETT

The same explanation does not hold for the difference between the
two kinds of f. typica, for vegetatively they are equally vigorous.
May not the yellow seedlings which occur in progenies of mut. bilonga
give a clew to an understanding of the situation? These yellow seed-
lings constitute a mutational type in which chlorophyll formation
can not take place, and therefore a type which can not persist more than
a few days after germination. It does not require a very great effort
of the imagination to conceive of physiological defects that might
originate by mutation and that might operate disadvantageously to
the organism possessing them at an even earlier stage in the life cycle
than failure to produce chlorophyll. May not the aborting seeds in
the polymorphic progenies represent one or more physiologically
defective classes of mutations, of which the zygotes are unable to
develop into mature embryos? Pending cytological study of the
abortive seeds, such a hypothesis seems to us much more plausible
than the alternative hypothesis that they are typica zygotes, eliminated
by some unknown selective process that leaves to develop the intrin-
sically weaker zygotes of the several mutational types.

MATROCLINIC INHERITANCE IN THE MUTATION CROSSES

In 1915 a complete series of mutation crosses was made, involving
f. typica and the three well-known mutations. One parent plant of
each form served for self-pollination and for crossing with the three
other forms. Each cross was made reciprocally. Two of the twelve
crosses, mut. semialtaX mut. debilis and mut. bilongaX mut. debilis,
failed, but the remaining ten were in varying degrees successful, and
progenies of all were grown in 1916. ‘The reader will find the four
parent plants of these crosses indicated by asterisks in figure I, and
may determine by reference to Table I that all gave rise to uniform
progenies in the following generation. It will be observed that the
phenomenon of mass mutation had not occurred in the direct line of
descent of the individual of f. typica chosen as a parent. Both-the
semialta and the bilonga parents belonged to first generation progenies
from primary mutations (7. e., mutations derived directly from f.
typica, and not from one of the other mutations). The former type
arises only as a primary mutation, but the latter is frequently derived
as a secondary mutation from mut. debilis. The individual of mut.
debilis used as a parent was an actual primary mutation in a poly-
morphic progeny, chosen because, at the time the other plants were in

MATROCLINIC INHERITANCE I29

condition for crossing, the uniform first generation culture of mut.
debilis did not contain a single plant on which enough flowers remained
to suffice for all of the crosses. The detailed analysis of the mutation
crosses is given in Table IT.

In brief, the results of the mutation crosses are as follows:

typica X semialta — typica

typica < debilis > typica

typica X bilonga > typica + yellow twin

semialta X typica > semialta

semialta X bilonga — semialta

debilis X typica — debilis

debilis X semialta — debilis

debilis X bilonga > debilis + bilonga

bilonga X typica > bilonga + yellow twin

bilonga X semialta > bilonga ++ yellow twin
With one exception the scheme of inheritance is strictly matroclinic.
The type of pollen used is immaterial, providing it does not come from
mut. bilonga. All progenies which did not have mut. bilonga as the
pollen parent were exactly the same as they would have been if the
mother plant had been self-pollinated. The fact has already been
mentioned that progenies of self-pollinated mut. bilonga consist of a
mixture of green and yellow plants. Every cross into which mut.
bilonga entered as the pistillate parent showed exactly the same
mixture of green and yellow plants, of which the former developed as
normal mut. dilonga and the latter died. It is obvious, however, that
the crosses with mut. bilonga as pollen parent constitute a real ex-
ception to the prevalence of matroclinic inheritance in the mutation
crosses. .

In the case of mut. debilis X mut. bilonga the progeny contained
both the maternal and the paternal types, the latter in such large
numbers that it was not possible to view them as having arisen de novo
by mutation from debilis eggs. Thus the progeny from the cross
contained 18 plants of mut. bilonga out of a total of 47 plants. By
way of contrast, the progeny of the pistillate parent, mut. debilis,
self-pollinated, included only two individuals of mut. bilonga in a
total of 62. Mut. bilonga was therefore roughly twelve times as
frequent in the cross as in the progeny resulting from self-pollination,
—a difference that one must ascribe to the pollen parent. Incident-
ally, it seems worth while to call attention to the fact, without attempt-

BARTLETT

ee

LA RUE AND H.

CARL D.

130

“Ajuo sqzueyd us018
dy} JO JUNOIL soyL} I1OJasoYy} PJOeI UOIJeUTUTIDS BY, “payuNod jou 1am pue Aypides Yo porp syurjd Moped oy, “Stoquinu
enba ynoqe ut Apwuaredde ‘syuejd Mo]jaA pue use13 Jo Auasoid pax e& 9AeS DsuOIIg “NU K VIIGK] “J SSOID BY} JO SPIVC x

, AUP IN
({hi#)1

SUCHEN lpsuoj7g

19410

S729 2

AyINye yA
0}
poeureyay

Woot | oi
Vea 109
%,0'0S..|- 2&
Ve | VS
Tk VI
CORG | alee
Oe OG) “eee
Ue Ola cok
%S'og | gsv
%e'vg | 198
0b "ON

UOT} eUIUL Ia)

10S
C1
Col
66
eI
gl
£6S
ore
69S
1c0'1

poejurlg
Sp22S

g1—b-£ 1-¢—-68
1-£S—6¢
zz—L$-£-68
gi—v—£ 1-£-68
I-¢S—68
z7z—€S$-L-68
I-€S—68
z7z7—£$-L—-68
bl i—-—CL1-L-6g

91—b-—L1-£-6g

qusleg

zz—CS-C-68
z7z—CS-C—-68
bl I 1-€—6g
PlI-C1-€ -68
vl1—-C1-L—-68
g1—-b—-£ 1-¢-68
gI—-v-¢ 1-¢—68
1-€¢—6g
I-¢S—698
I-¢S—6¢

jusie~g 6

DYDIWMAS K D8U0ILG
pa1gky * D8u0jrq
DaUojrg X $2rqap

DyyvIwWas X sriqap
pagay X SYyiqap

DsuojLg XK vyjvwmas

DILg KY XK DyyoIMas
DBUuojig K 191g K4
syiqap X 1924}
DYpIMas K DI1G KY

SSOID)

T QD, ut punoy aq Koy uoynuryod-fjas Kq may], worl paaisag Sawuasorg jo uoyrsodwmo) ay] pup
‘T Bh UL Systaisp KG paywIvpUT AAD Sassod] ayy JO SJUAID AY] “QIOI Ut UmOAD ‘sassodg uoyninyy fo sadnjngy ayy fo Kapmungy

I] A314vVL

MATROCLINIC INHERITANCE I31

ing to draw any conclusions from it, that seeds of mut. debilis were
much less viable than those of the crosses into which this form entered
as the pistillate parent.

The progeny obtained from the cross f. typica X mut. bilonga
showed the influence of the pollen parent in the large number of yellow
plants, the same, as far as superficial observation could indicate, as
the yellow twin that appears in progenies from self-pollinated mut.
bilonga. In view of the fact that the crosses typica X bilonga and
debilis X bilonga both resulted in twin types, it is interesting that the
third cross, semialia * bilonga, gave only plants of the maternal type,
aside from a few which appear to be satisfactorily accounted for as
derived from mutated gametes.

It will be observed from a scrutiny of Table II that in the foregoing
discussion we have tacitly assumed that the sporadic occurrence of
types in cultures where they would not necessarily be expected to
occur was to be ascribed to mutation. We have made no special
comment, for example, on the few individuals of mut. debilis that turned
up in the progeny of semialia X typica, for the reason that mut.
semialta always seems to produce some mutated gametes that give
rise on fertilization to mut. debilis. Furthermore, there is every ground
for the belief, on evidence furnished by the matroclinic progeny of the
cross debilis X typica, that the few debilis-yielding eggs of semzialta
would give rise to mut. debilis quite regardless of the source of the
male gametes, provided, of course, that the latter were not derived
from mut. bilonga, the one form of the series that,seems to give rise
to more than one type of sperms.

To return to the case of the cross semialia < bilonga, we have seen
that mut. semialta always produces some debilis-yielding eggs, and
that the cross debilis X bilonga yields a progeny containing both
parental types. Consequently we should expect that in the cross
semialta X bilonga some of the mutated eggs would give rise to mut.
debits and some to mut. bilonga, whereas in a progeny resulting from
self-pollination or from one of the other crosses the mutated eggs would
be represented by mut. debilis alone. The results of the crosses realize
this expectation. Although mut. semialia, whether self-pollinated or
crossed, has always given rise to mut. debilis, it has never given mut.
bilonga except in the case of the cross semialta x bilonga.

The results of the whole series of cultures are intelligible on the
supposition that Oe. Reynoldsii is one of the mutable species to which

132 CARL D. LA RUE AND H. H: BARTLETT

De Vries? would apply the term heterogamous. It has frequently
been found that crosses of the Oenotheras differ strikingly according
to the direction in which the cross is made. Often the reciprocal
hybrids from the same two parent plants are as unlike as the parents
themselves. De Vries has attributed such results to a difference in the
hereditary qualities of the male and female gametes, and has suggested
the term heterogamy for the condition of species in which such a
differentiation of gametes is found. There is much unpublished evi-
dence at hand which tends to show that heterogamy may exist in
some species without a sharp restriction of either type of gamete to
the eggs or sperms and on this account we shall use the term ‘‘heter-
ogamy’’ with no implication that the non-equivalent gametes may
not exist on both the male and female sides. The conception of
heterogamy so modified as to apply to results that have been obtained
in our experiments has been published’? in advance of the data which
suggested the modification.

Let us assume (1) that a heterogamous species such as Oe. Reynold-
sit normally produces two types of non-equivalent gametes, which
may be designated as @ and 8 respectively; (2) that the a gametes
carry most of the characters by which specific differentiation is effected ;
(3) that mutation occurs through the modification of a gametes, which
thus become a’, a’, a’’’, etc. Applying this conception to the par-
ticular case in hand, let us think of f. typica as the zygote af, mut.
semialta as a’B, mut. debilis as a’’B, and mut. bilonga as a’’’B. The
conditions imposed by the results of the various crosses are satisfied if

f. typica > a eggs + B sperms,

mut. semialia > a’ eggs + B sperms,

mut. debilis > a’’ and B eggs + 6 sperms,

mut. bilonga > a” and B eggs + a’” and 6 sperms.
Since the various forms are determined by the a gamete, all mutation
crosses must of necessity show matroclinic inheritance, except those
involving mut. bilonga, for this one form is the only member of the
series that produces any male a gametes. In order to be functional,
male a gametes must fuse with female 8 gametes, which are produced
only by mut. debilis. Therefore the cross debilis X bilonga is the
only one that yields both the maternal and paternal types. It will

® De Vries, Gruppenweise Artbildung, pp. 30-32.

10 Bartlett, H. H., The Status of the Mutation Theory, with Especial Reference
to Oenothera, Amer. Nat., 50: 513-529. 1916.

MATROCLINIC INHERITANCE 133

be remembered that mut. debdilis, when self-pollinated, was marked by
great seed sterility. This sterility was much reduced when pollen
from one of the other forms was used, and the effect of foreign pollen
was greatest of all when that of mut. bilonga was used. Doubtless
several factors are concerned with the increase of fertility on crossing,
but it seems not unwarranted to call attention to the fact that if our
hypothesis were true such an increase would be expected, because good
embryos would result from the fertilization of female 6 gametes by
male a gametes. All the bilonga individuals in the mixed progeny
from debilis X bilonga would be represented in a self-pollinated prog-
eny by aborted seeds.

On the whole, the facts point to the truth of the hypothesis of non-
equivalent gametes. The facts to be explained are sufficiently orderly
to demand more than a superficial criticism at the hands of those who
see in the mutation phenomena merely evidence of Mendelian segre-
gation. It seems to the writers that the work with Oe. Reynoldsiu
affords very convincing evidence of De Vriesian mutation.

QUANTITATIVE EVIDENCE OF MATROCLINIC INHERITANCE

Although no one who has had an opportunity to examine the
mutation crosses has doubted the fact of matroclinic inheritance, it
was of course essential to obtain quantitative data that would con-
vince one of the accuracy of our observations. Leaves and capsules
from self-pollinated and crossed progenies were therefore measured,
both in order to establish the fact that the several forms differed widely
from one another and to show that the mutation crosses resembled the
pistillate parent. In most cases a large enough number of plants was
at hand to give satisfactory data.

Mature stem leaves were measured from plants of all the pure
strains and mutation crosses, five leaves being taken at the same part of
the main stem from each plant as it came in the row, without selection.
The leaf lengths are summarized in Table III, the widths in Table IV.
The two tables are based upon the same material, but individual
leaves were frequently imperfect, so that one or the other measurement
could not be made. On this account the number of measurements
does not always tally in the two tables. It is very clear that the
modes of the variation curves lie very close together in the cases of all
progenies having the same pistillate parent. There are some dis-
crepancies, to be sure, the most notable being the failure of a closer

BARTLETT

H.

LA RUE AND H.

CARL D.

134

I
6 I Cc Q CC
Cee |r SECS che lor]
LY $9 Soe evs Va
: a6 : z6 O11 GL | $1] gg
; : I ¢ ne gl CL SQra Oc CoE
: 6 Cz ay Iv ce bL | 39) 9F
; ve Iv o LY sik oz | ov! 61
: : £6 €g VI 3 I z Lane
O7zI ol ce I eed aaa
9 CI ¢ gII zS €Z I Z 9
QI 89 Qc I 6 : 99 ZZ 6S ot eos
O1l.-|) SZI- |-.02 of 61 I LY C1 60 : :
gir | SZ goI | O11 cv OI Le Iz ase ic :
c¢ ge CV 69 If cl, fz TOin ae ; Zz Pe
6 ¢ I OI 1 88 S9 II 09 ER ; ;
ay : oe I VI ran Qz ell
me i Qz I
. ase _ ae
8 °ON
D DSU0ILG | S2719a ou} hued
2 SE oN | 92.°0ON 122 | S222 | eg “ON | Z€ on | 6r ony | TE “ON | O€ sony | Zr ‘on | 62 -on | gz ‘on | Zz “on |'ON -01g
vig | wsuoz D1IV vg $2729 psuoz vI1g DID DSUOT $2729 V4I0
“AL X | 2gxX -2ulasX| AL X | -ag X | -7g K | -AE XK |-tmeas X|. -2g XK | -ag X |-20as X
9£ ‘ON vE-ON pager x
vq7D1muLas X weUuojzrg * ee L
vsUuojlg WOoryewyl §2727997 Wolye ny DI]VIMAS UOeIN]A vI1iga gt, “ay

61z—O1z
60z7—00z
661-061
68I-Og1
6L1-OL1
691-091
6S1-oS1
6v1-ot1
6C1-Of1
6zZI-OZI
61I-OII
60I—OO1
66-06
62-08
6L-oL£
69-09
6S—oS
6t—ob
6£—of£

WI
url Yjsua'T

SASSOL) UOYDIN, pun suoyninpy sir ‘uspjoukay *acQ fo syysuaT f{oaT

Ill STavy

135

MATROCLINIC INHERITANCE

Zz I LS—9S
9 I 3 cc—rs¢
‘ 6 3 ZI ¢o-—z¢
I EG, 1z Le cite I 1$—oS
v oS vs UA Saag Re ab ies 6b—9P
I cL fe) 14 88 Cv | o1| 72 Lv—9v
c LI 0) L9 36 cS | 61; IZ Cv—bb
Li QZ O CY LE ZQ | 7€| 99 Cv—zv
vz CY e 6z VI ov -| Zz| ve 1v—ov
LE S9 61 ¢ av QI He|. Ot 6£-g¢
£9 LV IZ e 9 C1; ¢ LE-9¢
Q8 ze 6¢ I v Tr) av. Co_-Pe
6g of gt I fe) 4 oh er
LQ Cr 1S 3 I I 9 1¢-of
S I I c¢ l rag re I Mee 6z-gz
I ¢ ZI Z z 6¢ ¢ of I Lz-9z
61 oS ge CI 6 Li c CI _ Cz—be
gL L6 6 Qc Zz I : Zz ZI se ¢ €7-Zz
III £6 L6 64 SV or ; 8 I : I1c—0@
ol ss 60 z$ Ee, of 9 9 gt fe 61-gI
Iz GI 2 ze S v9 ge ec can : LI-91
Z 0 I 8 go tebe | ome Gu G1-b1
; I : ¢ gI C1 C1 61 C1-ZI
. ee . ee Zz Z ee ee I I-—OI
Z 2772 Ag
2 ss SE ‘on | oz ‘oN wsucned | $1199 €€ ‘on | z€ on | 6x ‘on | TE ton | OF “on | Zr ON | 62 “ON | Bz “ON | 42 °ON “ON Ku301g
ae OR IES SA prs PL RON AU WEE VeRO ae SO vt) se “ar
eee ; apie x Lae Z “ ae eee ee o ee wI21gay X Ur WPI
vsuopig “WIN 527299 “WIN vygvIULaS “NIN vga, A

SASSOAD UOYDINT pun suoyninypy siq ‘uspjoukay ‘acd fo syipryy, {0aT
AI a14VL

136 CARL D. LA RUE AND H. H. BARTLETT

agreement between semialta X semialta and semialta X typica, and,
conversely, the unexpected closeness of the modes for semialia X typica
and typica X semialia. It is believed, however, that the departures
from the expected are all explained by environmental non-uniformity.
The garden itself was relatively uniform, but the cultures were set
out at different times, so that some of the plants were subjected to
hot dry weather much sooner than others. The most rapid growth
took place early in the season, with the result that the last plants to
be transplanted were markedly the weakest. Without exception,
the greater deviations from the measurements which would have been
expected in matroclinic inheritance were correlated with the greater
intervals between time of transplanting, and, conversely, the best
agreements with expectation were found in the cases of cultures set
out on the same day.

A valuable evidence of shifting of the mode attributable to dif-
ference in date of transplanting was quite accidentally obtained in the
case of the cross bilonga X semialta. The culture had been partly
transplanted at the close of the day, and the next morning the remainder
was overlooked. It was set out a couple of weeks later. Leaves
from the two lots were collected separately, and the data are given
separately in Tables III and IV.

Considering the unfavorable experimental conditions, the measure-
ments provide as good a demonstration of matroclinic inheritance as
could reasonably be demanded. Moreover, it would be unfair not to
emphasize the fact that in the one or two cases where the measurements
might appear ambiguous, the plants were in general aspect true to the
expected type. Thus the cross semialia * typica had the whole ap-
pearance of self-pollinated semialta, from which it differed only in
being much more robust. Although we are of the opinion that en-
vironmental non-uniformity explains the difference, we shall carry
out more carefully controlled experiments to see if cross-pollination
produces any effect similar to the vigor of heterozygosis. Such an
effect might conceivably be measurable if environmental non-uni-
formity were eliminated, but if it exists it is obviously not large enough
to obscure the underlying phenomenon of matroclinic inheritance.

The data for capsule length in the mutation crosses are incomplete
because of the fact that some of the cultures of 1916 were just coming
into flower at the time of the first heavy frost and were destroyed.
Before examining the data for the crosses, the reader should glance at

MATROCLINIC INHERITANCE 137

Table V, which shows the range of variation in capsule length in each
of the four pure forms. The measurements were made in Maryland
in 1915. In the cases of f. typica and mut. semialta the ten lowest
normal capsules of the main inflorescence of each plant were used.
The capsules of mut. debilis were taken from the debilis plants of a
polymorphic progeny. They were collected at random, because very
few inflorescences bore as many as ten good fruits. Capsules of mut.
bilonga were measured from two different cultures, in order to demon-
strate the essential identity of this mutation, whether obtained directly
from f. typica, or as a secondary mutation from mut. debilis. The
measurements prove the anticipated identity, or, if anything, give a
false impression that the secondary mutation is stronger than the
primary. This impression is due to the fact that at the time the
measurements were made the secondarily derived mut. bilonga had
been so short a time in flower that ten full-grown capsules could not
be obtained from each inflorescence. Six fruits were therefore taken
from each, and the greater average length which they show in com-
parison with primarily derived mut. bilonga is due to their lower
position in the inflorescence. The progeny of primary mut. dilonga
was arbitrarily divided into a class of weak plants and a class of strong.
The capsules of these two classes, ten from each plant, were separately
measured, and the results are recorded in Table V for each class separ-
ately and for the two classes combined. ‘The discrepancy between
the two classes was much less than was expected, and indicated
clearly that the capsules respond less to environmental conditions
than vegetative parts of the plant. Measurements of any other part
would have shown a much more marked difference between the
arbitrarily selected strong and weak plants.

On account of its relative independence of environmental factors
the capsule length affords us a more conservative criterion of matro-
clinic inheritance than the leaf measurements which have already
been considered. In this respect it seems to be similar to the character
of flower size, which East!! has found especially useful in his studies of
inheritance of quantitative characters in Nicotiana. In Nicotiana
sylvestris he found that adverse environmental conditions, which
brought about a shortening of the leaf amounting to three fourths of
its normal length, did not appreciably change the size of the flowers.

“East, E. M., Inheritance of Flower Size in Crosses between Species of Nico-
tiana, Bot. Gaz. 55: 177-188. 1913.

138

CARL D. LA

RUE AND H. H. BARTLETT

TABLE V
Length of Capsule in Oe. Reynolds and its Mutations

Length in Mm.

14-15
LO-t 7
18-19
20-21
22-23
24-25
26-27
28-29
30-31
327-33
34735
36-37
38-39
40-41
42-43
44-45
40-47
48-49
50-51
0700
54-55
50-57
58-59
60-61
62-63
64-65
66-67
68-69
70-71
72-73
74-75

Goodspeed and Clausen’? have questioned East’s conclusions,

Typica
Progeny
No.7

Sentalta
No. 16

Debilis
No. 11

Bilonga
from
Debilis
No. 23

Bilonga
from
Typica:
Weak
No. 25

Bilonga
from
Typica:
Strong
No. 25

Bilonga
from

and

have demonstrated beyond question that the flower size does respond
to environmental changes, but their results really strengthen the
contention, that, as compared with changes in other parts, the flowers
are relatively little affected. Our own conclusion in regard to the
fruits of Oenothera is that they respond in size to varying environment,
but that the response is relatively much less than the response in height
of plant or size of leaves, and that the character of capsule length is
particularly significant as a criterion of matroclinic inheritance.

2 Goodspeed, T. H., and Clausen, R. E., Factors Influencing Flower Size in
Nicotiana, with Special Reference to Questions of Inheritance, Amer. Journ. Bot. 2:

332-374.

IQI5.

MATROCLINIC INHERITANCE 139

TABLE VI

Capsule Lengths of Oe. Reynoldsi f. typica and mut. semialta, and of Some of their
Mutation Crosses

The capsules measured were the lowest five capsules from each of two secondary
inflorescences from each plant; the progeny numbers refer to Tables I and II.

F, Zypica Mut. Semzalta
Lengthin Mm.| X 7y/zca ; ee : verte ;

sr eee | x rete Oe | ee
f6—17 Concens Loni eae sree 2) 2 2
18-19 bedagr ae ee aaees 5 17 14
20-21 eee: Bie eae sere 30 59 Bar|
22-23 Gea eos Sey eke 56 99 27,
24-25 I eos 3 I 114 147 44
26-27 2 are 5 12 97 124 36
28-29 18 tas 23 II 86 84 29
30-31 44 4 | 46 28 45 31 10
32-33 79 16 i) 54 12 3 7
34735 100 39 87 71 2 4 2
36-37 86 61 99 gO Sao a I
38-39 48 92 82 110 at ais eee, sees
40-41 13 EG Di 43

42-43 6 50 14 26

44-45 2 41 6 14

46-47 pie 17, 5 4

48-49 Kt a 9 Sane I

50-51 cose 3 4 Sy ae ae

Table V brings out clearly the fact that the forms of Oe. Reynoldsii
differ distinctly from one another in capsule length. It is not, however,
strictly comparable with Table VI. In the first place, the garden of
1915 was in Maryland, where climatic, cultural and soil conditions
were unlike those in Michigan. In the second place, the early frost in
1916 overtook the plants before the inflorescence of the main stem was
sufficiently mature to provide full-grown capsules. Since the in-
florescences of the long basal branches had begun to flower several
days earlier, their lower capsules were full-grown. The five lower
capsules from two branches of each plant were measured. ‘Thus each
plant provided ten capsules, but they were from two lateral branches
rather than from the main stem. The capsules of the terminal in-
florescence of the main stem are usually slightly larger than any others,
and on this account the modes in Table V ought to be higher than in
Table VI, as indeed they are. Perhaps the difference is not as great
as it would have been if the capsules of the lateral branches had not

140. CARL D. LA RUE AND H. H. BARTLETT

Fic. 2. Inflorescences showing mature fruits of the mutation crosses semtalta
x bilonga (left) and its reciprocal, bilonga X semialta (right). Each cross is matro-
clinic, and therefore the lengths of the fruits are in the ratio I: 2.

MATROCLINIC INHERITANCE I4I

Fic. 3. The typica series of mutation crosses. The inflorescences are alike,
- resembling in each case the pistillate parent, f. typica. From left to right, f. typica
(self-pollinated), typica X semialta, typica X debilis, and typica X bilonga.

I42 CARL D. LA RUE AND H. H. BARTLETT

had a lower average position in the inflorescence, for the lower cap-
sules, if normally developed, are usually longer than those higher up.

Notwithstanding the difference in material, it is instructive to
observe the close agreement between the two sets of measurements.
The difference between the modal lengths for f. typica and mut. semi-
alta in 1915 (Maryland) was just 12 mm., the same as the average
difference between the typica series of cultures and the semzalta series
in 1916 (Michigan).

The capsule measurements give thoroughly satisfactory evidence
of matroclinic inheritance. The data for the semzalta series of cultures
are particularly convincing. The modal lengths are the same for
pure semialta, semialta X typica, and semuialta X bilonga, being 25
mm.in each case. The ratio of the capsule lengths of the three pollen
parents is 2: 3:4, but the self-pollinated mut. semzalia is just like the
two crosses. Turning to the slightly less consistent data for the
typica series, the evidence is hardly less satisfactory. In the cross
typica X debilis the capsule length is actually slightly higher than
in pure f. typica, in spite of the shorter capsule of the pollen parent.
In the cross typica X bilonga the length is slightly greater than in self-
pollinated f. typica, but that the difference is not significant is certain
from the fact that the cross typica X semialta has capsules just as
long. In the one case the pollen parent has a longer, in the other case
a shorter, capsule than the pistillate parent, but the crosses are iden-
tical. The results of the capsule measurements, taken all in all, prove
that matroclinic inheritance is the rule in the mutation crosses under
consideration, and leads us to suspect that there may be such a thing
as increased vigor due to cross-pollination, independently of factorial
recombinations such as those that occur in Mendelian inheritance.

In the case of the very interesting dimorphic culture resulting
from the cross debilis X bilonga, it is especially unfortunate that
capsule measurements were not obtained, because the two forms differ
so much from one another. That measurements would have fully
substantiated the conclusions already drawn in regard to this cross is
certain from the few precocious plants that matured before frost.

Figure 2 shows typical plants of the cross semzalta < bilonga and
its reciprocal. Of the latter there were too few plants that matured
to make a series of measurements worth while. The fact of matro-
clinic inheritance is obvious from the photograph, however, for in
semialta < bilonga the capsules are only half as long as in bilonga

MATROCLINIC INHERITANCE 143

R : Sie ns >:
ARPA AA tx hohe + kht er Pp a

Fic. 4. The semialta series of mutation crosses. From left to right, mut.
semtalta (self-pollinated), semialia X typica, and semialta X bilonga.

144 CARL D. LA RUE AND H. H. BARTLETT

x semialta. . Figures 3 and 4 illustrate the ¢ypica series and the semt-
alta series of crosses, respectively.

CONCLUSIONS

1. Mass mutation in Oenothera Reynoldsit consists in the production
of inordinate numbers of mutations, belonging to several characteristic
types, by certain mass-mutant individuals, which may be looked upon
as having undergone a premutative modification.

2. Aside from their mutability, these mass-mutant individuals
resemble normal f. typica. The production by them of a large number
of abortive seeds may itself be looked upon as one of the manifestations
of mutability.

3. The characteristic mutations form a series, each member of
which may give rise to the succeeding member. Thus:

mut. semialta > mut. debilis,
mut. debilis > mut. bilonga.

4. Mut. semialta and mut. debilis appear to represent successive
reduction stages in the mutation series. Mut. bilonga, on the con-
trary, marks an advance over the other members of the mutation
series and also over f. typica.

5. With the exception of crosses involving mut. bilonga, the muta-
tion crosses are matroclinic.

6. The cross debilis X bilonga gives a mixture of the two parental
types.

7. The facts of inheritance are best explained by the hypothesis
that two types of non-equivalent gametes, designated as a and 6
gametes, are normally produced.

8. The a gametes are usually eggs, and the 8 gametes sperms, but
mut. bilonga produces both a and 8 sperms.

9. Mutation in Oenothera Reynolds consists in the modification in
a gametes of factors that have no counterpart in the 8 gametes.

10. Since the sperms of f. typica are B gametes, mutations appear
whenever a mutated a@ gamete is fertilized. They do not appear as a
result of segregation.

DURATION OF LEAVES IN EVERGREENS
VINNIE A. PEASE

While the duration of leaves in evergreens is not at all a new subject,
very little systematic work seems to have been done toward determin-
ing durations for an extended list of evergreen species. This work was
begun for the purpose of determining the leaf duration of the evergreen
species of trees and shrubs in western Washington. It soon developed
that the work would not be a mere cataloging of species with their
accompanying leaf durations, since a very superficial examination of
some of the coniferous evergreens growing under different conditions
on the University campus, showed a wide but constant difference in
the duration of their leaves. It was then decided to limit the species
discussed to those growing under varying conditions that could be
examined in the field, and to ascertain, if possible, the factors governing
the duration of their leaves.

The Pacific northwest is peculiarly an evergreen region. Sargent
(15) described the characteristic coniferous forests as the most luxuriant
if not the most diversified on the continent. His report in the Tenth
Census states that ‘‘Washington is covered with the heaviest con-
tinuous belt of forest growth in the United States. This magnif-
icent coniferous forest extends over the slopes of the Cascade and
Coast ranges, and occupies the entire drift plain surrounding the waters
of Puget Sound.’’ Evergreenness is not only characteristic of the
forests, but is equally typical of the forest undergrowth, and of a large
list of herbaceous species of the open fields. This is especially true
of the Puget Sound region, in which the mild climate affords a practic-
ally continuous growing season. This may be one reason why many
species elsewhere deciduous are here evergreen.

There are, in the state of Washington, according to Frye and Rigg
(2), 76 species of woody evergreens, 24 of which are gymnosperms, and
52 angiosperms. In western Washington there are 52 species, 16 of
which are gymnosperms, and 36 angiosperms. Of these the writer
has studied the following 9 gymnosperms and 22 angiosperms:

145

146 VINNIE A. PEASE

GYMNOSPERMS
1. Abies grandis 6. Pseudotsuga taxifolia
2. Juniperus scopulorum 7. Taxus brevifolta
3. Picea sitchensis 8. Thuja plicata
4. Pinus contorta 9. Tsuga heterophylla
5. Pinus monticola

ANGIOSPERMS

(a) Transitional Forms.—Those species which are deciduous under
certain conditions and under others partly evergreen.
10. Rhamnus purshiana 11. Vaccinium parvifolium

(b) Sub-evergreens.—Those holding the leaves of one season only
until the leaves of the next season are able to carry on the photosyn-
thetic work of the plant. These species are not noticeably affected by
external conditions.

12. Arbutus menziesu * 16. Rubus laciniatus
13. Ceanothus velutinus 17. Rubus pedatus
14. Linnaea americana 18. Rubus ursinus

15. Micromeria douglasu

(c) True Evergreens—Species which usually hold their leaves
longer than the second season. These are noticeably affected by
external conditions.

19. Arctostaphylos tomentosa 26. Kalmia polifolia

20. Arctostaphylos uva-ursa 27. Ledum groendlandicum

21. Berbers aquifolia 28. Oxycoccus oxycoccus tnter-
22. Berberis nervosa medius

23. Chimaphila menziesi | 29. Pachistima myrsinttes

24. Chimaphila umbellata 30. Rhododendron californicum
25. Gaultheria shallon 31. Vaccinium ovatum

Stark (16), in 1876, spoke of leaf duration as “‘not a new subject,”’
yet at the same time declared his inability to find anything bearing on
the subject in botanical literature. He made extensive observations
on the native and introduced conifers in his large private grounds in
the British Isles, and distinguished between true leaf fall, as shown in
Taxus and Abies, and the shedding bodily of twigs (cladoptosis) as
shown in Thuja, Pinus, and Sequoia sempervirens. He also remarked
that old trees of Picea and Abies held their leaves for a shorter time
than saplings.

Legget (10), in 1876, recognized the influence of climate on leaf
duration especially in transitional forms.

DURATION OF LEAVES IN EVERGREENS 147

Hoffman (7), for a series of years prior to 1878, carried on inves-
tigations with angiosperm evergreens in the Botanical Gardens at
Giessen. He tied tinfoil tags to the petioles of six or eight leaves on a
given plant and observed these individual leaves at stated intervals,
reporting for several species the leaf duration in months. The method
was too cumbersome to be applied on a very large scale, therefore
his general conclusions seem hardly justified.

Kraus (8), in 1880, published on the duration of evergreen leaves.
Unfortunately his work was not accessible to the writer.

Other writers, as Copeland (1) and Groom (4), also speak of leaf
duration. Galloway (3), in 1896, enumerates various factors which
may affect leaf duration in Pinus virginiana; but these references are
all incidental, and mentioned in connection with other problems, or in
general descriptions.

Sargent (13) (14) and Sudworth (17) in their descriptions of North
American and Pacific Coast trees mention the leaf duration of many
species, but their figures do not hold in some cases for the regions
under discussion, and they give no estimates for other species which
are quite common in this region.

The method of determining the age of a given leaf was simple. In
those species having covered buds, the scars marking the boundaries
of annual growth made it easy to count the years. In those species
with naked buds, as Thuja plicata and Juniperus scopulorum, free-
hand sections were made through the twig at the base of the given leaf,
and the annual growth-rings of the twig counted under the hand lens
or low power of the compound microscope. This method was also
used as a check in other doubtful cases.

When counting by means of terminal bud scars, the endeavor was
to make counts of 100 twigs, but that was not always possible. In
no case, however, did the count fall below 65. When counting by
means of sections the attempt was made to obtain counts of 50 twigs.
This was done in a majority of cases, and in no case did the count fall
below 24. These counts were made in the field whenever possible;
or the material was collected and carried to the laboratory, where the
counts were made immediately, before handling dislodged leaves, or
the unaccustomed dryness of the atmosphere caused them to fall.
The counts were afterwards tabulated, and the tabulations placed on a
percentage basis, the percent being calculated to the nearest whole
number. Finally, curves were plotted from these data (figs. 1-13),

148 VINNIE A. PEASE

the vertical axis representing the percent of twigs or branches ex-
amined which bore leaves persisting for the time in years denoted on
the horizontal axis. :

In making observations on gymnosperms, three chief points were
considered on each twig or branch: (a) the year in which leaf fall com-
menced; (b) the year of maximum fall, that is, the time when the
twigs were fully half bare; (c) the extreme duration of the last scattered
leaves. In angiosperm species it was considered sufficient to make but
one count for each twig or branch, and that to determine the age of the
oldest persisting leaf.

The factors considered as having an influence on leaf duration
were age of the tree, light, climate, and exposure to constant winds.
When studying gymnosperms, observations were made on mature
trees growing in the open and in close stands, as well as on saplings
growing in the open and under the forest cover. In angiosperm
species, observations were made from specimens growing in the open
and under the forest cover. The observations included natural gym-
nosperm forest, partially cleared land, and second growth stands.
These observations were made in the vicinity of Seattle, where the
winds are not strong and the annual rainfall is about 36 inches.

_ In order to get contrasting climatic conditions, the writer spent the

summer of 1915 at the Puget Sound Marine Station at Friday Harbor
on San Juan Island, Washington. This island is sheltered by the
Olympic Mountains, leaving the island an annual rainfall of less than
25 inches. The south slopes of the island are wind-swept, the trees
having the characteristic one-sided form common to such regions. On
this island the Seattle observations were repeated. Also observations
were made to see if leaf duration varied in the same species on the
leeward and windward slopes.

Several peat bogs in the vicinity of Seattle gave opportunity also
to observe the effect of bog habitat on leaf duration. The observa-
tions were made partly at the bog one-half mile east of Ronald, Wash-
ington; partly at the Mud Lake bog, near the west shore of Lake
Washington at 65th St., Seattle.

Since leaf duration varies with the conditions under which the
plant is growing, and since these conditions are matters of general
observation rather than of accurate measurements, it follows that the
results are general. The longest durations are for the poorest com-
bination of conditions; the shortest duration for the best combination

DURATION OF LEAVES IN EVERGREENS T49

of conditions; and the average duration merely the average of these
conditions as nearly as could be ascertained from all the observations
made. Mere general observation of the external and internal con-
ditions of tree are not sufficiently accurate to enable one to predict
with certainty just what one will find in a given tree.

1. Abies grandis Lindl. Shortest leaf duration observed, 2 years;
average, 4-10 years; extreme, 14 years. All observations were made
in the San Juan Islands, since the species is rare in the vicinity of
Seattle. Old trees have a longer leaf duration than do saplings;
shade tends to increase leaf duration; the leaves of wind-swept trees
have a shorter duration than those of protected trees (figs. 9, 10).

2. Juniperus scopulorum Sarg. Shortest duration of green color
observed, I year; average, 2-3 years; extreme, 4 years. The leaves,
however, persist after turning brown. This results in the following:
shortest leaf duration, 3 years; average, 4-6 years; extreme, I4 years.
West of the Cascades this species occurs at low altitudes only in arid
regions. It is quite common in the San Juan Islands. Two distinct
types of leaves are found. The juvenile type, which are long, awl-
shaped, and spreading, have a shorter duration than the adult, over-
lapping scale-like type. In all cases observed, the leaves lost their
green color from I—4 years before they fell, and were then gradually
sloughed off.

3. Picea sitchensis Traut. & May. Shortest leaf duration observed,
2 years; average, 9Q-II years; extreme, 18 years. In the vicinity of
Seattle this species was observed only in peat bogs. In the San Juan
Islands the trees observed stand at the head of a salt marsh which
extends up a creek bed from False Bay. Mature trees in ordinary
soil were not available and no saplings were observed, so that the
results given are by no means complete.

4. Pinus contorta Dougl. Shortest leaf duration observed, 2 years;
average, 4-6 years; extreme, 9 years. Leaf duration reported by
Sargent (13), (14), 7-8 years; by Sudworth (17), 6-8 years. In the
San Juan Islands, saplings in the open, and mature windswept trees,
showed the shortest leaf duration; mature trees, protected from the
wind, the longest duration. Trees introduced on the University
campus showed the shortest duration observed. Sudworth states
that “‘long persistence appears to belong more to young trees,”’ but
the writer found the opposite to be true.

5. Pinus monticola Dougl. Shortest leaf duration observed, I year;

150 "VINNIE A. PEASE

average, 3-4 years; extreme, 6 years. Leaf duration reported by
Sargent (13), 3-4 years. This species, found commonly in the Puget
Sound region in peat bogs, showed the shortest duration of any of the
gymnosperms studied. Mature trees show a tendency to hold their
leaves longer than do saplings.

6. Pseudotsuga taxtfolia Britton. Shortest leaf duration observed,
I year; average, 3-9 years; extreme, 16 years. Leaf duration reported
by Groom (4), Sargent (14) and Sudworth (17), about 8 years; by
Ward (19), 6-7 years. Observations showed that saplings have a
much shorter leaf duration than do mature trees; trees in the open have
a much shorter leaf duration than those in the shade; wind-swept
trees have a short leaf duration; a dry climate increases leaf duration;
a peat bog habitat increases the duration of leaves in saplings to a
greater degree than does a dry climate. No observations were made
on mature trees in peat bogs. A winter season of unusual ‘severity,
such as that experienced by the Pacific northwest in January and
February, 1916, when snow lay on the branches for several weeks,
seriously affects the duration of the leaves. Thirty-eight percent of
the branches examined showed partial loss of the leaves of the pre-
ceding season’s growth, whereas no such loss was observed on the same
trees during the same period of the preceding year. It was noted also
that, in specimens of Pseudotsuga taxifola growing in dense shade,
the annual thickening of the trunks was very slight, the leafy twigs
were very slender, and the needles small and comparatively few on a
year’s growth (figs. I-5).

7. Taxus brevifolia Nutt. Shortest leaf duration observed, 2
years; average, 5-I2 years; extreme, 23 years. Leaf duration re-
ported by Sargent (14), 4-5 years; by Sudworth (17), 6-9 years.

A summary of the effects of varying external conditions cannot be
given since not enough data could be secured. However, in ordinary
conditions of moisture for the Puget Sound region, and in densely
shaded locations in the drier climate of the San Juan Islands, the
duration of leaves has been found to be much greater than previously
supposed.

8. Thuja plicata Donn. Shortest duration of green color, I year;
average, 2-5 years; extreme, 7 years. Since the leaves persist after
losing their color the duration is longer than given above. Observa-
tions resulted in the following: shortest leaf duration observed, 3 years;
average, 4-7 years; extreme, 12 years. Leaf duration reported by

DURATION OF LEAVES IN EVERGREENS Pal

Sargent (14) and Sudworth (17), about 3 years. Observations seem
to indicate that the leaves of mature trees have a greater duration
than those of saplings; that leaves in the shade have a greater duration
than those in the open; that a dry climate seems to prolong the duration
of the leaves; that a bog habitat has the same effect as a dry climate.
Leaves remain on the tree for at least two or three years after losing
their green color, and then are gradually sloughed off by the increase
in size of the twig. Sudworth and Sargent also agree in saying that
the lateral branchlets, which are shed entire, fall in their second year.
The writer found that the duration of lateral branchlets also varies
with habitat. Full data were not taken, but observations showed
that under typical moisture conditions their duration was 2 to 3
years, while in bog specimens they persisted 4, 5 or 6 years (figs. I1, 12).

9. Tsuga heterophylla Sarg. Shortest leaf duration observed, 2
years; average, 4-7 years; extreme, I2 years. In general, mature
trees show a greater leaf duration than do saplings under the same
conditions of light and moisture. However, the shaded saplings
observed in the vicinity of Seattle showed a greater leaf duration than
that of mature trees growing under the same conditions. The saplings
observed grew on fallen logs in dense shade under the parent trees,
and had grown very slowly. Specimens which showed 20 annual
growth-rings were less than a meter high, and no thicker than an
ordinary lead pencil. The linear growth per year in many of the twigs
examined was less than a centimeter, the needles on each year’s
growth were few in number, and the individual leaves were very small.
There may be some correlation between this extreme slowness of
growth and the increased duration of the leaves. Saplings in a moist
climate show a longer leaf duration than saplings in a dry climate,
while the converse is true for mature trees.

Bog saplings, observed in the peat bog at Ronald, Washington,
were dwarfed and stunted in their growth to a much greater extent
than the shaded saplings previously described. As determined by
counting the annual rings under the low power of the compound micro-
scope, these saplings ranged in age from 5 to 32 years. They were
from 17 to 60 cm. high, but the height was not proportional to the
age. Both lateral and terminal shoots averaged less than a centimeter
per year in linear growth; and a year’s growth in many cases comprised
from 6 to 10 needles, which were much below normal in size. The
leaf durations in these bog saplings show a remarkable feature, which

152 VINNIE A. PEASE

was not observed in the case of any other species examined, under any
condition. All three curves, that is, for beginning of leaf fall, for
maximum leaf fall, and for extreme duration, show two maxima,
the first occurring in the fourth year in all three cases, and the second
in the sixth and seventh. This is probably due to variations in the
toxicity of the bog water in different parts of the bog (13). Mature
specimens from the bog also showed slow growth and small needles,
but the duration curves were normal and the maxima lay between
the two sets of maxima in the curves of the saplings. Peat bog
specimens, both saplings and mature, show an increased leaf duration
over specimens growing’ in the open in ordinary soil, the duration
more nearly approximating that of specimens from a drier climate
(figs. 6-8).

10. Rhamnus purshiana DC. Sudworth says that “‘in its northern
habitat the thin large leaves are shed regularly in the autumn, while
in the drier southern distribution to and through central California,
the leaves, which are smaller, thicker, and somewhat leathery, often
persist more or less during late autumn and winter.”’ Frye and Rigg
(2) state that the leaves are ‘‘deciduous except occasionally on very
young plants.’’ Sargent (14) says that “‘in Washington and Oregon
the leaves fall late in November, while farther south and near the Cali-
fornia coast they remain on the branches almost all winter, or until
the following spring.’”’ The writer has found that not only do seedlings
retain their leaves in the Puget Sound region, but that trees in moist
rich humus under the forest cover, up to ten years old, may retain at
least a part of their leaves well on into May, when the new season’s
leaves are fully expanded; and these persistent leaves seem not to
differ in size or texture from those shed in. the fall.

Il. Vaccinium parvifolium J. E. Smith. The small plants which
have germinated on fallen logs under the forest cover are almost in-
variably evergreen. The slender branches which arise from the root
crowns of older shrubs also bear leaves which persist from one to
several seasons. It was thought at first that evergreenness was con-
fined to branches near the ground, but later several specimens were
found which bore evergreen leaves from I to 2 meters above the ground,
on the upper branches. Two distinguishing characteristics present
themselves in regard to these evergreen leaves:

(a) The leaves are usually much smaller than the ordinary de-
ciduous leaves, and are borne on very slender, slow-growing branches.

DURATION OF LEAVES IN EVERGREENS E53

These branches attain a growth usually of less than 5 centimeters in a
season, and may bear no more than 3 leaves on a season’s growth.
However, leaves of 3 or 4 years’ growth have been found which were
from 20 to 30 mm. in length, while the species description, Frye and
Rigg (2), gives leaves 6 to 17 mm. long.

(b) While evergreen leaves are quite common, they are not usual
on mature shrubs, and there seem to be no definite external factors
which will explain their appearance or non-appearance. At best,
only a few branches bear evergreen leaves. Also, of two shrubs of
approximately the same age, growing under apparently the same con-
ditions, and standing only 3 or 4 meters apart, one may be entirely
devoid of leaves and the other have several branches bearing leaves of
3, 4, 5 or even 6 years’ duration. The extreme duration observed was
6 years.

12. Arbutus menziesu Pursh. Observations were made on the
campus of the University of Washington. The leaves begin to fall
early in June of their second year. Many of the trees put on a second
growth late in the summer, whose leaves are somewhat smaller and
lighter in color than the normal spring leaves, and this gives the
appearance of two seasons’ growth. During the extreme and unusual
cold weather of the past winter, many of the spring leaves were killed
by frost while the late summer leaves seemed to be scarcely affected.
This enhances still more the appearance of two seasons’ growth.

13. Ceanothus velutinus Dougl. Like Arbutus menztesi1, this nor-
mally holds the leaves of one season only until those of the succeeding
season are fairly matured; that is, for a period of about 15 months.

14. Linnaea americana, Forbes. This trailing vine, as a rule, does
not drop its leaves, but the leaves simply decay while attached, as
they lie against the damp moss or already decaying leaves of the sub-
stratum. They persist throughout the winter, and in many cases
until after the flowering season in the spring.

15. Micromeria douglasu, Benth. The same condition is found in
this as in Linnaea americana.

16. Rubus laciniatus Willd. This plant has escaped from cultiva-
tion, and is commonly known as the “evergreen blackberry.’’ Some
leaves persist at least until after the flowering season.

17. Rubus ursinus Schlecht. & Cham. This is common on logged-
off lands; according to Frye and Rigg (2) it is evergreen only in western
Washington.

yA: VINNIE A. PEASE

18. Rubus pedatus J. E. Smith. The writer found a single speci-
men, and that bore leaves of two seasons’ growth.

19. Arctostaphylos tomentosa Dougl. ‘The writer had access to only
one specimen, a shrub which has stood for several years in the north-
west angle of a 3-story building on the University campus. ‘This
showed a leaf duration of 4, 5 and 6 years on various branches.

20. Arctostaphylos uva-ursa Spreng. Shortest leaf duration ob-
served, 2 years; average, 3 years; extreme, 5 years.

21. Berberts aquifolium Pursh. Shortest leaf duration abscroet
I year; average, 2-4 years; extreme, 6 years. Not found usually in
shaded situations. A dry climate shortens its leaf duration.

22. Berberis nervosa Pursh. Shortest leaf duration observed, 2
years; average, 3-4 years; extreme, 8 years. Plants growing in the
shade show a longer leaf duration than those growing in the open. A
dry climate accents the difference in duration between leaves in the
open and those in the shade.

23. Chimaphila menziesii Spreng. Shortest leaf duration observed,
2 years; average, 4—5 years; extreme, 8 years. This species was found
only in a limited area on San Juan Island.

24. Chimaphila umbellata Nutt. Shortest leaf duration observed,
I year; average, 2-4 years; extreme, 7 years. A dry climate tends to
increase its leaf duration.

25. Gaultheria shallon Pursh. Shortest leaf duration observed,
I year; average, 2-4 years; extreme, 6 years. Shade plants under
typical moisture conditions have a shorter leaf duration than plants
in the open, while under dry conditions plants in the open have the
shorter leaf duration. Plants growing in sphagnum about the margins
of peat bogs resemble in growth-habit plants growing in the open in
ordinary soil, but have a decided tendency toward shorter leaf duration.

26. Kalmia poltfolia Wang. Shortest leaf duration observed, I
year; average, 2 years; extreme, 3 years. In contrast to Ledum,
shaded plants showed a tendency to shorter leaf duration, and plants
which had been growing for several years in the experimental gardens
were entirely bare of leaves when observed in December.

27. Ledum groenlandicum Oeder. Shortest leaf duration observed,
I year; average, 2-4 years; extreme, 5 years. Plants in an open peat
bog showed the shortest leaf duration. Plants which had been trans-
ferred to the experimental gardens of the university campus several
years ago showed a marked tendency to increased leaf duration.

DURATION OF LEAVES IN EVERGREENS 155

Plants growing in the shade about the borders of the bog were much
modified, being much taller; and with leaves larger, thinner, less
revolute, and less densely clothed with hairs on the under surface.
These leaves were of much longer duration.

28. Oxycoccus oxycoccus intermedius Piper. Shortest leaf duration
observed, I year; average, 2—3 years; extreme, 4 years. Plants parti-
ally shaded by the taller growth of Ledum about the hemlock hillocks
showed increased leaf duration.

29. Pachistima myrsinites Raf. Shortest leaf duration observed,
2 years; average, 3-4 years; extreme, 8 years. This species was ob-
served only in the San Juan Islands. Plants growing in exposed
locations on the windward side of the islands had a shorter leaf duration
than those on the leeward side.

30. Rhododendron californicum Hook. Shortest leaf duration
observed, I year; average, 2 years; extreme, 3 years. This plant was
observed only on the university campus, where it is used extensively
as an ornamental shrub.

31. Vaccinium ovatum Pursh. Shortest leaf duration observed,
2 years; average, 2-4 years; extreme, 7 years. Plants in the shade
show a decided increase in leaf duration (figs. 13).

‘It has already been noted that in many of the gymnosperms growing
under adverse conditions, that is, in dense shade or in peat bogs, leaves
are smaller and fewer in number on a year’s growth than on specimens
of the same species growing under more favorable conditions. While
the tendency is not so marked in all cases the same difference in size
was noted between the leaves of mature trees and those of saplings,
mature trees ordinarily having smaller leaves than those of saplings.

Kraus (9) observed that the length and vigor not only of the grow-
ing shoots but also of the needles vary in different seasons; and Reinke
(12) demonstrated that in transplanted evergreens the needles formed
during the growing season immediately following the transplanting
are conspicuously shorter than those formed during either the pre-
ceding or the following season. This was afterward confirmed by
Copeland (1), who measured the needles on transplanted evergreens
on the campus of Indiana University. Former observations are thus
extended to include the variation in size of leaves on trees of the same
species of different age, or growing in different habitats.

Groom (5) observed that though the individual leaf is small, the
aggregate leaf surface of the conifer often greatly exceeds that of the

156 VINNIE A. PEASE

dicotyledonous tree; and Copeland (1) in his study on the size of ever-
green needles found that in abnormal years, when the leaves are small,
“the number of needles compensates the plant for their lack of size,
sometimes furnishing an even greater surface of leaf than is borne on
the normal year’s growth of stem.’’ Following the same line of
thought, it may be that the longer duration of leaves on mature trees,
or on trees growing under adverse conditions, which is correlated with
a decrease in size, tends to keep up the total leaf area. With longer
duration and smaller leaves in dense shade as compared with open
situations, increased duration may be correlated with two factors.
Reduced size of the individual leaf, and reduced photosynthetic ac-
tivity, due to diminished light intensity, are both compensated by an
increased number of leaves; and increased leaf duration would furnish
this increase in the number of leaves.

In all angiosperm forms which were examined, both in the open and
in the shade, the leaves on shaded plants were much larger than those
on plants exposed to direct sunlight; and with the exception of Gaul-
theria shallon growing in the typical climatic conditions of the vicinity
of Seattle, plants in the shade held their leaves longer than those in
the open. Hasselbring (6), in commenting on his experiments with
Cuban tobacco grown under a cheese-cloth shade, states that “the
reduction in photosynthesis in the shade leaves was compensated by
an increase in leaf area, so that the production was not diminished.”
In various species under discussion, it is quite possible that the increase
in photosynthetic area, which compensates the decrease in light in-
tensity, is due not only to the increased size of the leaves but also to
their increased duration.

CONCLUSIONS

1. Leaf duration varies widely among the different evergreen
species, ranging from Rhamnus purshiana, which in young plants
sometimes holds part of the leaves of one season until those of the next
season are mature, to Taxus brevifolia, which has an extreme leaf
duration of 23 years.

2. Leaf duration varies widely in individuals of the same species of
different age or growing in different habitats: (@) Saplings have a
shorter leaf duration than mature trees in the same habitat. (0) Trees
or shrubs growing in the open have a shorter leaf duration than those
of similar age in the shade. (c) Trees or shrubs on a windward coast
have a shorter leaf duration than those on a leeward coast. (d) Gym-

DURATION OF LEAVES IN EVERGREENS 157

nosperms in a moist climate have a shorter leaf duration than those in
a drier climate. (e) A peat bog habitat has an effect similar to a dry
climate.

3. Those factors which cause slowness of growth, and thus only a
slight increase in diameter of the axis, are accompanied by an increased
duration of the leaves.

4. Under the same climatic conditions, those factors which cause
an increase in transpiration are accompanied by a decrease in leaf
duration, and thus by a decrease in the transpiring surface.

5. Those factors which cause a decrease in photosynthetic activity
are accompanied by an increase in leaf duration, and thus by an
increase in the photosynthetic area.

6. It is quite possible that the variations in leaf duration in a given
species may be due to differences in transpiration or photosynthetic
activity, caused by difference in age or habitat.

UNIVERSITY OF WASHINGTON, SEATTLE

BIBLIOGRAPHY

1. Copeland, E.B. The Size of Evergreen Needles. Bot. Gaz. 25: 427-436. 18098.
2. Frye, T. C., and Rigg, G. B. Northwest Flora. Seattle. Ig1o.
3. Galloway, B. T. A Rust and Leaf Casting of Pine Needles. Bot. Gaz. 22:
433-453. 1896.
4. Groom, Percy. ‘Trees and their Life Histories. London. 1909.
Remarks on the Oecology of Conifers. Annals of Botany 24: 241-269.
1910.
6. Hasselbring, Heinrich. Effect of Shading on the Transpiration and Assimilation
of the Tobacco Plant in Cuba. Bot. Gaz. 57: 257-286. 1914.
7. Hoffman, H. Ueber Blattdauer. Bot. Zeit. 34: 705-708. 1878.
8. Kraus, Gregor. Die Lebensdauer der immergriinen Blatter. Naturf. Ges.
Halle. Sitzber. 1880.
Abhandl. Naturf. Ges. Halle 16: 363. 1886.
10. Legget, W. H. Bull. Torrey Club 6: 125. 1876.
11. Piper, C. V. Flora of the State of Washington. Contr. U. S. Nat. Herb. 11.
1906. :
12. Reinke, J. Ber. Deutsch. Bot. Ges. 2: 376. 1884.
13. Rigg, G. B. Decay and Soil Toxins. Bot. Gaz. 61: 295-310. 1916.
14. Sargent, C.S. The Silva of North America. New York. 1894-1902.
Manual of the Trees of North America. New York. 1905.
16. ——. Report on the Forests of North America. Govt. Ptg.
Office, Wash. D. C. 1884.
17, Stark, James. On the Shedding of Branches and Leaves in the Coniferae.
Trans. Roy. Soc. Edinburgh 27: 651-666. 1876.
18. Sudworth, G. B. Forest Trees of the Pacific Slope. U.S. Dept. Agr. Forest
Service. 1908.
19. Ward, H. M. Trees, 2. Cambridge. 1904.

158 VINNIE A. PEASE

EXPLANATION OF FIGURES 1-13

Horizontal figures indicate years; vertical figures indicate number of specimens.
Unless otherwise stated, ...... is curve showing beginning of leaf fall; is curve
showing greatest leaf fall; - - - - is curve showing extreme duration of leaves.

Fic. 1. Pseudotsuga taxtfolia, on San Juan Island; mature trees, in the open, on
leeward slope.

Fic. 2. Pseudotsuga taxifolia, at Seattle; sapling, in the open.

Fic. 3. Pseudotsuga taxtfolia, at Seattle; mature tree, in the open, after unus-
ually cold weather.

Fic. 4. Pseudotsuga taxifolia, on San Juan Island; mature tree, in the open, on
windward slope. ,

Fic. 5. Pseudotsuga taxifolia, at Seattle; mature tree, in the open.
Fic. 6. Tsuga heterophylla, at Seattle; mature tree, in peat bog.
Fic. 7. Tsuga heterophylla, on San Juan Island; mature tree, in the open.
Fic. 8. Tsuga heterophylla, at Seattle; mature tree, in the open.
Fic. 9. Abtes grandis, on San Juan Island; sapling, in the shade.
I

Fic. 10. Abies grandis, on San Juan Island; sapling, in the open.

Fics. II AND 12. Thuja plicata, on San Juan Island; mature trees; Fig. I1 in
the shade, Fig. 12 in the open. ...... is curve of loss of green color; is curve of
beginning of leaf fall; - - - - is curve of extreme leaf duration.

Fic. 13. Vaccinium ovatum, at Seattle. is curve of extreme leaf duration
in shade; - - - - is curve of extreme leaf duration in open.

DURATION OF LEAVES

50
40
30
20

/0

:
Ps Gay Ss Bir 7 Oe EG 10 oe

70
Z) 30
70 20

60 10

1

IN EVERGREENS

So
70
60
$0

Yo

20

/0 ?

Q

159

PEASE

VINNIE A.

160

THE RELATION BETWEEN EVAPORATION AND PLANT
SUCCESSION IN A GIVEN AREA

FRANK C. GATES?

As a result of an investigation into the relative amounts of evapora-
tion from the chamaephytic or ground layer of certain genetically
connected, adjoining plant associations at Havana, Illinois, during
the summer of 1910, Gleason and Gates (1) concluded: ‘‘ that succes-
sions between associations are not caused by any conditions of evapor-
ation.’” In conclusion to a much more extensive series of investiga-
tions, bearing on the same subject, Fuller (2) concludes: “‘ the decreased
rate of evaporation . . . is the direct cause of successions between
different associations.’’ Weaver (3) concludes: “A study of the dif-
ferences of the rates of evaporation in the various plant formations
and associations shows that these differences are sufficient to be im-
portant factors in causing succession, at least through the earlier stages,
where light values are usually high.”

Each investigation dealt with neighboring associations in a limited
area, thereby accentuating the action of local factors and minimizing
the obscuring interference of climatic factors. An inspection of the
pertinent data obtained in each of these investigations shows that they
are similar; yet diametrically opposite conclusions are drawn.

To obtain new data on the relationship between evaporation and
plant succession, three series of experiments were carried on during the
summers of 1915 and 1916, at the University of Michigan Biological
Station at Douglas Lake, Michigan. During 1915, twenty-six stan-
dardized Livingston atmometers were employed for a period of 40
‘days, inclusive of the time of maximum evaporation during the year.
In 1916, sixteen newly standardized instruments were employed during
varying periods inclusive of the severest summer evaporation in years.
Each instrument was set up in close proximity to certain plants. The

1 Contribution from the University of Michigan Biological Station at Douglas
Lake, Michigan, No. 41.

2 Owing to the press of duties attendant upon the establishment of the University

of Michigan Botanical Garden, Dr. H. A. Gleason was unable to collaborate, as
planned.

161

162 FRANK C. GATES

experimentation and the calculation of the results to a standard basis
followed the normal methods used for such work.

The object of this experimentation was the determination of the
relationship between evaporation and plant succession in a local area.
Douglas Lake region presents an admirable opportunity for such ex-
perimentation. A detailed discussion of the vegetation of the area
will be found in Gates (4). A brief resumé of the pertinent facts is as
follows: Aside from a few small associations, local along streams and
around lakes, the vegetation of the region falls readily into three di-
visions, each characterizing a soil type. Bog associations, particularly
the Chamaedaphne, Larix, Picea and Thuja associations, occupy the
low wet soil. The sandy uplands were dominated by the pine asso-
ciation—now, following lumbering and fire, largely replaced by the
aspen association. Clayey soil on the uplands is occupied by the
hardwood or beech-maple association, except where it has been de-
stroyed by lumbering or fire.

Experiments were carried on separately with the vegetation of
each soil type. The Thuja association, chosen for the bog experi-
mentation, is typically composed of a large number of trees of Thuja
occidentalis, growing close together. The ground vegetation in a dense
patch of Thuja is virtually nil. In open places, as along roads and
trails, ericads and ericad-like plants are conspicuous. A few of the
most abundant species are Ledum groenlandicum, Streptopus amplext-
folius, Moneses uniflora, Pirola asartfolia incarnata, Muitella nuda,
Rubus triflorus, Cornus canadensis, Carex spp., Habenaria obtusata,
Chamaedaphne calyculata and Vaccinium oxycoccus.

The pine type—once represented by Pinus strobus and Pinus
resinosa, now by scattering seedlings, small trees, and a few old trees
of the same species mixed in with the aspen association—was investi-
gated during 1915. At least 96 percent of the trees in the aspen
association belong to the following four species: Populus tremuloides,
Populus grandidentata, Betula alba papyrifera, and Prunus pennsyl-
vanica. Among the higher shrubs are Salix rostrata, Rhus glabra,
and Viburnum acerifolium; among the lower shrubs, Dvervilla lonicera
(which is frequently exceedingly abundant), Vaccinium pennsylvan-
acum, Gaultheria procumbens, Rubus idaeus aculeatissimus, and Rubus
allegheniensis are quite common. The fern, Pteris aqutlina, is fre-
quently more abundant than any of the shrubs. With the shrubs are
seedlings and small trees of Quercus rubra, Acer rubrum, Acer sac-

RELATION BETWEEN EVAPORATION AND PLANT SUCCESSION 163

charum, Fagus grandifolia, Tilia americana, Pinus resinosa, and Pinus
strobus. Among the herbaceous species are several grasses (Panicum
xanthophysum, Danthonia spicata, Poa pratensis, Agrostis hiemalis,

Cc. in

No. Location. 40- Day
Hardwood Series, Period.

x. Cutover in 1914-15. 590
2. Cutover in 1914-15. 561
3. Open place in old cut. 453
4. Cutover in 1913-14. 425
5. Cutover in 1913-14. 416
6. New margin of woods, 378
7. Within margin go meters. 240
8. Pine Point hardwoods. 210
g. Opening in dense woods. 187
to, Within dense Aspen thicket. 179
uz, Dense hardwoods, 175
12. Very dense hardwoods. 147

13. Burntover hardwoods near Bryant’s. 473
14. Bare ground near laboratory. 524 §

15. Burntover pine land near Bryant’s. 513

Pine in Aspen Series.

16, Exposed place in upper flat. 347
17. Exposed place on hill. 321
18, Exposed place in lower flat. 315
1g. Exposed place on hill. 310
20. Aspens in lower flat. 288
21. Crest of slope in Aspens. 238
22. Aspens in upper flat. 198
23. Aspens in upper flat. 197
24, Aspens in middle.-flat. 196
25. Aspens in middle flat, 187
26. Foot of slope in Aspens. 178

Fic. 1. Diagram showing the total evaporation and the rate per day for 40
days (July 10 to August 19, 1915) from different stations in hardwood and pine land
in the vicinity of Douglas Lake, Michigan.

Agrostis alba, and Oryzopsis aspertfolia), a very few sedges, and other
plants, such as, Convolvulus spithamaeus, Aster laevis, Hieracium sca-
brum, Hieractum venosum, Solidago canadensis, Melampyrum lineare,
Fragaria virginiana, Smilacina stellata, besides such common weeds as,
Erigeron canadensis, Rumex acetosella, Lepidium virginicum, Epilobium

164 FRANK C. GATES

angustifolium, and Erigeron ramosus. Overtopping all other vegeta-
tion are a few scattered giant trees of Pinus strobus and Pinus resinosa.

HARDWOOD SERIES: 1915..,
1,Cut over in 1914-15.
3,Open place in old cut,
5,Cut over in 1913-14.
6,Margin of woods,
7. Nipty meters within margin.
9,Cpening in dense woods,
11.Dense woods,
12.Very dense woods,
1{3.Burnt over near Bryant's,

Fic. 2. Diagram showing the daily rate of evaporation in cc. and the pre-
cipitation in cm. for the intervals between readings from certain stations in the
hardwood series, I9I5.

The principal trees in the hardwood or beech-maple association
are Fagus grandifolia, Acer saccharum, Tsuga canadensis, Betula lutea,
and Tilia americana. Shrubs occur largely in openings, Acer penn-
syluanicum being most abundant. A large number of herbaceous
species grow near the ground. Among the more frequent of these
are Araha nudicaulis, Maianthemum canadense, Trillium grandiflorum,
Trientahis americana, Aster macrophyllus, Sireptopus longipes, Strep-
topus roseus, Medeola virginiana, Clintonia borealis, and Actaea alba.

Clearings made in different years, now covered with mixtures of
vegetation, furnish series from bare ground up to the hardwood as-
sociation. Another series leads from bare ground, through aspens, to

RELATION BETWEEN EVAPORATION AND PLANT SUCCESSION 165

the pine association; while a third series leads from open water to the
Thuja association. In wet soil, seedlings of Thuja are present under
many conditions. On sandy land, healthy seedlings of Pinus strobus
and Pinus resinosa occur under a large number of conditions.
Similarly on the better soil are seedlings of Acer saccharum and
Fagus grandifolia: On other portions of the lumbered land such
seedlings have become small trees, with every prospect of reproducing
the original forest. For the purposes of the present investigation,
young seedlings, 15 to 25 cm. in height were chosen, as these are in a
most critical stage.

Cc. in
No. Location. 47-Day
East Point Series. Period.
27. Edge of dune. 618
28, Dense Thuja Bog, 330
29. Very dense part of bog. 225
30. Densest part, no green plants. 155
Reese’s Bog Series.
31. Clearing. 587
32. Layered Thuja in opening. 460
33. Dense, no green plants. 435
34. Marginal foss. 391
35. Thuja seedling in dense part. 345
36. Bog in Aspens, 335
Series East of Douglas Lake.

37. Thuja seedling in slashed bog. 344

38. Thuja seedling in slashed bog. 307

Bryant’s Bog Series.
39. Edge of Chamaedaphne Assoc, 590

40. Picea, Pinus and Larix seedlings in 555
Chamaedaphne Assoc,

Fic. 3. Diagram showing the total evaporation and the rate per day for 47
days (July 8 to August 24, 1916) from different bog stations.

In 1915 three atmometers were run at the level of Acer saccharum
seedlings in the dense hardwood forest under three conditions: at-
mometer No. 11 under ordinary shade (Fig. 6), No. 12 under very
heavy shade, and No. 9 in an opening caused by the removal of a
large tree. Atmometers No. 6 and No. 7 were run in a large patch of
vigorous I to 3 year old seedlings near the edge of the forest. The
edge was the result of the preceding winter’s clean-cut lumbering—
therefore exposed to light and wind (Fig. 5). Atmometer No. 6
was placed at the very edge, while No. 7 was run about 90 meters in
the forest. Atmometers No. 1 and No. 2 were run alongside of Acer

166 FRANK: C.sGATES

seedlings, growing unprotected in the open sun in an area cleared
during the preceding winter (1914-15). In this same area three at-
mometers were started in 1916. The second year had allowed the
brambles to encroach upon the fireweeds—clothing the ground with a
dense covering of vegetation. The maple seedlings were likewise one
year older and their vigor was positive proof that they were amply
and easily meeting conditions. ‘The introduction of cattle into the
area in the middle of the summer necessitated the withdrawal of the
atmometers. The healthy condition of the seedlings in the fall,
however, was evidence that these seedlings could withstand even such
an extremely dry summer as that of I916.

BOG SERIES: 1916,
East Point Series, Reese's Bog Series.
27,Edge of -dune,. 31,.Clearing.
29.Very dense part of bog. 33.Dense, no green plants,
30.Densest part, no green 34, Marginal foss,
plants. 35.Thuja seedling in dense part,
41.Beach (One week),
37.Slasht bog east of Douglas Lake,

ae Ce Je

29 a en

a ~

Ete oe === ye 33
Siew ae Sc ere yy pews
7 eS ee ee eee
iapigte aye, =
“

Fic. 4. Diagram showing the daily rate of evaporation in cc. for the intervals
between readings from certain bog stations, 1916. The daily precipitation is shown
on the same scale in cm.

In 1915, atmometers No. 4 and No. 5 (Fig. 7) were run near maple
seedlings in an area cleared in the winter of 1913-14. Weeds and
brambles were also present. Atmometer No. 3 was run by maple
seedlings in an open place in a thicket-tree growth—long since cut and
lightly burnt—into which brambles have entered thickly. Atmometer
No. 8, the last of this series, was run on Pine Point in a mixture of hard-
wood and cedar in which all the large Thujas had been cut out.

RELATION BETWEEN EVAPORATION AND PLANT SUCCESSION 167

A similar series of experiments was run in connection with the
establishment of pine plants in the aspen association. Pine seeds are
furnished by large trees bordering the lake and scattered sparingly in
the main body of the pine land. The ground conditions are various.
Open sandy soil may be quite plantless where fire damage has been
very severe. The ground is sometimes covered with a dense carpet
of moss or sod, which makes seeding ineffective.

Fic. 5. Margin of hardwoods, the result of clean cut lumbering. July 14, 1916.

Eleven atmometers were set out on a line running back from the
lake near the Biological Station under conditions as follows: No. 20
in an open growth of aspen, the ground covered with Pteris; No. 18
near the preceding in a growth of Pteris under the open sky; No. 26
at the foot of a slope in a dense aspen thicket, in which the ground was
entirely obscured by the luxuriant growth of Pteris; No. 21 about 20
meters from the preceding at the crest of a slope where the ground
flora was predominantly formed by Gaultheria procumbens under a
fairly open aspen thicket (Fig. 8). Atmometers No. 24 and No. 25
were run in a dense aspen thicket, where Pteris was also luxuriant.
This thicket was separated from the uplands by a steep partially
cleared slope about 10 meters high. Atmometers No. 17 and No. 19
were run on this slope. On the uplands there were fewer pine seed-
lings, both because of the distance from seed trees and the greater

168 FRANK C. GATES

fire damage. Three atmometers were run in close proximity to small
pine seedlings, two of which, No. 22 and No. 23, were under a fairly
dense aspen stand, while No. 16 was exposed to the sky.

Fic. 6. Floor of a hardwood or beech-maple forest showing atmometer No. 11
in a dense mass of Acer saccharum seedlings. Seedlings of Acer pennsylvanicum are
also present. July 22, 1915.

Until the winter of I91I-12, south of Bryant’s hotel, there was
a patch of hardwood. East of it was pineland, now vegetated with a
very open growth of aspen. A north and south ravine sharply separ-
ated these two areas of different vegetation. As the area to the east
is in line with the prevailing westerly winds, it has had abundant
opportunity to become thoroughly seeded with Acer saccharum and
other hardwood plants. The hardwood was cut in the winter of I911-
12 and fireswept in May, 1915. ‘To determine whether there was any
particular characteristic of evaporation which possibly could have
influenced the fact that Acer seedlings were not present in the pineland,
although present on the hardwood land, two’atmometers were run—
No. 13 in the burnt-over hardwood land and No. 15 about 200 meters
distant in the pine land. |

The evaporation conditions attendant upon the establishment of
Thuja seedlings in boggy soil were investigated with 16 atmometers in
1916. Seed trees of Thuja are smaller, less abundant, and more local-

RELATION BETWEEN EVAPORATION AND PLANT SUCCESSION 169

ized in their distribution than pine or maple, which explains why
Thuja was not found in some of the smaller bogs. Atmometers No. 39
and No. 40 were run in a smal! Chamaedaphne bog in which Larix and
Picea mariana were conspicuous invaders. This bog has been thor-
oughly fireswept and no Thuja is present. Atmometers No. 41 and
No. 42 were started by Thuja seedlings on the beach and at the edge
of the beach thicket respectively, but after the first week had to be

Fic. 7. A view in a hardwood area cut over in 1913-14, showing atmometer
No. 5. The conspicuous weed is Erigeron canadensis. July 22, 1915.

discontinued. In a small slashed bog along a little stream east of
Douglas Lake, atmometers No. 37 and No. 38 were run in moderately
open conditions near healthy Thuja seedlings. At East Point there
are several bogs in different stages of development. Atmometer
No. 27 was run near a Thuja seedling at the edge of the fringing dune,
exposed to winds from the lake, No. 28 near Thuja seedlings at the
inner edge of the bog, No. 29 in the densest part of the bog in which a
Thuja seedling could be found growing, while No. 30 was run, in
August, in the deepest and darkest spot which could be found. Thuja
seeds but no Thuja seedlings were present. Atmometer No. 36 was
run in a small bog in the aspens south of the Biological Station.
Larix, Thuja and Picea were present, but fire had seriously damaged
the vegetation.

170 FRANK C. GATES

Reese’s bog, the largest bog in the vicinity of the Biological Station,
is a well-developed Thuja bog. Atmometer No. 34 was run near a
Thuja seedling in the marginal foss at the foot of a hill, where the
soil was very wet. Although exposed to the sun, the opportunity for
free circulation was poor. Atmometer No. 35 was near a Thuja

Fic. 8. View showing atmometer No. 21 near a pine seedling at the crest of a
hill in an open aspen growth. The ground is carpeted with Gaultheria procumbens.
August 9, I9I5.

seedling in a dense thicket of 10—-20-foot saplings in very wet soil—
likewise hemmed in from the wind. Atmometers No. 32 and No. 33
were in Thuja on slightly higher ground where the soil was dry at the
surface and the circulation good—No. 32 in a slight opening in which
a layered sprout was healthily growing and No. 33 in a very dense
thicket of small trees under which was no green ground vegetation
(Fig. 9). Ungerminated composite and Thuja seeds were found in the
layer of dead Thuja leaves. Atmometer No. 31 was run by a Thuja
seedling in a good-sized clearing where the seedlings were exposed to
full sunlight.

Atmometer No. 14 represents the evaporation conditions of the
bare ground near the lake in the immediate vicinity of the laboratory,
in 1915.

In each experiment, unless otherwise noted, the atmometer was

RELATION BETWEEN EVAPORATION AND PLANT SUCCESSION I7I

run in immediate proximity to a young healthy seedling of maple,
pine, or white cedar and represents the conditions successfully met by
those seedlings. Where virgin hardwood forest is cleared during a
winter, the vegetation in the following spring consists of such forest
species as can withstand the new conditions. This includes the seed-
lings of Acer saccharum. Weeds appear later in the season, but not
in great abundance during the first year. During this time maple
seedlings have little or no protection from the full sun, yet large

Fic. 9. View in a Thuja bog, showing atmometer No. 33 in the center of the
background where the shade is so dense that no green ground vegetation is present.
August 12, 1916.

numbers of them survive. Is a downward change in evaporation a
necessary prerequisite to succession or is the evaporation changed asa
result of succession? If the former is the case, since Acer saccharum
seedlings are normal to the floor of the climax vegetation where the
rate of evaporation is very low, it might be logical to suppose that
maple seedlings will not be found except where the rate of evaporation
is much less than that over bare ground. If the latter is the case,
maple seedlings will be found growing wherever the soil is suitable,
regardless of the rate of evaporation of the habitat and regardless of

any change that their development may subsequently have upon the
habitat.

172 FRANK C. GATES

The results and their interpretation follow: Taking up the hard-
wood series first, the following results were obtained. In the area cut
over during the winter of 1914-15, where sufficient time had not yet
elapsed for weeds to invade and change the evaporating conditions of
the ground layer, the rate of evaporation was 590 and 561 cc. for 40
days in the middle of the summer of 1915. This rate was 3.37 times
as great as that from the floor of the normally dense hardwood forest
in this region. In the area cut during the winter of 1913-14, where
weeds and brambles had entered in quantity, the evaporation rates
were 416 and 425 cc. from two stations. A relative slowing up of the
rate of evaporation even during the season was plainly evident in
atmometer No. 4, as the development of weeds during the course of
the season came to protect the instrument and the Acer seedling to a
greater and greater degree. In fact this protection from weeds was
sufficient to cause a lower rate of evaporation than was obtained from
atmometer No. 3 run in an open weedless spot in an area where hard-
woods had made considerable progress in revegetating a former cut.
There, the rate was 453 cc. during the same length of time. At the
edge of the woods, where atmometer No. 6 was stationed in a luxuriant
growth of Acer saccharum seedlings, a rate of 378 cc. was obtained for
the period of experimentation. Ninety meters in from the margin,
the rate had decreased to 240 cc. Within the woods the rate was 175
cc. in a spot of average density, 147 cc. in a very dense situation, and
187 cc. in a small opening in the dense forest. These results show a
wide range of conditions from bare ground without shade—the severest
conditions maple seedlings could be called upon to withstand—to the
mature forest with its dense shade. Seedlings in the open received
sunlight. Under more advanced conditions in the vicinity it was seen
that such seedlings were developing into trees, while the vast majority
of the multitudes of seedlings in the dense forest did not persist for
more than a year or two, unless they were in openings.

This is a clear case in favor of the contention that the seedlings of
the dominant species of certain associations become established ir-
respective of the evaporation conditions—in fact, with the addi-
tional advantages accruing from an increased amount of sunlight,
seedlings of mesophytic species thrive better under more xerophytic
conditions than that which the mature forest furnishes.

In the presence of sunlight, Thuja seedlings readily develop in either
sandy or boggy soil, having a sufficient supply of water, under the

RELATION BETWEEN EVAPORATION AND PLANT SUCCESSION 173

entire range of evaporation conditions present in the region. Thuja
seedlings commence development on the open beach, but are de-
stroyed by ice action. On the low fringing dune, where the evapora-
tion was 618 cc. from atmometer No. 27, Thuja seedlings were more
frequent.

At Bryant’s bog, where conditions were intermediate between the
sand dune and a normal cedar bog, atmometers No. 39 and No. 40
gave 590 and 555 cc. respectively for the season of 1916. As pre-
viously noted, this bog has been repeatedly devastated by fire and
there are no Thuja seed trees in the immediate vicinity. The absence
of Thuja, therefore, can not be attributed to the conditions of eva-
poration.

In certain of the East Point bogs, conditions pre-eminently suitable
for the development of Thuja prevail. Although Thuja seedlings are
found under a wide range of evaporation conditions, there are places
in the bog where it is too dark for them to grow. Darkness is here
attended by low evaporation. With an increase in light, evaporation
is increased. Since a certain amount of light is necessary for the
development of the Thuja seedling, low evaporation is not in itself a
sufficient reason for the absence of Thuja seedlings. Darkness results
from the dense canopy formed by the trees, but even in the darkest
places Thuja seeds may be found. ‘The evaporation from such a spot
where no Thuja seedlings were present was 155 cc. for the season of
1916. Inasmall opening nearby, where Thuja seedlings were actively
growing, the evaporation was 225 cc. The increased rate of evapora-
tion in itself could hardly be held responsible for the presence of seed-
lings in one case and not in the other. The development of seedlings
in openings tends to restore a dense canopy and thus to lower the
evaporation from the chamaephytic layer. When a clearing of con-
siderable size is made, the evaporation is increased to a much greater
extent, as in the case of atmometer No. 3, in Reese’s clearing with an
evaporation of 587 cc. Many Thuja seedlings were present.
~ Reese’s bog occupies a low rolling site at the head of Burt Lake.
A road and several trails improve its circulation. A comparison of
atmometers No. 33 and No. 35 brings out the effect of circulation.
Atmometer No. 33 on the ground beneath a canopy of Thuja so dense
as to prevent ground vegetation, gave 435 cc., a higher rate than 345 cc.
from No. 35 in the crown of a small Thuja seedling in an opening
nearby. In the latter case, the development of edge conditions in the

174 FRANK C. GATES

foliage of the trees around the opening greatly checked the circulation.
Likewise atmometer No. 34 by a small Thuja seedling in the marginal
foss at the foot of a high ridge, where air drainage was poor, gave 391
cc., a lower result than the 435 cc. from No. 33, which was further in
the bog, but free from the influence of ground vegetation owing to
the dense canopy of Thuja saplings.

Atmometer No. 36, run in a small relic bog in the aspens north of
Reese’s bog gave 335 cc. and atmometers No. 37 and No. 38, run ina
slashed bog to the east of Douglas Lake, gave 344 and 307 cc. re-
spectively. In each case Thuja seedlings were developing at a rapid
rate. Based upon one week’s record, the evaporation near a Thuja
seedling on the beach for the season of 1916 would have been 562 cc.
and 487 cc. at the edge of the beach thicket.

In the pine series, investigated during 1915, the evaporation varied
from 347 cc. in an open spot in an aspen grove, through 310 and 321
cc. on an exposed hillside, 197, 198, 196, and 187 cc. in the ordinary
aspen association to 178 cc. at the foot of a slope in the densest part
of the aspens. In each case the results express the conditions with-
stood by one to three-year-old pine seedlings of which there were large
numbers throughout the aspens. Pine seedlings easily withstand as
wide a range of conditions as the region presents. In no case therefore
could it be said that evaporation conditions were the determining
factor in their ecesis. The presence of all ages and sizes of pine trees
is excellent evidence of how well the pine is developing and in con-
sequence the succession is progressing. Aspen seedlings are abundant
in the open sandy ground. As they develop, the increasing shade and
the checking of the wind are instrumental in causing a decrease of
evaporation from the chamaephytic layer, for example, atmometers
No. 18 and No. 24, with rates of 315 and 196 cc. respectively.

The two atmometers run during 1915 in connection with the area
south of Bryant’s, the one in pine land and the other in hardwood land
—each of which was very openly vegetated—gave the following
results: The evaporation from the immediate vicinity of a pine seedling
in pine land was 513 cc. Atmometer No. 13, run in hardwood land
devastated by fire, one and one half months previous, gave a rate of
473 cc. for the same period. The fact that the evaporation rate was
473 cc. in the hardwood land, where maple seedlings were present,
and 513 cc. in the pine land, where pine seedlings were present, whereas
maple seedlings developed successfully under the highest rate (590 cc.)

RELATION BETWEEN EVAPORATION AND PLANT SUCCESSION 175

obtained in the region, means that the evaporation from the chamae-
phytic layer is not the fundamental factor in the ecesis of such seedlings.

The fact that the rate of evaporation from the chamaephytic layer
is decreased in the development of mesophytism has been demon-
strated by many investigators: Transeau (5) at Cold Spring Harbor,
Gleason and Gates (1) in Central Illinois, Fuller (2) near Chicago,
Weaver (3) in southeastern Washington and adjacent Idaho, and the
present investigation in northern Michigan all strongly bring out the
same conclusion. If there is a causal relationship between evaporation
from the chamaephytic layer and succession, which I believe no one
disputes, either the decreased evaporation causes plant succession
or plant succession causes a decrease in evaporation. Dr. Gleason
and I (1) made the latter interpretation. Fuller (2) says: “the de-
creased rate of evaporation caused by the heavier vegetation is the
direct cause of succession between different associations.’ The
data of the present investigation indicate that evaporation is changed
in the course of succession and not preceding it.

In the succession towards mesophytism a conspicuous feature is
the fact that the seedlings of the dominant species of a genetically
higher association commence their development under the conditions
furnished by the existing association. Instead of a change of evapora-
tion preceding the development of a different vegetation, that which
is controlling and changing the rate of evaporation from the chamae-
phytic layer is the invading dominant species which have successfully
withstood the conditions imposed upon them by the existing asso-
ciation. It is quite obvious that they can not change nor control
conditions before they are present.

An increase in density of an association, itself, likewise causes a
decrease in the rate of evaporation from the chamaephytic layer.
Except a change of dominant species obtain, however, succession has
not taken place. Variations in evaporation from typical stations of a
given association in a given area are not likely to be as great as the
difference obtained between two genetically related associations. If
one should add to Fuller’s statement, previously quoted, to have it
read: ‘‘The evaporation thus controlled and changed is one of the
principal factors in permitting the development of a different lower
story vegetation,” its validity could be readily appreciated for those
secondary species whose physiological limitations precluded their
development in the lower genetic association. The fundamental

176 FRANK C. GATES

thing in succession is the replacement of the dominant species of the
existing association by those of the invading association. Changes
in the flora of the ground laver are secondary events.

With these facts in mind, one can not dodge the issue that, in a
given local area, invasion takes place under the existing conditions.
With the development of the invading species the evaporation con-
ditions of the ground layer are changed, which is usually accompanied
by a change in the ground flora. In other words, a change of evapora-
tion conditions of the ground layer is the result and not a fundamental
cause of succession.

SUMMARY

1. Experimentation was carried on in the vicinity of Douglas Lake,
Michigan, during the summers of 1915 and 1916, with 42 standard
Livingston atmometers. The usual methods of experimentation and
calculation of data were employed.

2. As the investigation was carried on in a small area, the influence
of edaphic factors was not obscured by the action of broad climatic
factors.

3. Invasion, which is the initial stage of succession, must take
place under the conditions already existing.

4. The change of conditions coincident with mesophytic succession
brings about a decrease in the rate of evaporation in the ground or
chamaephytic layer.

5. In a given area, the differences in the amount of evaporation
under which seedlings develop are largely due to the surrounding
vegetation, which by its size and density controls the evaporation
beneath it.

6. The complete range of evaporation conditions present in this
region, namely, from bare ground to the mature forest, is completely
within the physiological limits of the seedlings of Acer saccharum,
Pinus strobus, Pinus resinosa, and Thuja occidentalis. Given suitable
soil conditions, maple seedlings will develop under evaporation con-
ditions at least 337 percent more xerophytic than the normal hardwood
forest, or 400 percent more xerophytic than the very dense forest.

7. Within their soil requirements and in the presence of light,
the establishment of the pine, beech-maple and Thuja bog associa-
tions—three of the most important tree associations in northeastern
North America—is independent of any particular conditions of eva-
poration. Consequently a decrease in evaporation is not a prere-

RELATION BETWEEN EVAPORATION AND PLANT SUCCESSION 177

quisite to succession. A change in dominant species in an area is
fundamental to succession.

8. The change in the rate of evaporation from the chamaephytic
layer is produced by the development in density of the invading vege-
tation. Being coincident with and not antecedent to it, the change in
evaporation is a result and not a cause of succession.

9. While it is necessary for certain species to develop under existing
conditions to bring about succession, other species, of narrower physi-
ological limitations, can not develop until conditions are brought within
their range. Such species are secondary species, unable to cause suc-
cession.

10. Even though evaporation conditions are within suitable limits,
succession will not take place unless the disseminuls of the dominant
species of a higher genetic association arrive and develop.

11. The average evaporation from the chamaephytic layer of the
average aspen association for 40 days during the summer of 1915, at
Douglas Lake, Michigan, was 4.9 cc. per day; for the normal density
of the beech-maple forest, 4.4 cc. per day; while the highest average
rate for the season obtained from open ground was 14.7 cc. per day.
For a single week the highest rate was 21.6 cc. per day.

For 47 days during the summer of 1916, the average evaporation
from the chamaephytic layer of a densely developed Thuja bog was
4.8 cc. per day. A rate of 26.6 cc. per day was recorded from an at-
mometer in open ground at the crest of the low bluff a short distance
from the laboratory.

List OF PLANTS MENTIONED, WITH AUTHORITIES
(Using “ Gray’s Manual,” 7th Edition. Where different, the names
used in Britton and Brown “Illustrated Flora,’’ 2d edition, are given
in parentheses.)

Acer pennsylvanicum L.

Acer rubrum L.

Acer saccharum Marsh.

Actaea alba (L.) Mill.

Agrostis alba L.

Agrostis hiemalis (Walt.) B.S.P.
Aralia nudicaulis L.

Aster laevis L.

Aster macrophyllus L.

Betula alba papyrifera (Marsh.) Spach.

(B. papyrifera Marsh.).

Betula lutea Michx. f.

Chamaedaphne calyculata (L.) Moench.

Clintonia borealis (Ait.) Raf.

Convolvulus spithamaeus L.

Cornus canadensis L. (Chamaepericly-
menum canadense Asch. & Graebn.).

Danthonia spicata (L.) Beauv.

Diervilla lonicera Mill. (D. diervilla
MacM.).

Epilobium angustifolium L. (Chamaener-
ton angustifolium Scop.).

178

Erigeron canadensis L. (Leptilon cana-
dense Britton).

Erigeron ramosus (Walt.) B.S.P.

Fagus grandifolia Ehrh.

Fragaria virginiana Duchesne.

Gaultherta procumbens L.

Habenaria: obtusata (Pursh) Richards.
(Lysiella obtusata Richards.).

Hieracium scabrum Michx.

Hieracium venosum L.

Larix laricina (DuRoi) Koch.

Ledum groenlandicum Oeder.

Lepidium virginicum L.

Maianthemum canadense Desf.
folium canadense Greene).

Medeola virginiana L.

Melampyrum lineare Lam.

Mitella nuda L.

Moneses uniflora (L.) A. Gray.

Oryzopsis aspertfolia Michx.

Panicum xanthophysum A. Gray.

Picea mariana (Mill.) B.S.P.

Pinus resinosa Ait.

Pinus strobus L.

Pirola asarifolia incarnata (Fisch.) Fer-
nald (variety not given in Britton and
Brown).

Poa pratensis L.

Populus grandidentata Michx.

Populus tremuloides Michx.

(Unt-

CARTHAGE COLLEGE,
CARTHAGE, ILLINOIS.

FRANK C. GATES

Prunus pennsylvanica L.f.

Pteris aquilina L. (Pteridium aquilinum
Kuhn).

Quercus rubra L.

Rhus glabra L.

Rubus allegheniensts Porter.

Rubus idaeus aculeatissimus (C.A.Mey.)
Regel & Tiling (Rubus strigosus Michx.).

Rubus triflorus Richards.

Rumex acetosella L.

Salix rostrata Richards.
Sare.):

Smilicina stellata (L.) Desf.
stellata Morong).

Solidago canadensis L.

Streptopus amplexifolius (L.) DC.

Streptopus longipes Fernald. (Included
with the following species in Britton
aid Brown).

Streptopus roseus Michx.

Thuja occidentalis L.

Tilia americana L.

Trientalis americana (Pers.) Pursh.

Trillium grandiflorum (Michx.) Salisb.

Tsuga canadensts (L.) Carr.

Vaccinium oxycoccus L. (Oxycoccus oxy-
coccus MacM.).

Vaccinium pennsylvanicum Lam.
cinium angustifolium Ait.)

Viburnum acertfolium L.

(Salix bebbiana

(Vagnera

(Vac-

LITERATURE CITED

. Gleason, H. A. and Gates, F. C. A Comparison of the Rates of Evaporation in
Certain Associations in Central Illinois. Bot. Gaz. 53: 478-491. 1912.

. Fuller, G. D. Evaporation and Soil Moisture in Relation to the Succession of
Plant Associations. Bot. Gaz. 58: 193-234. I914.

. Weaver, J. E. Evaporation and Plant Succession in Southeastern Washington
and Adjacent Idaho. Plant World 17: 271-294. 1914.

. Gates, F. C. The Vegetation of the Region in the Vicinity of Douglas Lake,
Cheboygan County, Michigan, 1911. Rep. Mich, Acad. Sci. 14: 46-106.

TOU:
. Transeau, E. N. The Relation of Plant Societies to Evaporation. Bot. Gaz.
45: 217-231. 1908.

€. Stan Gone puncerigy it

ee ao ee

vies

st A. oot

"i Sapublie a

oc we ae » “se ada

ND “ Le a x 2 yee sal
Some iu becordered wh en: pee oc

Ca

Gaels of ‘salen, The Seater will ee Senta num
_ have been lost i in the. mails. 4 E
er ae pacha | a nee ec

Co z ni ie C

“ansonian Insiip ~
rf cS

% MAY 10 pe}
AMERICAN \_. ee
JOURNAL OF BOTANY

VoL. IV APRIL, I9I7 No. 4

THE RELATION OF SOME RUSTS TO THE PHYSIOLOGY
OF THEIR HOSTS!

E. B. MaAINns

LY INTRODUCTION

The relation of the rusts to their hosts has long occupied the at-
tention of many workers, not only because of their economic impor-
tance, but more especially by reason of the extremely interesting
biological problems which they offer. Not only have the rusts afforded
a wide field for the study of the questions of immunity, susceptibility,
physiological varieties, heteroecism, etc., but they, together with a
few other groups such as the Peronosporales and Erysibaceae, make
up part of the group of fungi which de Bary has called obligate para-
sites. This group of fungi is characterized by the requirement of a
living host as the source of food supply. Saprophytes, on the other
hand, obtain their food from dead organic material. Between the
two classes are the intergrading facultative parasites and facultative
saprophytes, determined by the degree a fungus is independent or
dependent upon a living host. The saprophytic and facultative para-
sitic fungi have long been studied with attention to their food relations,
but most of the work upon the obligate parasites has been confined to
other lines, since the parasitic condition itself puts great difficulties
in the way of an investigation of the nutrition of the fungus.

One must not overlook the fact that there are two conditions in
obligate parasitism. In the first, we have the problems concerned
with immunity and susceptibility, a condition which is common to all
parasites whether obligative or facultative as well as to the facultative
saprophytes. The other condition is that which goes to produce the

1 Paper No. 156 from the Botanical Department of the University of Michigan.

179

180 E. B. MAINS

obligative relation and make it seemingly impossible for the fungus
to grow elsewhere than upon its host or hosts. Much work has been
done and a number of theories developed with reference to the first
condition; but concerning the second only a little work has been at-
tempted, and but few theories advanced. Some authors in an en-
deavor to explain such parasitism have expressed the idea that the
solution of the problem might be sought in the requirement of the
fungus for some special nutrient which only its own particular host
would be able to supply. What this nutrient might be, if such is the
case, would be of extreme importance. Failing to determine this, it
would be of not much less importance that some idea of its nature be
obtained. Since the obligate parasites are distinguished by the abso-
lute need of a living host for their food supply, it is from the host that
the evidence for the solution of such a problem must be sought, and
it is through the control of the various physiological activities of the
host that one may hope to do this. It was to this end that this work
was undertaken with the object of obtaining more data regarding the
factors which control the obligate condition and determining, if pos-
sible, the substances or class of substances which are necessary for
parasitism of this kind.

The work was carried on in the Cryptogamic Laboratory of the
University of Michigan during the years 1914, 1915, and 1916 at the
suggestion and under the direction of Dr. C. H. Kauffman to whom I
am under deep obligations for many helpful suggestions and stimu-
lating criticism.

Eo sHIsTORICAE

The early history of the parasitism of the rusts has been well sum-
i.arized by de Bary (1853), who was the first to study the rusts and
smuts with scientific accuracy. According to de Bary, early naturalists
such as Pliny, Theophrastus, Malpighi, Duhamel, Tillet, Tessier, and
Plenk considered rusts not as the cause but as the result of a diseased
condition brought about by atmospheric conditions. In the course
of time, they were looked upon as foreign material which was partly
the cause and partly the result of the disease. Later, the rusts were
recognized as fungi by Linnaeus and Persoon, but they were still con-
sidered as the product of a diseased condition due to an injury such as
the sting of an insect, etc. Unger (1834) believed that the rusts were
produced by disarrangements in the respiratory organs of the plant
due to which sap exuded into the intercellular spaces and there coagu-

&

RELATION OF SOME RUSTS TO THEIR HOSTS I8I

lated, thus forming the rust. The next great step in the direction of a
true understanding of the nature of the rusts was the recognition that
they were the cause of the disease. The believers in this theory first
concerned themselves with the study of the rusts as fungi and their
entrance into their hosts. Léveillé (1839) showed that the rusts did
not differ from saprophytic fungi in their development of mycelium and
fruiting bodies except that they were within the living host. Prevost,
according to the Tulasne brothers (1847), first observed the germina-
tion of rust spores. The Tulasne brothers and de Bary (1853, 1863)
showed that the germ-tubes of the rusts enter through the stomata of
their hosts and in some cases (the germ tubes of basidiospores) through
the cell wall.

It was de Bary (1853), however, who finally definitely established
that the rusts were parasites and that they were the cause and not the
result of the disease. He concluded that the ‘‘ Brandpilze,”’ in which
group he included both smuts and rusts, are to be considered as para-
sitic growths, since they arise from spores whose germ-tubes penetrate
the host, develop a mycelium within the host’s tissue, form spores,
and finally break through the epidermis and infect other plants. De
Bary in Die Brandpilze (1853, p. 109) defined a parasite as, ‘“‘solche
Pflanzen oder Thiere, welche auf lebenden Geschépfen existiren, und
ohne diese nicht bestehen kénnen, welche durch den Reiz, den sie
verursachen, durch die Nahrung, die sie dem Wohnorganismus ent-
ziehen, St6rungen in dessen Organfunctionen hervorrufen; diese
schwinden, sobald der Parasit entfernt oder getédtet wird.” In view
of this definition, de Bary’s work on the ‘Brandpilze”’ was hardlv
sufficient to establish the rusts as parasites, since he did not show th
they could not exist outside of living organisms. we

A rather exhaustive search of the literature of this period does not
reveal that any attempts were made to grow the rusts saprophytically.
The general opinion which is now held appears to have arisen from
the earlier idea that the rusts were diseased products of the host, first
non-living and finally living products. In part this assumption of the
obligate character of the rusts is due to the fact that they, unlike the
facultative parasites, are never found in nature growing on other than
living plants.

Among the later workers upon the obligate parasitism of the rusts
is Brefeld (1883, 1908), who believes that the growing of rusts sapro-
- phytically is merely a matter of technique. He was able to obtain

182 E. B. MAINS

secondary and tertiary spores from the basidiospores of some rusts,
but further development was prevented by contaminations present in
his cultures. Carleton (1903) used different media and a substratum
as nearly like a wheat leaf as possible and obtained only a little dif-
ference in the length of the germ-tubes. Carleton, however, does not
give an account of his methods. Ray (1901, 1903) has claimed to
have cultivated a number of rusts upon decoctions of the host and on
sterilized carrot and reports that in one case teleutospores were formed
from the mycelium which was produced. Ray, however, gives only a
very imperfect account of his methods. This coupled with the fact
that he does not give either the species of rust or the kind of spores
used subjects his results to criticism.

The germination of rust spores has received considerable attention
especially from later workers. Plowright (1889), Eriksson and Hen-
ning (1894), Ward (19026, 1903), Melhus (1912), Johnson (1912) and
myself (1915) have found that temperatures between 10° and 30° C.
are necessary for good germination and that the optimum temperature
is between I2 and 18° C. Fromme (1913) has shown that a saturated
atmosphere is necessary for abundant infection.

The factors controlling the inoculation of the host after the germi-
nation of the rust’s spore have been principally investigated by Ward
and his students. Miss Gibson (Ward, 1905) found that the germ-
tubes of a number of rusts are able to penetrate into the inter-
cellular spaces of plants other than their host without infecting.
Furthermore Ward (1905) noticed this in the case of immune varieties
of plants. He (19020) also observed that the germ-tubes of Puccinia
dispersa had a tendency to be negatively heliotropic and suggested
that this may be a factor aiding in inoculation. Robinson (1914) in
the case of Puccinia Malvacearum and Fromme (1915) and myself
(1915) in the case of Puccinia coronata have shown a similar reaction
of the germ-tubes. Balls (1905) believes that inoculation is brought
about by a hydrotropic stimulus which causes the germ-tube to enter
the stoma of its host. |

The relation of the rusts to their host after infection has occupied
the attention of anumber of workers. De Bary (1887) and Jost (1907)
have expressed the opinion that the predisposition of certain hosts
for certain parasites is to be sought in the nature of the food which that
host offers to them. De Bary (1887) and Tubeuf (1897) both have
remarked that the rusts appear to adapt themselves to their host cells,

RELATION OF SOME RUSTS TO THEIR HOSTS 183

causing but little injury, at least up to the time of spore formation.
Tubeuf, in the case of hypertrophies caused by rusts, thought that
there appeared to be ‘‘a closer symbiotic relation between the fungus
and its host branch than between the host branch and its main branch.”
He also noticed that in the case of some rusts the infected portion of
the leaf remained alive after the death of the surrounding tissues and
he looked upon this as a condition resembling that of some lichens.
Ward (1890, 1902c) suggested that the relation between Puccinia
dispersa and its host tends towards symbiosis and that the rust is not
destroying the protoplasm of the host, but is robbing the host of its
food supplies. Ward (1905) found that when the rust does attack
the host so vigorously that the protoplasm is killed it brings about its
own death and immunity for the host is produced. This condition,
he discovered, can be duplicated by starving the host and by so doing
starving the rust. His earlier work (1902a, I902c) had already
pointed towards this conclusion since he found that immunity did not
depend upon anatomical features such as number and size of the
stomata, hairiness, etc., and that mineral starvation, although it did
not produce immunity, reduced the number of spores which were
produced.

The relation of the rusts to the carbohydrate supply has been noticed
by a number of workers. Halsted (1898) found in the case of Puc-
cinia Podophylli that there is a collection of starch in infected regions
of the leaf. The centers of such areas however contained much less
starch than the margins. Robinson (1913) noticed much less starch
in those areas of the leaf which are infected with Puccinia Malvacearum.
McAlpine (1906) quotes the work of F. T. Shutt, who found that the
grain and straw of rusted wheat contained more protein and less
carbohydrates than the rust-free plants. Tischler (1912), working
with Uromyces Pis1, discovered that the portion of the host containing
the most mycelium of the rust also contained the greatest amount of
"sugar.

The effect of environmental factors such as soil, moisture, tempera-
ture, and light upon the relation of the rusts to their hosts has been
studied by a number of workers. McAlpine (1906) noticed that
nitrogenous manures retarded the ripening of grain, while phosphate
of lime brought about early maturity and enabled the grain to escape
the rust to some extent. Sheldon (1905) has reported that soils favor-
able to the host are also favorable to the. rust of carnations (Uromyces
Caryophyllinus). :

184 E. B. MAINS

Considerable difference of opinion has arisen concerning the effect
of moisture on the development of the rusts. -De Bary (1863) found
that vegetative development of Uromyces appendiculatus and its pro-
duction of spores was greatly increased by a humid atmosphere.
McAlpine (1906) reported that drainage increased the yields of wheat,
but did not decrease the rust. He also considered that irrigation late
in the season tended to make the grain soft and brought on the rust.
Stone and Smith (1899) and Blaringhem (1912) claimed that the rusts
‘ were favored by dryness. Sirrine (1900) and Buchet (1913), on the
other hand, claimed that the rusts were favored by wet soils. Smith
(1904) considered that a dry atmosphere retarded the development of
the rust within the host while a dry soil favored development.
Fromme (1913) found that for Puccinia coronata after infection has
taken place moisture appears to have no effect upon the length of the
incubation period.

The effect of temperature upon the development of rusts in their
hosts has been but little studied. Sheldon (1902) found that the
incubation period of Puccinia Asparagt was longer at an average
temperature of 69° F. during the winter months than at an average
~ temperature of 76° F. during the summer. It is likely that these
results were in part due to the difference in the amount of light present
in the two seasons. Fromme (1913) discovered that for Puccimia
coronata a temperature between 20° and 30° C. brings about a shorter
incubation period than a temperature of 14° to 21° C.

Fromme (1913) appears to be the only one who has definitely in-
vestigated the effect of light upon the development of the rusts. He
found that when oats inoculated with Puccinia coronata were placed
in darkness for a few days and then returned to the light, the incuba-
tion period was lengthened approximately by the time that the oats
were in the dark. Fromme states that this may mean a dependence
of the rust upon the transition products of photosynthesis and that
this may explain the obligate parasitism of the rusts.

II]. EXPERIMENTAL’ WORK

MATERIAL

Two rusts were employed, Puccinia coronata Cda. and Puccinia
Sorght Schw. Puccinia Sorghi was used in all of the experiments since
its host maize (Zea Mays L.) was easier to work with. Puccimia

RELATION OF SOME RUSTS TO THEIR HOSTS 185

coronata was employed wherever oats (Avena sativa L.) could be used
advantageously, and whenever time allowed. These two rusts were
kept going on their host plants and thus stock material was always on
hand. ‘The method used was to make a spore suspension of the uredo-
spores and spray them on the host by means of an atomizer. The
pots were then covered with a belljar and placed where the temperature
ranged between 14° and 25° C., a temperature of 20° C. being used
whenever obtainable. The belljars were removed after 24 hours.
When maize was used it was found to be advantageous to draw the
leaves gently between the fingers before inoculating as the leaves are
covered with a wax-like substance which causes the spore suspension
to roll off without wetting them. In some experiments, definite areas
were inoculated by placing spores on them with a spear-pointed needle
after the plant had been atomized. During the winter, the cultures
were kept in a greenhouse, where Puccinia coronata required renewal
-about every three to five weeks and Puccinia Sorght about every two
months. ‘This was done by inoculating freshly grown plants. During
the summer, the stock cultures were kept in a garden in the open,
where the rusts propagated themselves.

DEVELOPMENT OF THE Rusts
Puccimia coronata

The first signs of infection show in from five to seven days after
inoculation, when light green areas are formed on the leaves. About
seven to eleven days after inoculation, pustules appear in these areas
as small, yellowish swellings, which soon break through the epidermis
liberating the mass of uredospores. Teleutospores develop in about
twenty-nine to thirty-six days after inoculation, when uredospore
production has about ceased and the leaf is slowly dying and drying up.
They show as blackish zones usually at the margins of the infected
areas and their first appearance occurs towards the apex of the leaf,
which is also the part of the leaf which first begins to die.

Puccimia Sorght

' The first signs of infection on maize are also light-colored areas on
the leaves. These appear in about six to seven days after inoculation
and pustules develop soon afterward, in usually seven to ten days.

186 E. B. MAINS

The pustules make their appearance first on the lower side of the leaf
and are more abundant and larger there, often becoming confluent.

The development of the rust within the host can be followed by
sectioning day by day after inoculation. The method employed,
which gave very good results, consisted in sectioning the infected leaf
between pith. Very thin sections can be obtained in this way, if the
leaf is cut up and a number of thicknesses of the leaf are placed between
pieces of pith. The sections obtained in this way were mounted in
Stevens’s chloral hydrate and iodine (Stevens, 1911). The sections
are cleared by this mixture so that the hyphae of the rust stand out,
showing clearly the haustoria in the cells of the host. Chlorophyll,
which in untreated sections interferes with the determination of the
mycelium, is thus bleached out, and any starch which may be present
can be recognized. At the end of the third day, mycelium was already
found to be well developed. At this time, the amount of starch, which
was normally present in the parenchyma sheaths of the vascular
bundles, was only slight, or it was entirely absent. By the fourth day,
the mycelium had formed dense masses in the intercellular spaces of
the mesophyll of the leaf. None was found sending haustoria into
the epidermal cells, nor was any mycelium found in the vascular
bundle and its parenchyma sheath. The cells of the parenchyma
sheaths in the infected areas did not show at this time so large a quan-
tity of starch as those in the non-infected areas. By the fifth day, the
epidermal cells were invaded by haustoria from neighboring hyphae,
and the amount of starch was noticeably less in the parenchyma sheaths
of the infected areas. About the sixth day, pustules appeared. These
were formed from masses of mycelium just below the epidermis. No
mycelium was found in the vascular bundles or their parenchyma
sheaths at this time. Some starch was found in the parenchyma
sheaths of the infected areas, but these did not stain so darkly with
iodine as those of the non-infected areas.

The development of the rust progresses by a continued spread of
its mycelium and the production of more pustules. The infected areas
are however always limited in extent, varying from a millimeter to
about a centimeter in diameter. Often, in the case of heavy infec-
tion, whole leaves may become covered with pustules due to the union
of many infected areas. Mycelium is found throughout the inter-
cellular spaces of the infected areas, where it sends its hyphae into the
neighboring cells, forming the branched, finger-like haustoria, which

RELATION OF SOME RUSTS TO THEIR HOSTS 187

have been well illustrated by Evans (1907). The vascular bundles
are apparently never invaded by the mycelium and the parenchyma
sheaths which surround them only occasionally have haustoria in
their cells.

When limited areas of the leaf are infected, an interesting phenome-
non takes place. After the tissue has been infected for some time—
in some cases in so short a time as nine days—the ends of the leaf
beyond the infected areas begin to die and the regions immediately
surrounding become yellowish, while the infected areas retain the
green color of healthy tissue. The infected areas, of which there may
be two or three on the same leaf, often become surrounded by dead
tissue, except perhaps for the midrib itself. The infected areas them-
selves still retain their original green color, and sections show that the
cells of these regions have all the appearance of normal cells, except
for the presence of haustoria within them. They are turgid and filled
with green chloroplasts. The neighboring tissue, on the other hand,
is brown and the cells are shrivelled up and dead.

EFFECT OF TEMPERATURE UPON DEVELOPMENT
Puccinia coronata

Experiments 1 and 2.—Two experiments were carried out to dis-
cover the effect of temperature upon the development of Puccinia
coronata within its host. In each, six pots of oats were used. These
were inoculated by spraying with a spore suspension of the uredospores,
and, after being left for 24 hours under belljars, four pots of each set
were removed to a well-lighted room where the temperature averaged
15° and 13° C. respectively for the two experiments. Two from each
set were kept in a similar room where the temperature averaged 20° C.
The incubation period of the rust at 20° C. in both experiments was
g days. The incubation period of the rust at 15° C. in the first ex-
periment was 13 to 15 days, and the incubation period for the rust at
13° C. in the second experiment was 15 days.

The results of these two experiments thus indicate that low tem-
peratures retard the development of the rust in its host.

Puccinia Sorght

Experiment 3.—This experiment was carried out with Puccinia
Sorght on corn in the same manner as were the two preceding experi-

188 E. B. MAINS

ments. Three pots of corn were kept at a temperature averaging 20° C.
and three at a temperature averaging 13° C. The incubation period
of the rust at 20° C. was 7 days while that at 13° C. had an incubation
period of 13 days, showing that low temperatures retard the develop-
ment of Puccinia Sorght in its host. .

Experiment 4.—Six pots of corn were inoculated with uredospores
of Puccinia Sorghi and kept under a belljar at a temperature of 18° C.
for12 hours. Twoof the pots were placed under belljars in an Eberbach
electric incubator at 40° C. The outer door of the incubator was left
open, allowing light to enter, and the incubator was placed in an east
window. Ina similar manner two pots were kept in an incubator at
30° C. Even with the outer doors open these incubators maintained
a temperature varying only a few degrees. The other two pots of
corn were placed under belljars in a box about the size of the incubators
with the open side facing the window. These two pots were at room
temperature, which according to a thermograph averaged 18° C. The
belljars were removed once a day from all the plants in order to renew
the oxygen supply.

At the end of the fourth day, the two pots of corn at 40° C. were
dead. Pustules appeared on the plants at 18° C. in seven days.
The only sign of rust on those at 30°C. at this time was the greenish
spots mentioned before as remaining in infected areas when the rest
of the leaf is yellowing. Sections through these areas showed a my-
celium which was only sparingly developed. At the end of fourteen
days, no pustules had formed on the plants at 30° C. At this time,
most of the leaves were dead, only the upper still retaining a green
appearance.

These results show that a temperature of 30° C. or higher prevents
the development of Puccinia Sorghi in its host.

EFFECT OF HUMIDITY UPON DEVELOPMENT

The work, as far as carried out, was done with Puccinia Sorght.

Experiment 5.—Fourteen plants of maize were inoculated with the
uredospores of Puccinia Sorght and kept under belljars at 18° C. for
twelve hours. Four plants were then placed in a south window under
belljars, by which means they were kept in a nearly saturated atmos-
phere. The remaining ten plants were placed without belljars on the
table beside the other four. The earth in the pots of five of these ten

RELATION OF SOME RUSTS TO THEIR HOSTS: 189

was kept saturated with water. The other five were watered just
enough to prevent the plants from wilting. The humidity of the room
varied between 20 and 36 percent and the temperature averaged
24° C. At the end of nine days after inoculation the number of pus-
tules on the plants was counted.

The plants in dry air and moist soil averaged 51 pustules per plant.
These pustules, however, were small and not as well developed as in
the others. Plants in a dry atmosphere and wet soil averaged 151
pustules per plant. These pustules were large and well filled with
spores. Plants in a saturated atmospheric and wet soil averaged 371
pustules per plant. These pustules were large and difficult to count
as many of them had become confluent.

At the end of 25 days, the infected leaves of plants in dry air and
moist soil were all dead and dried up, and the new leaves were free
from the rust. The plants in dry air and wet soil had a few pustules
on a few old live leaves, but most of the infected leaves were dead.
The new leaves upon these plants were free from the rust. Upon the
dry, dead leaves of ‘these plants green areas surrounding the pustules
were still evident. The tissue surrounding these areas had the brown
appearance of tissue whose cells had disintegrated before they had
dried. This would indicate that the green areas surrounding the
pustules had died because of the drying out rather than because of the
effect of the rust. The plants in wet soil and saturated atmosphere
still had a number of live leaves heavily infected with rust. Some of
the leaves however were dead or dying. On the latter, the infected
areas remained green surrounded by yellow, sickly tissue or by brown
tissue composed of dead cells. The new leaves had a small number of
pustules showing that by this time some reinfection had taken place.

The development of Puccinia Sorghi, as shown by the number of
pustules produced, is thus favored by the saturated atmosphere on
the one hand and by the wet soil on the other. The length of the
incubation period, however, is not much influenced. In dry air, the
plants finally become free from the rust by the drying up of the infected
leaves; and reinfection does not take place since spore germination is
prevented in dry air. In a saturated atmosphere, the infected leaves
live for a longer time since they do not dry up; and reinfection also
takes place to some extent, since the spores produced are able to
germinate in the humid atmosphere.

190 E. B. MAINS

EFFECT OF MINERAL SALTS UPON DEVELOPMENT

Two experiments were carried out to discover the effect of mineral
salts upon the development of Puccinia Sorght. In the first, water
cultures were used, while in the second pure quartz sand watered with
the solutions was employed.

Experiments 6 and 7.—The solutions used in the first of the two
were Knop’s full mineral nutrient solution, as given by Jost (1907)
and Miss Wuist (1913), and solutions in which one of each of the eight
principal elements (Mg, Ca, K, Fe, S, N, P, and Cl) were lacking.
The full nutrient solution consisted of the following in 1,000 cc. of
distilled water:

MgSO, PER ree Stok Ue nee MAIN LPP, eM TON vet nrar ain ON AN Wed i 25g
Ca (NOs )o% keiths mars otc cee ee en aa 1.00 g
KO POR ie hs Rite os eG, ee eee 25g
|. @) geome Cree ee ON TRAM ieee eM eh Soa I2g
BEC eos oh oP aaat oka as ee ea cee trace

The nutrient solution minus calcium was made by substituting
KNOs for Ca(NOs)e. KsSOs was substituted for MgSO, to form a
solution minus Mg. Ca(HePO:s)2 and MgCle were substituted for
KH2PO, and KCl to form a solution minus K. MgCle was substituted
for MgSO, to form a solution minus S. Ca(HePQOu.)e2 was used in place
of Ca(NOs)e to give a solution minus N. KNOs was used in place of
KH2PO, to give a solution minus P. KNOs: and FePO, were used in
place of KCl and FeCl; to give a solution minus Cl. And FeCl; was
left out of the full solution to form a solution minus Fe.

In the second experiment, maize was planted in quartz sand which
had been thoroughly washed with distilled water and then dried.
Knop’s full mineral nutrient solution of three times the ordinary
strength was used. The other nutrient solutions were made up a little
different than in the preceding experiment in that, when an element
was omitted, the concentration of the other elements in the solution
was maintained except for the element it was replaced by. Thus for
example, in — Mg solutions .75 gm. of MgSO, was replaced by 1.04 gm.
K.SO, so that there was as much SO, present as before. The amount
of the replacing element, K, naturally increases. In the —Ca solution,
Ca(NOs)2 was replaced by Mg(NOs)2. and KNO3 so that there would
not be too great a preponderance of either Mg or K in the solution.

Kernels of corn were planted in battery jars containing 1.2 kg. of

RELATION OF SOME RUSTS,. TO THEIR HOSTs IQI

quartz sand prepared as above and the jars were then placed in the
greenhouse. The various solutions were used to water the plants.
After 35 days, the plants were inoculated and placed in a moist chamber
at 20° C. for twelve hours.

In both experiments sections through the infected areas showed
that from about .5-1.2 mm. on each side of the pustules, there was
no starch in the parenchyma sheaths, although it was present
in considerable quantities in the rest of the leaf. Beyond this for
about .2-.6 mm. on each side, the concentration of the starch gradually
increased until it reached the full concentration of the rest of the leaf.

TABLE I[
Effect of Mineral Starvation upon the Development of Puccinia Sorghi

Number of Number of Average No.

Condition of Plants Plante’ Used Plants Infected Pustules per

: Plant
Solution
:

Water Culture | Sand Culture Ba eae Renee puiteg | eons Sand Culture
Full solution. .| green green 2 5 2 4 197
== Cay eae yellowish dead 2 5 I O O
== INS asia P light green 4 5 4 4 31
al Sa ee z 7 ‘ 4 5 4 2 170
San [roa eae i . 4 5 4 2 8
i oes VER Rie green 4 5 2 a 22
Se: dark green a e 4 5 4 2 Te
BN Jyed oe ns i a ¢ 5 3 2 II
ced One cee. | green S 4 5 2 2 1

|

The results as given in Table I show that mineral starvation does
not prevent infection of Puccinia Sorghi but only that the amount of
rust as shown by the number of pustules is less. Starch is prevented
from forming in the immediate vicinity of the pustules.

EFFECT OF LIGHT UPON DEVELOPMENT

The effect of light upon development of the rusts was studied in a
set of experiments, the results of which are given below. A number of
pots of the plants were inoculated under the same conditions. Some
were then placed under belljars in the light, and the rest were covered
with dark cylinders. After a few days, the plants under the dark
cylinders were placed in the light and the incubation periods of each
recorded.

192 E. B. MAINS

Puccinia Coronata

Experiments 8, 9, and ro.—In these experiments inoculation was
accomplished as stated above and the results obtained are given in the
following table.

TABLE II

Effect of Light upon Development of Puccinia Coronata

Experiment | Pot No, | Time in Dark | Time in Light Incubation Retardation

Carat 10 days 10 days

Car 72 LO TOn

C192 5 days LO. EGY hai 5 days
8 Cra AS. oe: FON. Tove Bias a:

C175 TS4 died, no infection

C176 ES uae died, no infection

C177 | TOE 7-10 days

C178 LO 7-10 |“

Crr79 7a On ae 12-13°- "0, CaF.
9 C7 S107 Via ten, ors 12-13“ Ong

CAU7U se eOtee no infection

CENT C2 aAl., (207 oa" ig i

Carrey tl TOW? 10 days

Cri alt Hil iis

Cc L70s 7 “c 5 ‘c i ‘6 2-3 c
10 C1716" 4 Yai Ca ie eee 2-3 «C

COT alll 2OU ee no infection

C178 20— i. + “

The results of these experiments show that in the absence of light
the development of Puccinia coronata is retarded and if left in darkness
too long, the rust is killed.

Experiment 11.—In the preceding experiments, the plants were all
placed in a dark moist chamber after spraying with the spore suspen-
sion. Fromme (1915) and I (1915) have found that the germ-tubes of
the uredospores of this rust are negatively heliotropic. It seems pos-
sible from this that the retardation of the appearance of pustules
might have been due to a failure of the germ-tubes to enter the host
while in the dark. Then, when brought into the light, inoculation
might have taken place from spores whose germination had been
delayed. To test this a fourth experiment was set up in which some -
of the plants, after being sprayed with the spore suspension, were left
under belljars in the light for from one to five days and were then
placed under dark cylinders for various periods, after which some of
them were returned to the light. After spraying with the spores,

RELATION OF SOME RUSTS TO THEIR HOSTS 193
others were kept in the dark for various periods of time and were then
_ brought into the light.

TABLE III

Effect of Light Upon the Development of Puccinia Coronata. Comparison of Plants
Inoculated in Light and in Darkness

Pot No. Time in Light | Time in Dark Tocintene oma Retardation
OUT eE Oe. : I day 4 days 6 days 11 days 2 days
Cay 200). Pets Ais Gre 18 ee 2a
Care 7s... es 19+ “ no infection
he a Sie 5+“ S 2
ine 223 5 eae Ones Bay i. 14 days eee ey
Orig 24e Bre Ger: co 14 . Bee
CM 2S ols. op 9
CIV 726. ss Orig Or Si,

Como 7h Gales gr

C728 cc. Buea rae Toe! ue
Ci717-20 5 0k Ti Sees Omen;

(ONS 37 (oa Wri | Sie qr

Bis teas peas Bite 65°" Lhe se 2
Ot by Ae > ea 20+ “ no infection

Car73 ee, eee: Se 9 days

GSA es... bey Tae eae ae
Gear 25 hoc sh Cae rons: oh ae ge
C726... 6 Bie ke Ta i Ton jee a

Following this experiment, plants were inoculated in areas marked
with India ink and covered with dark cylinders. After two days, the
inoculated areas were sectioned and the sections treated with chloral
hydrate and iodine. Mycelium was found in some of the inoculated
areas, but had not developed to a very great extent.

From these results it is evident that infection of Avena sativa by
Puccinia coronata takes place in darkness as well as in light, although
apparently the amount of infection is less in darkness.

Puccinia Sorght

The first experiments upon the effect of light upon the development
of Puccinia Sorght were carried out in the same manner as with Puc-
cinia coronata. The results obtained in these first experiments
(Experiments 12-19) were not as clear cut as those obtained with the
latter. Seven out of eleven plants which were in the dark three to
eight days before being placed in the light had their incubation period
lengthened one to two days. The other four had no retardation of
their incubation period Four out of thirteen plants which were in

194 E. B. MAINS

the dark throughout the experiment became infected. Their incu-
bation period was lengthened only two days. The remaining nine of
“the thirteen, however, remained uninfected.

Experiment 20.—In this experiment the procedure was the same as
in the previous experiments, except that the plants were kept in the
dark for five days to exhaust them as much as possible of carbohy-
drates. In this case, two out of the seven plants which were put in the
dark for seven days did not have their incubation period lengthened
at all. The other five had their incubation period lengthened from
2 to 4 days, which is shorter than the time they were in the dark.
Infections occurred on two out of twelve of the plants in the dark for
the entire time and in these cases the incubation period was lengthened.
Since these results more nearly agree with those obtained with oats
(Avena sativa), it would appear that the reserve food supply of the
maize is to be considered as the cause of the disagreement. In the
case of maize, the endosperm furnishes considerable food to the plant
during the first month and by the time this is exhausted, the plant is
of such a size that considerable reserve food is stored up in the stem
and other organs of the plant. The next experiment was carried out
with the object of exhausting this reserve as nearly as possible before
inoculating.

Experiment 21.—In this experiment, young plants were used. In
order to control the reserve food supply of the host as much as possible,
maize was germinated in a moist chamber and after four days the
endosperms were dissected away. The plants were then planted in
quartz sand which had been moistened with Knop’s solution and were
left in the light until the leaves were out and the plants had taken on a
_ green color. Following this they were removed to the dark for three
days to exhaust the carbohydrates manufactured during this time.
All were finally inoculated with uredospores of Puccinia Sorght and
four kept under belljars and eight under dark cylinders. Five of the
latter were removed at various intervals and placed under belljars.
Three were left under dark cylinders throughout the experiment.

From the results of this experiment, it is evident that when the
reserve food supplies of the host are cut down to the minimum, the
incubation period of the rust is lengthened to a period corresponding
to the time that the host was placed in the dark and that when the
host is kept in the dark, there is no development of the rust. Not
only does this indicate a direct relation to the carbohydrate supply

RELATION OF SOME RUSTS TO THEIR HOSTS 195

TABLE IV

Effect of Light Upon Development of Puccinia Sorght. Plants Exhausted as Nearly as
Possible of Soluble Carbohydrates

Age of | Incubation Retard-

Plant No. Plant Endosperm Time iu Dark Time in Light Period ee
FC-E 50....| 13 days) without 5 days
(died) no infection
LE Wea rat eairepea I oe Sa - 7 days 7 days
P52 EG le | (ae reo a
15 Os Oona ee ea a 3°" S
MEE S4:....(13° "| : 3 days a 46 to. 3 days
FC-E Bona a. 13 im “6 3 6 6 éé 9 “6 2 66
FC-E BO lsat: 12 66 66 2 66 Gh 66 ine) é6 a 66
PC-E56...2|13. °°" . Geo 7 ite no infection
(died)
RC—E60, .2.)13) ce OF as 6 days F250aySs (50a
CB Ol. s.. (13° a Gis:
(died) no infection
| Go Dy er al a ee Da i: 10 days | :
ECLE By. 13 6b “ 13 ‘i ‘6 66

of the host, but the apparent exceptions in the previous experiments
due to the presence of reserve food in the host only strengthen this
conclusion the more.

EFFECT OF THE LACK OF CARBON DIOXIDE UPON DEVELOPMENT

The preliminary experiments to show the effect of the lack of
carbon dioxide upon development were not satisfactory, since the
plants used possessed an endosperm and derived their food supply
from it as was shown by the plants in a minus carbon dioxide atmos-
phere developing as well as those of the check. Infection occurred at
the same time as in the checks or a few days later. Since the experi-
ments with light, which were being run at the same time, pointed to
carbohydrates as factors in the development of the rusts, it was evident
that the plant must be deprived as nearly as possible of carbohydrates.
This was attempted in two experiments.

Maize was germinated in a moist chamber at 30° C. and at the end
of six days, when the plumule had reached the length of five to six:
centimeters, the endosperm was dissected away and the plants were
planted in small bottles filled with quartz sand and moistened with
Knop’s solution. The plants were grown in the light for a few days
until the leaves were expanded and chlorophyll had developed. They
were then placed in a dark chamber for three days to exhaust them of

196 E. B. MAINS

the carbohydrates which had been formed while in the light. Follow-
ing this, the plants were inoculated with uredospores of Puccinia
Sorght and placed in large-mouthed liter bottles.

A carbon-dioxide-free atmosphere was obtained in these bottles
by placing a strong solution of potassium hydroxide in the bottom of
each and the oxygen supply was maintained by connecting the bottles
with U tubes which contained a mixture of pumice-stone moistened
with a KOH solution and pieces of KOH. Checks were run which
were set up similarly, with the exception that KOH was omitted
(Plate IV, figure 1). All the joints and corks were coated with paraf-
fine. |

Two experiments were conducted in this manner, the results of
which are given below. |

Experiment 22.—In the first experiment, plants treated as above
were inoculated and six (FC—E 60-65) were placed in bottles having
a carbon-dioxide-free atmosphere and two (FC-—E 66 and 67) in check
bottles. The set was kept in the dark for 24 hours to allow the ap-
paratus to become free from carbon dioxide and then was placed in
the light of an east window at an average temperature of 22° C. In
four days, numbers FC—E 60-65 began to show the effects of the lack
of carbon dioxide by their sickly appearance. Numbers FC—E 66
and 67 remained fresh and healthy. Pustules first showed on Num-
bers FC-E 66 and 67 in eight days. The plants in carbon-dioxide-free
atmosphere at this time showed no signs of infection. Numbers
FC-E 64 and 65 were dead and the upper parts of the leaves of Num-
bers FC—E 60-63 were also dead.

The experiment was finished on the eleventh day. At this time
no infection had taken place upon any of the plants in the carbon-
dioxide-free atmosphere.

Experiment 23.—Eight plants (FC—E 68-75) which were treated
as in the preceding experiment were inoculated with spores and placed
under a dark cylinder for 12 hours at 18°C. Six of these (FC—E 68-73)
were then placed in bottles with a carbon-dioxide-free atmosphere
and two (FC-E 74, 75) were used as checks. The bottles were all
kept in the dark for 12 hours so that the plants might not manufacture
carbohydrates while containing carbon dioxide. Pustules appeared on
both the checks in six days and on the seventh day pustules appeared
on FC-—E 69 and 73 in small numbers. These two plants, however,
appeared fully as healthy in every way as the checks, while FC-E

RELATION OF SOME RUSTS TO THEIR HOSTS Oy

68, 70, 71, 72 plainly showed the lack of carbon dioxide in their lighter
green leaves, which were dying back from the tip.

_ The experiment was finished at the end of the twelfth day. There
was no infection on FC-E 68, 70, 71, 72, which, although they showed
indications of approaching death, were still alive. Not even greenish
spots showed on the yellowing leaves such as often appear upon dying
leaves in infected areas. FC—E 69 and 73 at this time had a number
of open pustules and the plants themselves were in every particular
as healthy as the checks, showing that the apparatus in these cases
was defective and did not eliminate entirely all the carbon dioxide,
or that the plants possessed a large enough amount of food at the
beginning of the experiment to supply them.

These two experiments show, as do those with light, that there is a
relation between the development of the rust and the carbohydrate

supply.

EFFECT OF SUPPLIED CARBOHYDRATES UPON DEVELOPMENT

Since the preceding work indicated a relationship of the rust to the
carbohydrate supply, further experiments were undertaken in order
to study this relationship more thoroughly. To do this it is necessary
to supply carbohydrates to the host which has been deprived of them
as nearly as possible, since the experiments with light indicate that
the host normally contains more or less carbohydrates in available
form for the use of the rust.

This work divides itself into two parts according to the manner of
supplying the carbohydrates to the host. In the first case, carbohy-
drates were supplied to the plant through the scutellum of the maize
seedling and through the roots. Van Tieghem (1873) has shown that
embryos can be developed upon starch. Brown and Morris (1890)
have found that not only do excised embryos develop normally upon
starch, but that they will also do so with sugar solutions, especially
cane sugar, and that when both starch and sugar are present, the sugar
is used up before the starch is attacked.

A number of authors have investigated the ability of plants to take
up carbohydrates through their roots. Mazé and Perrier (1904) ob-
tained a good growth of maize in I percent glucose and sucrose.
Acton (1889) found that in this way acrolein, acrolein-ammonia, :ally]
alcohol, glucose, acetic aldehyde, ammonia, glycerine, laevulenic acid,
calcium laevulinate, cane sugar, inulin, dextrins, glycogen, and extract

198 E. B. MAINS

of natural humus can be used by a number of plants when in a carbon-
dioxide-free atmosphere. J. Laurent (1897, 1898, 1904) investigated
this subject thoroughly for corn. He found that glucose, invert sugar,
sucrose, soluble starch, and starch can be taken up by corn through
its roots and utilized in a carbon-dioxide-free atmosphere and to a
somewhat less degree in a normal atmosphere in the dark.?

In the second case, cut pieces of leaf were floated upon various
carbohydrate solutions. A number of workers have shown that por-
tions of plants may utilize sugar from solutions in which they were
placed. Boehm (1883) showed that cut pieces of leaf of Phaseolus
multiflorus upon solutions of cane sugar and glucose form starch in the
dark. E. Laurent (1886) has shown that etiolated potato sprouts
placed in the dark with their cut ends in a Io percent cane-sugar solu-
tion grow for more than five months and form starch. A. Meyer,
according to Acton (1889), found that shoots, when supplied with
dextrose, glycerine, sucrose; and inulin, can form starch. A number
of other authors have carried out similar experiments with like results.

CARBOHYDRATES SUPPLIED TO SEEDLINGS

In order to grow plants in carbohydrate nutritive solutions, it is
necessary to have both the plants and solutions as nearly sterile as
possible or the solution will be quickly filled with the growth of a
number of saprophytic fungi. The only feasible way to obtain sterile
seedlings is to sterilize the seed. A number of methods have been
used by different workers for sterilizing seeds, but almost all of these
methods have been found inefficient in some particular by other
workers.

Ward (1902d) used ‘‘various antiseptics’’ and also heated the seeds
to 60-70° C. for the purpose of sterilizing brome seeds. These were
then placed in sterile petri-dishes to germinate. Laurent (1897, 1904)
used .2 percent HgCl. solution for 1144 to 3 hours for corn. J. K.
Wilson (1915) has recently reported good results for a number of seeds
from the use of a solution of chlorine obtained from bleaching powder.
A number of other antiseptics have been used, such as Hz SOs, CuSO,,
H2O2, phenol, HNOs, etc., none of which have been found generally
or uniformly successful. Of these H2SOu, HgCle, and the calcium
hypdchlorite method of Wilson were tried by myself.

2 Since the above was written Kundson (1916) has also shown that maize is
able to take up through its roots dextrose, laevulose, maltose, and sucrose from
their 2 percent solutions with an increase in growth and dry weight.

RELATION OF SOME RUSTS TO THEIR HOSTS 199

In the use of the first, grains of yellow dent corn were cleaned and
dropped into concentrated H:SO. for ten minutes. They were then
removed to a capsule of sterile distilled water and washed once or
twice with sterile distilled water. They were then placed in sterile,
moist chambers. This treatment with H:SO. did not appear to
injure the seed in any way. On the other hand, it hastened germina-
tion of the seed. But it was not effective in killing the particular
fungus spores which were present on these occasions and in nearly
every case, fungi developed in the moist chambers.

The use of HgClze was found to be much more satisfactory and it
has been used entirely in this work where sterile corn plants were
needed. The method used was to clean the seeds and drop them into
.5 percent HgCle solution where they were left for thirty minutes.
The HgCle solution was then poured off and replaced several times
with sterile distilled water. The corn was then removed with sterile
forceps and placed in large (4.3 X 25 cm.) sterile test tubes, which were
placed in an incubator at 27°-30° C. This method gave very good
results throughout the work and only a small number of contaminated
seedlings were found.

Wilson’s calcium hypochlorite method was tried, but it did not
give very good results. Ten grams of ‘“Acme’’ chloride of lime
(bleaching powder) containing 30 percent chlorine were mixed with
140 cc. of water according to Wilson’s directions and maize was treated
for 9 hours with the filtrate. Only very slight germination and poor
seedlings were obtained by this treatment. The corn used however
was a little over a year old and this may account for the failure, al-
though it gave very good germination when treated with HgCl:, as
stated above. To test this method still further corn was taken out of
the sterilizing solution every hour for eight hours and placed in sterile
moist chambers and then placed in an incubator at 27° C. Good
germination took place with corn treated for one and two hours, but
when treated beyond that time the germination was poor. All the
moist chambers contained more or less contaminated seedlings. This
method when used with oats and wheat gave good results, while HgCl.
as outlined above was unsatisfactory with oats and wheat.

Although the number of disinfecting solutions used was not very
great, yet the results obtained point to the uselessness of trying to
obtain a disinfectant which will work for all seeds under all conditions.
Not only will the kind of seed but also the age of the seed, the amount of

200 E. B. MAINS

water contained in it, the permeability of the seed coat, and the kind
of fungus spores on it—all variable factors—have important bearing
upon the effect of sterilizing solutions upon the seed. It is very un-
likely that any one solution will work effectively under all combi-
nations of these conditions. Such a situation means that each worker
must select the agent which is the best suited to meet the requirements
of his particular conditions.

The corn after having been sterilized with HgCl. was germinated
in large sterile test tubes. The necessary moisture was maintained
in the tube by having absorbent paper saturated with distilled water
in the bottom of the tube when sterilized. The tubes were kept in
the dark in incubators at a temperature of 27°-30° C. for about seven
days, at which time the plumule of the corn had attained the height
of 5-15 cm.

The endosperm of corn contains a quantity of nutriment which
can nourish the plant for about a month and in fact when the plant is
grown in the light, the endosperm often lasts much longer, even two
months. It is therefore necessary to remove the endosperm before
placing the plants in nutrient solutions, since it would furnish all the
plants of the experiment with a large carbohydrate supply. This is
done by removing the plant from the test tube with sterile forceps
and making a longitudinal cut through the endosperm down to the
scutellum. The action of diastase has by this time dissolved away
the portion of the endosperm lying next to the scutellum and the two
halves of the endosperm are easily removed with sterile forceps, leaving
the scutellum surface exposed. The plants are then placed in their
nutrient solutions.

The following solutions were used: Cane sugar 15, I2, 6, and 3
percent; cane sugar 10 and 3 percent plus Knop’s mineral nutrient; |
cane sugar 10 percent plus Knop’s mineral nutrient minus nitrogen
(see Experiment 6); starch jelly 15 percent; starch jelly 15 percent
plus Knop’s mineral nutrient; starch jelly 15 percent plus Knop’s
mineral nutrient minus nitrogen; dextrose 3 percent; dextrose 3 percent
plus Knop’s mineral nutrient; maltose 3 percent; maltose 3 percent
plus Knop’s mineral nutrient; dextrin 3 percent; dextrin 3 percent plus
Knop’s mineral nutrient; Knop’s mineral nutrient; Knop’s mineral
nutrient minus nitrogen; distilled water.

Erlenmeyer flasks of 150 cc. capacity were used to contain the
solutions. These were stoppered with cotton plugs and autoclaved

RELATION OF SOME RUSTS TO THEIR HOSTS 201

at 110° C. for 30 minutes. It was found that there was much less
contamination of cultures if the flasks were autoclaved just before
using and allowed to cool in the autoclave. If this was not possible,
the flasks were usually wiped off with a corrosive sublimate solution
before using, since the roots of the corn plants often touched the out-
side of the flasks while they were being placed in the solution due to
the small neck of the flask. The cotton plugs used to stopper the
flasks were replaced after flaming and as they were made rather loose
and somewhat larger than the necks of the flasks, they fitted rather
closely around the corn stems (Plate IV, figure 2).

The flasks with plants thus prepared were placed in a moist dark
chamber. This chamber was prepared by covering a galvanized tank
a% X11 X 3 ft.) with a cover made of heavy black paper. The
cover was a little larger than the tank and reached to the bottom on
the sides, so that light was excluded and ventilation permitted. A
layer of water was kept in the bottom of the tank while the work was
being carried on in the greenhouse; this was enough to maintain a
saturated atmosphere in the chamber. Later when the experiments
were conducted in the drier air of the laboratory, coarse woven cloth
which was kept wet was spread over the tank under the black paper
lid in order to maintain the saturated atmosphere,

After the plants had been in this dark chamber for several days,
they were inoculated by spraying with a spore suspension and in
addition a small quantity of spores was placed on certain leaves.
They were kept at 20° C. for 24 hours, at which temperature the uredo-
spores of Puccinia Sorght germinate vigorously.

The results of Experiments 24 to 31 are given in the following table.
In all the cases where infection took place upon plants in Knop’s
nutrient solution or distilled water, the pustules were poorly developed
and few in number. In the three experiments where such infection
occurred, the infected portion of the leaves were cut off and the plants
were reinoculated. No infection took place the second time on such
plants in distilled water or Knop’s nutrient. Plants in sugar solution,
however, were infected, although to a somewhat less extent than in
the previous experiments, The reinfected plants were left until all
of them died. It was found that the plants which were infected in
Knop’s and distilled water in the original experiments lived as a rule
longer than the others. This would indicate that infection in the
original experiments upon these plants was due to a supply of food

202 E. B. MAINS

TABLE V
Effect of Carbohydrate Supplied to the Seedling Upon the Development of Puccinia
Sorght
‘ No. of Plants | No. of Plants | Av. Growth per
Solution Used Infected Plant per Day
Starch. Tsong cece ate eee 10 9 3.4 mm
Starch 1506" Kop Stine eee eee 10 3 EO: 2)
Starch 15° =p nop.s = Nie Ae eer: 10 5 Ser a
Cane Suga ris Tots. shveina: «ee yee 12 3 Os
Cane Suvari? Joi. 22078 aout eaten 6 3 er ae
Cane ‘sugar 109, Knop-se 2. eine 12 6 cs ie ele
Cane sugar: 10% --.Knop s.— N «vey e 13 5 2.7 tam
Cane SUC AT OU sui -ae hie ade tee oe ete 6 4 Ono
Wane sugar, 20. shinier Oe eee ee 21 13 TOES
Cane sugar 3%. = KnOp Sais A see one 21 Ea Ebsaraiy
Naltose:3 Oi aa © pach ee acca era een eee 7 5 AO
Maltoser305 --aiGnoprss fh van pee 7 4 Q.rsi"
IDEXEEIING V/A 2 tee Ws Gi ee Ee 6 3 TIO-ony
Dextrin. 3% KnOp Seis gon cee ee 7 2 12.6)"
Dextrose 39/7 le Seka sere eee ee 18 5 FN a
Dextrose:305.4— Knoop Si hs seikesa cass 18 9 Ossi Ve
Knop smutrent ta 30 atee a nee bane ae 50 3 20s.
Knop sinutrient-—"°Ne Sear ee ea eae 13 0) ee
Distilled: waterce¥ Ancacin cea Ss se cae 45 8 IAS a kt

present in the host upon which the rust as well as the host was able to
draw. When reinoculated this was exhausted and there was no infec-
tion of hosts in either distilled water or Knop’s solution. The results
of these experiments indicate that soluble carbohydrates are necessary
for the development of the rust.

CARBOHYDRATES SUPPLIED TO PIECES OF LEAF

In the earlier part of this work contaminations occurred which
were due to working with imperfectly sterilized leaves and especially
with rust spores having saprophytic fungus spores mixed with them.

It was evident that to obtain trustworthy results not only sterile
host plants were necessary, but that pure cultures of the rust must
also be obtained. This was done as follows:

Pure Cultures of Puccinia Sorght

The only worker who has given any account of a method to grow
rusts in pure culture appears to be Marshall Ward (1902d) working
with Puccinia dispersa upon the bromes. His method consisted in
obtaining sterile cultures of the bromes by sterilizing the seed by

RELATION OF SOME RUSTS TO THEIR HOSTS 203

“steeping in various antiseptics, or by heating to 60-70° C.”’ The
sterile seeds were placed in sterile drying towers, supplied with Knop’s
mineral solution and aerated with a continuous current of air or were
placed in large sterile test tubes which contained the solution. When
the first leaf was well developed it was inoculated with uredospores of
Puccinia dispersa. Good infection was obtained in the inoculated
area.

Ward’s object in developing this method was to be sure he was
working with only one race of Puccinia dispersa and not so much to
free the rust from saprophytic fungus spores. He does not say that
the spores with which he inoculated his sterile plants were free from
other fungus spores yet he assumes that he had a pure culture as far
as fungi were concerned. He says (p. 459), however, that the method
does not exclude harmless bacteria. In order to be sure that spores
of saprophytic fungi were not present in the sowing of the rust spores,
it would have been necessary to sow the spores of the resulting rust
upon nutrient media. This is necessary, since many saprophytic
fungus spores do not germinate except in the presence of such nutrient
media and so would remain dormant upon the surface of the plant in
the infected area and be removed with the spores of the resulting rust.

Two methods have been developed for obtaining pure cultures of
the rust. The first method is a modification of Ward’s. Large test
tubes (30 X 5 cm.) were prepared by filling the lower end with absor-
bent paper and adding Knop’s mineral solution. These were then
stoppered with cotton plugs and autoclaved at 10 pounds pressure for
30 minutes. ‘Two or three seeds sterilized in .5 percent HgCle solution
for 30 minutes were placed in each test tube and the tubes placed in a
well lighted window. After the seedlings had developed one or two
leaves, they were inoculated with uredospores of Puccinia Sorght
(Plate V, figure 1).

The uredospores were obtained from well-developed pustules of
Puccinia Sorght and were placed in a capsule of sterile distilled water
and thoroughly mixed up. The uredospores of Puccinia Sorght, as
well as most other rusts, are much lighter than water and float on the
surface, from which they can be removed with a looped platinum wire
and placed upon the leaves of the corn. The first trials to obtain
infection in this way were failures because of the lack of adhesion of
the water drops to the waxy surface of the corn leaves. In ordinary
infection work, the leaves are gently drawn between the fingers before

204 E. B. MAINS

inoculation to remove this waxy substance. Since this method would
cause contamination in this case, other means were resorted to. It
was noticed that drops of water often condensed upon the leaves during
cool nights. Spores were placed in these with resulting infection. As
carried out later, the leaves were rubbed with sterile cotton wrapped
around sterile forceps and soaked in sterile distilled water. In the
drops left adhering to the surface of the leaves, a loop full of uredo-
spores was placed. In this way very good infection was obtained. In
most cases, the spores used, after the combined dilution and washing
to which they were subjected, were probably sterile, but in order to
make sure that there was no contamination, spores from the resulting
infection upon these plants were used to inoculate other sterile plants.
These spores were taken from the other side of the leaf from that on
which the original inoculation had been made. By this means they |
were obtained from a sterile surface which had not been touched in
the original inoculation. The plants of this second series, when once
infected, will be reinfected by the spores produced on them as long as
they are in good condition, if kept at a temperature favorable for
spore germination (Plate V, figure 1).

A second means of obtaining pure cultures of Puccinia Sorght was
by means of cut pieces of leaf themselves. Uredospores which had
been removed from clean parts of infected plants were diluted in
_ sterile distilled water. Drops from this spore dilution were placed on
the surface of pieces of leaf which had been cut from sterile corn plants,
and floated upon carbohydrate solutions. A few capsules were.con-
taminated by saprophytic fungi, but more often capsules were obtained
which were free from these and the rust produced in such capsules
was used to inoculate pieces of leaf in like manner. By inoculating
fresh cultures about every two or three weeks, a pure culture of the
rust can be kept on hand (Plate V, figure 2).

Experiments 32—41.—These experiments were carried out with
pure cultures of both host and rust. Cut pieces of sterile corn leaf
were floated upon sterile solutions of 3 and 6 percent cane sugar, 3
percent cane sugar plus Knop’s, 3 percent dextrose, 3 percent
dextrose plus Knop’s, 3. percent maltose, 3 percent maltose plus
Knop’s, Knop’s, and’ distilled water. After one to three days,
uredospores from pure cultures of Puccinia Sorght upon sterile plants
were diluted in sterile distilled water and a drop of this spore suspen-
sion was placed on each piece of leaf. The capsules were placed in a
dark chamber at 20° C.

RELATION OF SOME RUSTS TO THEIR HOSTS 205

The results of these experiments are given in the following table.

TABLE VI

Effect of Carbohydrates Supplied to Cut Pieces of Leaf Upon the Development of
Puccinia Sorghi

: ‘ No. of Pieces of | No. of Pieces |

Solution | Leaf Used Infected Remarks
WanersiOaiiO on os .6s es Se hack | I2 6 _ pustules large
Gane seals ot fotene Gs oS eae | 31 IZ iy
Cane'sugar 3% -- Knop’s..:..... | 36 8 | , i
WDEMUGOSE: Bis, Seem eislie wunte Oe walle bs | 7 3 | a. ee
Dextrose 3% + Knop’s.......... 7 4 a os
IVUIEOSCRA GG crt Alen, chee ane ere | 9 5 a sf
Maltose 3% + Knop’s........... 10 2 3 us
noms mutiient./.-. 2. 2... vee ne 30 2 | pustules few, small
MD ISEMNCE WAtER a aren 1y) satin ole oO dueleea | 30 I 3 pustules, smal!

Of the pieces of leaf on Knop’s nutrient solution which were in-
fected, one piece had one very small pustule, while the other had three
very small pustules which were light in color, very unlike the brown
color of a vigorous rust. The infection upon carbohydrate solutions
varied from a few large brown pustules to a mass of pustules which
covered nearly all the area of the piece of leaf (Plate V, figure 2).
All the pieces of leaf which were on carbohydrate solutions had the
cells of the mesophyll and parenchyma sheaths filled with starch.
Pieces of leaf on Knop’s solution and distilled water showed no sign
of starch. At the end of 14 days, most of the pieces of leaf were alive,
as was shown by plasmolyzing with strong KNOs.

These results agree with those obtained with plants in nutrient
solution.

EFFECT OF NUTRITIVE SOLUTIONS UPON SPORE GERMINATION AND
CONTINUANCE OF GROWTH

In this work, the uredospores of Puccinia Sorght were sown upon a
number of different nutritive solutions. Compounds were used which
it was thought would likely be utilized by the rust in its host.

The method employed in the first two experiments was to remove
uredospores from an infected plant as carefully as possible and float
them on the surface of a sterile solution of the nutrient material to be
used. Hanging drops of this spore suspension were then made.

In order to avoid the work necessary for making a large number of
van Tieghem cells, the hanging drops were made upon the lid of a

206 E. B. MAINS

sterile Petri dish. Evaporation from the drops was prevented by
either having absorbent paper moistened with the nutrient solution
used in the bottom of the dish or by having a small amount of the
solution alone. If absorbent paper was used, a V-shaped piece was
cut out so that the microscope could be used to observe the develop-
ment of the rust in the hanging drops, each of which in turn could be
brought over the V-shaped opening by turning the cover upon the
bottom part of the dish.

The germination could be watched under the 16 and 8 mm. objec-
tives with clearness and the growth and condition of the germ-tubes
could be easily followed. Each Petri dish had ten hanging drops on
its lid and since three Petri dishes were used for each solution thirty
hanging drops were employed for each nutrient medium.

Some contamination resulted in these cultures, but most of the
hanging drops showed only a slight growth of saprophytic fungi during
the short time that cultures were run.

Experiment 41.—-The nutrient media used in this experiment were
conductivity water, cane sugar I percent, cane sugar 5 percent, maltose
I percent, maltose 5 percent, leucine I percent, asparagine saturated
solution, asparagine I percent and peptone (Witte’s) 1 percent. Be-
sides these, a mineral solution, and carbohydrates plus the mineral
solution were used. The mineral solution used was Duggar’s standard
nutrient solution (1909) for fungi minus the sugar. It consisted of the
following, dissolved in 100 cc. of water:

NH,NO3 eg Nr eee PO SAS RR yk Ris lat RO rms oe A NEPA Cy 1.00 gm
ICED RO aby Gata citar As has se eee eee 5 gm
MeSOn ies kc tern Oke ee seer | eee eres 25 gm
BC [ise Sitti 8 Pare rain Sol etre ke Pee Onan, Sam we trace

The cultures were kept at 17° C. during the experiment. The
following table gives the results of the experiment.
Since a dense mass of hyphae was produced during the germination
of the spores, the number of germinated spores could not be accurately
counted and the amount of germination was estimated by the appear-
ance. Inthe column under remarks the germ-tube is described whether
it produced short side branches or was unbranched. In all these
solutions, the rust was dead in about four days.

Experiment 42.—Since the preceding experiment indicated that
strong concentrations were injurious to spore germination and that

RELATION OF SOME RUSTS TO THEIR HOSTS 207

TABLE VII

The Effect of Nutritive Solutions Upon Spore Germination and Continuance of Growth
of Puccinia Sorght

Solution Germination ‘Length of Germ-tube Remarks
Condtctivity waters -.24...2%.. good | 400-500 micr. unbranched
Meneral nutrient... cil. once oe slight | 160 agi i
OamemstIC al <5 Ai us on as ac arene fair i =EOO—S00, |
WamMersUCat Los vi. ibs Scere: good P-400-s00 * | e
Canesugan 5% + nut....5 04... none |
MAIEOSE 5 Wiest ts ek kde a denn a's fair about 100“ os
INIEMIEOSeet Opi eet) leek wes good 300-800 ‘‘ | branched some
Maltose 5% + nut............ very slight about 100 ‘“ | unbranched
CDEOREET V6 O60 5.5 04 ee cite See fair ; * “500 -“ branched much
eptone 195 +-- nut... ii... eo none
Duggar’s nut + .5% peptone....| none
Asparagine saturated........... slight “160 “ | branched
ING MATACIME Ty tie. gence 4 6 odes fair HS -80=100. © ‘
J (eit CIWS A ar ne ry aa slight | 100-400 “| st slightly

they shortened the length of the germ-tube, an experiment was under-
taken to study the germination and development of the rust at low
concentrations. Two solutions, I percent cane sugar and I percent
cane sugar plus mineral nutrient (see Experiment 41) were used as the
basis. of this series. Dilutions of 1/2, 1/8, 1/32, 1/128, 1/512, -and
1/1024 of this strength were made. Hanging drops of uredospores
of Puccinia Sorght were made in these solutions as in the preceding
experiment.

No difference was noticed in the amount of germination in these
solutions. The length of the germ tube was the greatest (400-800
micr.) in 1/8, I/2 and 1 percent cane sugar. In the other solutions,
the length varied between 160-400 micr. ‘The rust died in all solutions
in about three days.

Experiment 43.—In this experiment, plant extracts were used.
A decoction of leaves of corn was made by autoclaving pieces of corn
plants (1 part by vol. to 5 parts of distilled water) at 10 pounds pres-
sure for 30 minutes. An uncooked extract of the plant was made by
cutting up sterile plants as finely as possible and adding sterile distilled
water to them and then letting the mixture stand for 24 hours. A
third extract was made from sterile germinated seeds which were cut
up in sterile distilled water. Two or three seeds were used for 25 cc.
of water. Uredospores of Puccinia Sorght from pure cultures of the
rust were sown in hanging drops and in capsules of the solutions.

The germination in all cases was good. In the decoction of the

208 E. B. MAINS

host, the tubes were long (about 800 micr.) and somewhat branched.
In the other two solutions, the germ-tubes were long and abundantly
branched. Death took place in all these solutions in about four days.

Experiment 44.—Cane sugar 3 percent, cane sugar 3 percent plus
Knop’s nutrient, dextrose 3 percent, dextrose 3 percent plus Knop’s
nutrient, Knop’s nutrient, and distilled water were used in this experi-
ment. Pieces of corn leaf were floated on these solutions and the
solutions were then autoclaved for 30 minutes at 10 pounds pressure.
This culture was run at the same time as Experiment 39 and inocula-
tion with uredospores of the rust was made in the same way.

Infection took place upon the living leaves upon the carbohydrate
solutions in Experiment 39 in eight days. No development of the
rust occurred upon any of the autoclaved leaves.

PVe) (DISCUSSION:

The first question of interest concerns itself with the condition of
the tissues in and around the region invaded by the rust. In the
development of Puccinia Sorghi, it is noticeable that, although most
of the cells of the leaf may be invaded by the large haustoria, yet no
harmful effect is shown by the host until after some period of time.
The rust sends its mycelium through the intercellular spaces and then
its haustoria into adjacent cells. The invaded cells retain the char-
acteristics of cells of uninfected tissues. The first sign of effect upon
the host is seen in the gradual disappearance of starch from the paren-
chyma sheaths in the invaded region. Since the parenchyma sheaths
serve as a storehouse for the assimilated material from the adjacent
region, and since they are not invaded for some time, it would appear
that this loss of starch is due, not to a withdrawal of starch from the
parenchyma sheath by the fungus itself, but to the utilization by the
fungus of the material formed in the neighboring region before it
reaches the parenchyma sheaths. That, even at this stage, the rust
is not attacking the host vigorously is shown by the development of
more or less starch in the parenchyma sheaths of the invaded region
depending upon the conditions of photosynthesis at the time of obser-
vation. ‘That the rust is having some effect is shown however by the
paler color when the parenchyma sheaths of the infected areas are
stained with iodine.

This condition prevails up to the time of spore formation. At this
time, the rust begins to draw more heavily upon the host in order to

RELATION OF SOME RUSTS TO THEIR HOSTS 209

obtain the necessary materials for spore formation. This is evident
in the smaller amount of starch present in the parenchyma sheaths
in the immediate region of the pustules. Oftentimes the parenchyma
sheaths are here entirely devoid of starch, while in the neighboring
region starch is present to a considerable extent. Yet even at this
time, the cells of the host do not show an injury such as one would
expect if the protoplasm itself was attached vigorously by the fungus.

It is only after the number of pustules have increased and spore
formation has continued for some time that the host begins to show
the effect of the rust’s presence. The effect of the rust, even now, is
not apparent in the tissues containing the rust, but in the neighboring
tissues as is shown by the green color of the infected areas and the
lighter green or yellow of the surrounding tissue. The green tissue
‘of the infected areas even at this time, may contain small amounts of
starch, but the neighboring dying regions have no indications of
starch. It would thus appear that the rust instead of attacking and
killing the cells of the tissue in which it is situated has a very different
effect upon them. While it is withdrawing food, at the same time it
stimulates the infected tissue so that this loss of food is in turn com -
pensated by the withdrawal of food from neighboring uninfected tissue.
It would appear that the rust thus destroys the symbiotic balance
between the cells of the host and causes some of them to have parasitic
relations with the rest. Marshall Ward (19020) and Tubeuf (1897)
observed this effect and considered it as evidence of a symbiotic rela-
tion between the rusts and their hosts.

As this withdrawal of food goes on the yellowing of the leaf extends
farther and farther from the green infected area, the cells of the region
gradually die and shrivel up, and the tissue takes on a brown appear-
ance similar to that of cells which have died due to a decomposition
of their contents. The infected areas, however, still remain alive for
some time, but in these areas death results from two causes. The
first of these is the cutting off of the food supply due to the death of
the surrounding tissue. This, however, is not probably the principal
cause as the green cells of the infected areas could furnish food to
prolong their life and that of the rust until by a process of gradual
starvation both would die. The principal cause which appears to
bring about the death of these areas is the drying up of the leaf as a
whole. As Sachs has pointed out the loss of water from dead tissue
is much greater than from living. The great evaporation from the

210 E. B. MAINS

surrounding dead tissue naturally withdraws water from the green
infected areas, which have their water supply from the root diminished,
and brings about death through drying out.

If conditions at this time are unfavorable for spore germination,
such as a low humidity or a too low or too high temperature, the corn
plant will be freed from the rust, since the spores formed upon the
old leaves will not be able to infect the young newly formed leaves.
With the death of the old leaves the host becomes free from the rust.
In the same way, oats may become free from their rust.

The work upon the effect of temperature upon the development of
the rust also throws some light upon the relation between the rusts
and their host. From Experiments 1 and 2 and especially 3 and 4,
it is evident that in a saturated atmosphere the development of both
Puccinia coronata and Puccinia Sorghi is retarded by low temperatures.

It is difficult to say just how much these results are due to the
direct effect of the temperatures upon the rust, since the rust must be
studied in connection with its host. A search of the literature shows
that but little work has been done upon the effect of temperature upon
the growth of parasitic fungi in their hosts. Sheldon (1902) found that
during the winter months the incubation period of Puccinia Asparagt
was longer than during the summer months when the temperature
was higher. Fromme (1913) found a shorter incubation period for
Puccinia coronata at temperatures between 20° and 30° C. than at
lower. Ward (19020) explains the non-infection in some of his experi-
ments by the high temperature, although the host seemed to be un-
harmed.

The effect of temperature upon the development of saprophytic
fungi has received considerable attention. In such experiments,
however, the nutrient media remained unchanged. In experiments
with the rusts, other conditions besides the temperature alter, since
the physiological conditions in the host are altered. It is consequently
hard to determine how much of the effect produced on the rust is due
directly to the temperature. Lehenbauer (1914) has shown that for
corn the optimum temperature for the growth was situated between
29 and 32° C. Sachs (1882) gives 27.2° C. as the optimum tempera-
ture for the growth of the root. Besides the effect upon the growth,
two of the physiological processes of the host are especially affected.
These are respiration and photosynthesis. The respiration of plants
increases with the increase in temperature until the injurious effect of

RELATION OF SOME RUSTS TO THEIR HOSTS 211

the high temperature brings about a disorganization of the vital func-
tions of the plant. Photosynthesis according to Pfeffer (1900) in-
creases with the temperature up to an optimum, which is approximate
to that of growth and then it falls. Matthaei (1905), on the other
hand, gives a curve for photosynthesis which resembles that of respira-
tion. This curve, however, is a curve of maximal pnotosynthesis for
each temperature. At high temperatures, the maximal photosyn-
thesis is maintained only for a short time. The higher the temperature
the sooner the decline sets in and the steeper its slope. The full
development of photosynthesis also needs the best conditions of illumi-
nation while respiration is not much affected. Consequently in
Experiment 4, the food supply available for the fungus at 30° C. is
much reduced from that which would be available for it at 20° C.,
since photosynthesis has fallen off and respiration has increased. At
the lower temperatures, although the respiration is lowered, photo-
synthesis is also reduced and to a much greater degree so that the food
available for the fungus is less than at the optimum temperature for
photosynthesis.

Although the growth of the fungus is thus influenced to some extent
by the effect of temperature upon the amount of food supply available
in the host, it is probably the direct effect of the temperature upon the
rust itself which is the most important factor in determining the de-
velopment of the rusts in Experiments I—4; especially since the tem-
peratures obtained for growth also correspond very well to those for
spore germination. The actual temperature of the fungi in these ex-
periments, however, was undoubtedly higher than that recorded, since
Matthaei (1905), Ehlers (1915), and others have shown that the in-
ternal temperature of leaves is one to ten degrees higher than the sur-
rounding atmosphere, depending upon the amount of illumination, the
presence or absence of air currents, and the amount of transpiration.

The evidence concerning the effect of moisture upon the develop-
ment of the rusts is rather conflicting. Blaringhem (1912, 1913) and
Stone and Smith (1899) claim that the rusts are favored by a dry soil.
Buchet (1913) believes that a wet soil is favorable. Sirrine (1900)
considers that dew is the controlling factor. Smith (1904) finds that
a dry atmosphere is unfavorable to Puccinia Asparagi, while a dry
soil is favorable. From the results of Experiment 5, it is evident that
for Puccinia Sorght wet soil and moist atmosphere bring about an
increase in vigor, as shown by the greater number of pustules and the

212 E. B. MAINS

consequently more abundant spores. A humid atmosphere also
lessens the transpiration from the dying portions of the leaf and the
evaporation from the dead areas so that the infected leaves in a moist
atmosphere have a much longer life. In drier air, the infected leaves
dry up and the plants become rust free.

The effect of mineral starvation upon the host has been investi-
gated principally by Marshall Ward (1902c). Sheldon (1905) and
McAlpine (1906) have made some observation upon the effect of soils
and manures upon the development of some of the rusts.

Results similar to Marshall Ward’s were obtained in Experiments
6and 7. Infection was obtained upon some plants in all of the mineral
nutrients. Table I shows that the best infection was obtained upon
plants which were furnished with a full mineral nutrient. A deficiency
of an element, however, does not bring about immunity for the host,
but it causes a smaller number of plants to be infected and a lessening
of the amount of rust as shown by the number of pustules.

That this effect is produced upon the rust is due partly to a lack
of these elements for the host and a consequent slow mineral starvation
of the rust. This, however, only explains a portion of the effect pro-
duced, since the rust can probably obtain these elements in small
quantities from the host as long as the host is alive. A portion of the
effect produced is at least to be referred to the effect upon the physi-
ology of the host and its greatly reduced ability to manufacture the
proper food materials for the rust. That this is true is shown in
Table V. In this table it is seen that in solutions in which the host
was supplied with a mineral nutrient solution, but was not supplied
with carbohydrates and was kept in the dark to prevent their manu-
facture, the number of plants infected were few. That even this
number was infected was due to the fact that the host was not ex-
hausted of soluble carbohydrates, such as the sugars, before infection.
The infection of plants supplied with carbohydrates in all cases far
outnumber that of the plants without carbohydrates.

The need of carbohydrates is also shown in Experiments 8-21
conducted upon the effect of light upon the development of the rusts.
Of the two rusts Puccinia coronata upon oats shows the closest relation
in this regard. Thus, in Table II, it is shown that the retardation in
infection approximates closely the period that the host was left in
darkness and consequently the period during which carbohydrates
were not being formed. In all cases where the host was kept contin-
uously in darkness after inoculation there was no infection.

RELATION OF SOME RUSTS TO THEIR HOSTS 213

That these results are due to the prevention of photosynthesis and
not either to non-inoculation due to the negative heliotropic germ-tube
of Puccinia coronata or to the effect of the lack of light upon the develop-
ment of the mycelium of the rust is shown in Experiment 11. Since
it has been shown that the germ-tubes of the uredospores of Puccinia
coronata are negatively heliotropic, the explanation of the retardation
of the incubation period might be an inability of the germ-tube to enter
the stoma while in the dark. This is not the case, as shown by the
fact that when plants (C 17.19-C 17.27, Table III) after having been
inoculated in the light were placed in the dark for a short period, the
incubation period of the rust on them was also lengthened. The
demonstration of the presence of mycelium in the leaves of plants
inoculated and kept in the dark finally establishes this.

In Experiment 11 there is one case (C 17.35) where the retardation
of the incubation period was greater than the time during which the
host was in the dark. In this case, the oats were in the dark for eleven
days and were consequently so starved and their physiological proc-
esses so disarranged that when returned to the light, they carried on
their physiological processes poorly and so were able to furnish the
rust with only a small amount of food.

The relation of Puccinia Sorght to the carbohydrates formed in the
light is not at first glance so striking as in the case of Puccinia coronata.
Heckel (1912) and Blackshaw (1912) have however shown that corn
plants contain from 6 to 9 percent of sugars. Besides this, corn forms
starch in the parenchyma sheaths so that under ordinary conditions
it contains quite a considerable reserve of carbohydrates. Oats, on
the other hand, contain no such reserve and consequently the rust
quickly shows the effect when the daily supply is cut off. When means
were adopted to decrease the sugar content of corn the results compare
more nearly with those obtained with Puccinia coronata on oats. It is
very evident from Experiments 12-20 that the rust itself does not have
any direct relation to light, for infection took place and the rust devel-
oped to spore formation in the dark. The same is also shown much
better in experiments where carbohydrates were supplied to the host
in the dark.

The relation of the rust to the carbohydrate supply is further seen
when the host is deprived of its carbon dioxide supply (Experiments
22 and 23). For when the host is prevented from manufacturing
carbohydrates by surrounding it with a carbon-dioxide-free atmos-
phere, there is no development of the rust.

214 E. B. MAINS

A comparison of the infection of corn when supplied with car-
bohydrates alone and when supplied with carbohydrates plus Knop’s
nutrient shows varying results. In spite of the varying results, the
experiments clearly show that in all cases there is a much greater
infection when carbohydrates are supplied than when there are no
carbohydrates present. It is probable that with the slow growth of the
host the carbohydrates taken up by the host united with the nitrogen
compounds formed in the metabolism of the host and formed proteins
for the use of both host and fungus and so the rust did not feel the loss
of the mineral elements to a very great extent.

It appears from experiments in which the host was deprived of its
soluble carbohydrates that the rust was not able to live upon the host
even though the host was alive and consequently did not use the pro-
toplasm of the host, at least directly, as food. Marshall Ward (1904),
especially, among the workers upon the rusts has- noticed that the
protoplasm of the host does not appear to be affected by the rust. In
his Croonian lecture (1890), he points out that the relations of the
rusts to their hosts are very different from those of a facultative para-
site such as Botrytis. The rusts he considers as merely tapping the
food supply of the host, establishing a relation which approaches
symbiosis.

The results obtained agree with Ward’s view. The development
of the rust upon seedlings or cut leaves of the host furnished with
carbohydrates, and the non-development except in a few cases, upon
hosts furnished only with distilled water or mineral nutrient indicate
that the rust is dependent upon the food supply of the host and does
not live upon its protoplasm. The healthy development of host tissue
in the infected regions compared with the surrounding dying tissue also
shows that not only does the rust not live upon the protoplasm of its
host but it even stimulates it to greater activity.

It is possible that the lack of carbohydrates might produce prod-
ucts in the host which are toxic to the rust. Thus amino acids and
other products of metabolism might form which in the presence of
carbohydrates would unite with them to form proteins again. These
products might then inhibit the development of the rust. |Experi-
ments 38-40 indicate, however, that such is not the case. In these
experiments where cut pieces of leaf were floated upon water and
mineral nutrient such products would have had plenty of chance to
diffuse out of the leaf. Yet on those solutions deprived of carbohy-
drates, there was no infection.

RELATION OF SOME RUSTS TO THEIR HOSTS 215

Thus the food material for which the rust depends upon the host
appears to be the sugars or some of the compounds which they enter
into during the formation of proteins, or what is more likely both.
Such being the case it should be feasible to grow the rusts saprophyt-
ically. So far as I know only two attempts have been made to do
this. These have been made by Carleton (1903) who obtained negative
results and Ray (1901, 1903) who reports having grown several in
culture. As has been pointed out above Ray’s results are open to
criticism due to the incomplete account of his methods and material.

In the case of Puccinia Sorght, I have not been able to find that
various nutrient media had any appreciable effect towards developing
a mycelium. Various carbohydrates and organic nitrogenous prod-
ucts with and without mineral salts and at different concentrations
showed an effect only upon spore germination and the length of the
germ tube (Experiments 41-43). Sterilized pieces of leaf upon solu-
tions such as were used successfully with living pieces of leaf, sterile
decoctions of the host and water extracts of the host gave no better
results. In all cases, the germ-tube produced by the spore lived only
a few days.

Since Puccinia Sorghi does not appear to be able to use the sugars
directly, but must have them supplied to the host for their develop-
ment, it appears that it is not the stable carbohydrates or proteins
which are to be considered as essential for the metabolism of the rusts.
Rather, it is probable that the rust is dependent upon some transitory
products in the formation of these substances, as Fromme (1913) has
suggested, or it may be that the rust is able to utilize such compounds
only in their nascent state, so to speak, when these complex organic
compounds are not in a condition of equilibrium. Even among the
saprophytic fungi, we have many which prefer certain stereoisomers
and it would not be at all surprising to have in the rusts a group of
fungi which needs certain isomers for their development. It is in
some such explanation very likely that the obligate parasitism of the
rusts is to be sought.

V. SUMMARY

1. The optimum temperature for the development of Puccinia
coronata and Puccinia Sorght is situated at about 20° C. and the maxi-
mum for Puccinia Sorght is about 30° C.

2. While Puccimia Sorght is not prevented from developing upon
the host under conditions of dry air and soil, moist soil and a humid

216 E. B. MAINS

atmosphere favor the development of the rust and increase the number
of spores formed.

3. Puccinia coronata and Puccinia Sorghi do not appear to injure
the cells of the infected area. The injury produced appears first in
the areas surrounding the infected regions. This is probably due to a
starvation brought about by a withdrawal of food from them by the
infected areas.

4. Astarvation of the host of various elements does not bring about
immunity from the rust, but reduces the quantity of the rust produced.

5. Light is not necessary for the development of Puccinia coronata
and Puccinia Sorght when the host is able to obtain a good food supply.

6. When deprived of carbohydrates, light is necessary for the
development of Puccinia coronata and Puccinia Sorght in that it is
necessary for the formation of carbohydrates by the host.

7. When deprived of carbon dioxide, the development of Puccinia
Sorghi is stopped due to a lack of carbohydrates in the host.

8. Pure cultures of Puccinia Sorght can be maintained upon both
sterile seedlings and upon pieces of Zea Mays leaf floated upon car-
bohydrate solutions.

g. Puccinia Sorghi develops and forms spores upon seedlings or
cut pieces of corn leaf when these are supplied with starch, cane sugar,
dextrose, maltose, and dextrin in the dark.

10. When either seedlings or pieces of corn leaf are exhausted of
carbohydrates and supplied only with mineral nutrient or water,
Puccinia Sorghi is not able to develop in the dark.

11. Puccinia Sorghi is not able to use maltose, dextrose, cane sugar,
asparagine, leucine, peptone with and without mineral salts, or decoc-
tions of. the host when supplied to it directly.

12. The obligate parasitism of the rusts is probably explained by
their requirement of some transitory or nascent organic products
related to the carbohydrates which they obtain in the Jiving host.

VI. LITERATURE CITED
Acton, E. H.

1889. The Assimilation of Carbon by Green Plants from Certain Organic
Compounds. Proc. Roy. Soc. London 47: 150-175.
Balls, W. L.
1905. Infection of Plants by Rust Fungi. New Phytol. 4: 18, Ig.
Bary, A de.
1853. Untersuchungen tiber die Brandpilze. Berlin.
1863. Recherches sur le développement de quelques champignons parasites.
Ann. Sci. Nat. IV. Bot. 20: 1-148.

RELATION OF SOME RUSTS TO THEIR HCSTS 217

1887. Comparative Morphology and Biology of the Fungi, Mycetozoa, and
Bacteria. English edition. Oxford.
Blackshaw, G. N.
1912. South Afric. Journ. Sci. g: 42-48. Abstract in Exp. Sta. Record 30: 14.
Blaringhem, L.
1912. Observations sur la Rouille des Guimauves (Puccinia Malvacearum
Mont.). Bull. Soc. Bot. France 59: 765-773.
1913. Sur la transmission héréditaire de la Rouille chez la Rose termiere
(Althaea rosea). Compt. Rend. 157: 1536-1538.
Boehm, J.
1883. Ueber Starkebildung aus Zucker. Bot. Zeit. 41: 32-38, 49-54.
Brefeld, O.
1883. Untersuchungen aus dem Gesammtgebiete Heft 5. Die Brandpilze,
Leipzig.
1908. Untersuchungen aus dem Gesammtgebiete Band 14. Die Kultur der
Pilze. Miinster.
Brown, H. T. and Morris, G. H.
1890. Researches on the Germination of Some of the Gramineae. Journ.
Chem. Soc. 57: 458-528.
Buchet, S.
1913. Sur la transmission des rouilles in général et du Puccinia Malvacearum
en particulier. Bull. Soc. Bot. France 60: 558-565, 520-524.
Carleton, M. A.
1903. Cultural Methods with Uredineae. Journ. Appl. Micr. Labr. Meth.
6: 2109-2114.
Duggar, B. M.
1909. Fungous Diseases of Plants.
Ehlers, J. H.
1915. The Temperature of Leaves of Pinus in Winter. Amer. Journ. Bot. 2:
32-70.
Eriksson, J. and Henning, E.
1894. Die Hauptresultate einer neuen Untersuchung tiber die Getreideroste.
Zeitschr. Pflanz. 4: 66-73, 140-142, 197-203, 257-262.
Evans, I. B. P.
1907. The Cereal Rusts. I. The Development of their Uredo Mycelia.
Annals of Botany 21: 441-466.
Fromme, F. D.
1913. The Culture of Cereal Rusts in the Greenhouse. Bull. Torrey Club
40: 501-521.
1915. Negative Heliotropism of Urediniospore Germ-tubes. Amer. Journ.
Bot. 2: 82-85.
Halsted, B. D.
1898. Starch Distribution as Affected by Fungi. Proc. Amer. Assoc. Advanc.
Sci. 47: 408, 409.
Heckel, E.
1912. De linfluence de la castration male, femelle et totale sur la formation

218 E. B. MAINS

du sucre dans les tiges du Mais et du Sorgho sucré. Compt. Rend.
155: 686-690.
Johnson, E. C.
1912. Cardinal Temperatures for the Germination of Uredospores of Cereal
Rusts. Abstract in Phytopathology 2: 47.
Jost, Ludwig.
1907. Lectures on Plant Physiology, English edition. Oxford.
Knudson, L.
1916. Influence of Certain Carbohydrates on Green Plants. Cornel! Agr.
Exp. Sta. Memoir 9.
Laurent, E.
1886. Starkebildung aus Glycerin. Bot. Zeit. 44: 151.

Laurent, J.
1897. Sur l’absorption des matiéres organiques par les racines. Compt. Rend.
125: 887-889.
1898. Absorption des hydrates de carbone par les racines. Compt. Rend. 127:
790; 7076

1904. Recherches sur la nutrition carbonée des plantes vertes a l’aide de |
matiéres organiques. Rev. Gen. Bot. 16: 14-48, 66-80, 96-128, 155-
166, 188-202, 221-247,
Lehenbauer, P. A.
1914. Growth of Maize Seedlings in Relation to Temperature. Physiol. Res.
I's 247-288.
Léveillé, J. H.
1839. Recherches sur le développement des Urédinées. Ann. Sci. Nat. II.
Bot. 11: 5-16.
Mains, E. B.
1915. Some Factors Concerned in the Germination of Rust Spores. Mich.
Acad. Sci. Rep. 1'7: 136-140.
Matthaei, Miss G. L. C.
1905. Experimental Researches on Vegetable Assimilation and Respiration.
Phil. Trans. London ser. B 197: 47-85.
Mazé, P. and Perrier, A.
1904. Recherches sur l’assimilation de quelques substances ternaires par les
végétaux superieurs. Compt. Rend. 139: 470-473.
McAlpine, D.

1906. The Rusts of Australia. Melbourne. .
Melhus, I. E.
1912. Culturing of Parasitic Fungi on the Living Host. Phytopathology 2:
197-203.

Plowright, C. B.
1889. A Monograph of the British Uredineae and Ustilagineae. London.
Ray, J.
1901. Cultures et formes atténuées des maladies cryptogamique des végétaux.
Compt. Rend. 133: 307-309.
1903. Etude biologique sur le parasitisme: Ustilago Maydis. Compt. Rend.
136: 567-570.

RELATION OF SOME RUSTS TO THEIR HOSTS 219

Robinson, W.
1913. On Some Relations between Puccinia Malvacearum and the Tissues of
its Host Plant (Althaea rosea). Mem. Proc. Manchester Lit. Phil.
Soc. 57: no. TI.
1914. Some Experiments on the Effect of External Stimuli on the Sporidia of
Puccinia Malvacearum (Mont.). Annals of Botany 28: 330-340.
Sachs, J.
1882. Textbook of Botany. English edition.
Sheldon, J. L.
1902. Preliminary Studies on the Rusts of the Asparagus and Carnation.
Sciencé.n: ser. 16: 235-237.
1905. The Effect of Different Soils on the Development of the Carnation
Rust. Bot. Gaz. 40: 225-229.
Sirrine, F. A.
1900. Spraying for Asparagus Rust. N. Y. Agr. Exp. Sta. (Geneva) Bull. 188.
Smith, R. E.
1904. The Water Relation of Puccinia Asparagi. Bot. Gaz. 38: 19-43.
Stevens, W. C.
1g1r. Plant Anatomy. Philadelphia.
Stone, G. E. and Smith, R. E.
1899. The Asparagus Rust in Massachusetts. Mass. Agr. Exp. Sta. (Hatch)
Bull. 61.
Tieghem, Ph. van.
1873. Recherches Physiologiques sur la Germination. Ann. Sci. Nat. Bot.
Series 5, 17: 205-224.
Tischler, G.
1912. Untersuchungen iiber die Beeinflussung der Euphorbia Cyparissias durch
Uromyces Pisi. Flora 104: 1-64.
Tubeuf, K. F. von.
1897. Diseases of Plants. English edition.
Tulasne, L. R. and Ch.
1847. Memoire sur les Ustilaginées comparées aux Urédinées. Ann. Sci. Nat.
Pl, Bot? 7; 12-127.
Unger, F.
1834. Die Exantheme der Pflanzen und einige mit diesen verwandte Krank-
heiten der Gewaechse etc. Review in Ann. Sci. Nat. II. Bot. 2:
193-215.
Ward, H. M.
1890. Croonian Lecture. On Some Relations between Host and Parasite in
Certain Epidemic Diseases of Plants. Proc. Roy. Soc. London 47:
393-443. .
1902a. On the Question of ‘‘Predisposition’’? and ‘‘Immunity”’ in Plants.
‘ Proc. Cambridge Phil. Soc. 11: 307-328.
1902b. On the Relations between Host and Parasite in the Bromes and their
Brown Rust, Puccinia dispersa (Erikss.). Annals of Botany 16: 233-
315

1902c. Experiments on the Effect of Mineral Starvation on the Parasitism of

220 E. B. MAINS

the Uredine Fungus, Puccinia dispersa, on species of Bromus.
Roy. Soc. London 71: 138-151.
1902d. On Pure Cultures of a Uredine, Puccinia dispersa.
London 69: 451-466.
1903. Further Observations on the Brown Rust of the Bromes, Puccinia
dispersa (Erikss.) and its adaptive parasitism. Ann. Mycol. 1: 132-
I5I.
1904. On the History of Uredo dispersa Erikss., and the ‘‘Mycoplasm”’
Hypothesis. Phil. Trans. London B196: 29-46.
1905. Recent Researches on the Parasitism of Fungi.
I-55.
Wilson, J. K.
1915. Calcium Hypochlorite as a Seed Sterilizer.
427.
Wuist, Miss E. D.

1913. Sex and Development of the Gametophyte of Onoclea struthiopteris,
Physiol. Researches 1: 93-132.

Proc.

Proc... Roy. Soe.

Annals of Botany I9:

Amer. Journ. Bot. 2: 420-

EXPLANATION OF PLATES IV AND V

PLATE -TV

Fic. 1. Development of Puccinia Sorght in carbon dioxide free atmosphere.

a. Plant in carbon dioxide free atmosphere (uninfected). 06. The check (infected).

Fic. 2. Development of Puccinia Sorghi upon plants supplied with carbo-
hydrates. a. Cane sugar I2 percent (infected).
fected). c. Cane sugar 3 percent (infected).
Knop’s mineral nutrient (uninfected).
f. Distilled water (uninfected).

b. Cane sugar 6 percent (in-
d. Cane sugar 3 percent plus
e. Knop’s mineral nutrient (uninfected).

PEATE “V:

Fic. 1. Pure culture of Puccinia Sorghi and its host.

Fic. 2. Pure culture of Puccinia Sorghi upon cut pieces of corn leaf floated
upon carbohydrate solution (looking down upon the capsules containing the leaves).

AMERICAN JOURNAL OF BOTANY VOLUME IV, PLATE IV.

MAINS: RELATION OF RUSTS TO PHYSIOLOGY OF THEIR Hosts

ae 7 - Va re Vy x - + poets, + 4 ” e
‘ = ; : Po oe : Ae, bite yoy BEER Gansticim ae Ines Riko en 5
28 AWS ANRC SIH NN HA FeRAM nn AIR Mo at Lawn AW organ si ctp Nicaea | ALA te a BDNDOSE A SE AEE an VAGIEN, MUR ORAS HORNS 4 eich se Mek RESO 2 ge / Savinvioes F WON EE bs SPE Ls hee AEE a RON AD es we EAN ee 4
= - ree Pa : : ; = c

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE V.

MAINS: RELATION OF RUSTS TO PHYSIOLOGY OP THEIR Hosts

THE DEVELOPMENT OF SOME SPECIES OF AGARICS

A. W. BLizzARD

The species of Agaricaceae whose lamellae are endogenous in their
origin have in recent years been given considerable attention. The
structural variation of a number of forms has been observed and
studied. As a result a number of interesting morphological features
have been explained and their development demonstrated. But
very little attention has been given to those species whose lamellae
are exogenous in their origin. It is important that the morphological
characters of both forms be studied from their origin, beginning with
the young, undifferentiated basidiocarp and tracing their development
to the mature fruit body.

3 Hoffmann was the first to give serious attention to the develop-

ment of Agaricaceae. In 1856 (13) he described Panus torulosus,
showing that the lamellae are exogenous in their origin. | He observed
the hyphae of the young basidiocarp to diverge at the apical end, and
noted the subsequent development of the pileus.

In 1860 (14) and 1861 (15) he followed this work with a description
of a number of forms, the greater portion having an exogenous origin
of the lamellae. He observed the early stage of the palisade layer,
preceding the origin of the lamellae, to be level.

DeBary in 1859 (6) and 1866 (7) described Nyctalis asterophora,
N. parasitica, Collybia dryophila, and others. In the main he agreed
with Hoffmann’s observations, with the exception of the condition of
the palisade layer just previous to the development of the lamellae.
DeBary stated that this layer was folded from the first. He later
(8, 9) agreed with Hoffmann.

Fayod (11) in 1889 described a number of forms. He concludes
that the pileus primordium is endogenous in origin in all Agaricaceae.

Since Fayod’s work there are no published studies on the develop-
ment of exogenous forms. In view of these facts, it appeared to me
that it would be interesting to study the development of certain species
whose lamellae are exogenous in origin.

Mater1al.—Basidiocarps of three species in all stages of develop-

221

222 A. W. BLIZZARD

ment were collected, August and September, 1915, near Seventh Lake,
Adirondacks, N. Y. One species, Clitocybe adirondackensis, was
collected late in September of the same year in Coy Glen, near Ithaca,
IN NG

Omphalia chrysophylla was found growing ona coniferous log which
was not very far along in decay. The season being unusually wet,
quite sufficient moisture was present, which produced a very luxuriant
growth. The cells and structure of this species stand out with unusual
clearness, due, very likely, to the firmness of the cell walls.

Clitocybe adirondackensis was found growing gregariously among
leaves on a steep hillside under coniferous trees. | The whitish myce-
lium growing through and among the decaying leaves, spread over a
space equal to about three square feet.

Clitocybe cerussata was growing in leaf-mold under coniferous trees.
The mycelium was very abundant and spread in all directions.

Clitopilus noveboracensis was growing in leaf-mold in a mixed forest.
Its habit is gregarious. The white mycelium in forms of hyphal
strands permeated the substratum, covering an area equal to about
four feet square.

The basidiocarps were fixed immediately in Carnoy’s fluid, cleared
in cedar oil, embedded and sectioned in paraffine.

OMPHALIA CHRYSOPHYLLA FRIES

Basidiocarp Primordium.—The youngest stages obtained of this
species were small, elongate bodies, averaging 50u in diameter and 780
in length (Fig.1). They are larger at the base and gradually taper toa
blunt point at the apex. At this stage of development they consist of
a homogeneous weft of slender threads, measuring 3.54 in diameter.
Their general direction is parallel with the axis of the young fruit body.
The number of hyphae is increased by branching which takes place
more abundantly toward the base in the young basidiocarp. The
more peripheral hyphae end on the surface at varying distances from
the tip, while central ones converge at the apex (Fig. 18), thus giving
to the young fruit body its slender form.

Stipe Primordium.—The stipe primordium develops by continued
growth of the hyphae that compose the undifferentiated basidiocarp
and is not differentiated as such, until the origin of the pileus primor-
dium; even then there is no definite line separating it from the pileus

lites):

THE DEVELOPMENT OF SOME SPECIES OF AGARICS 223

In Figure 2, a slightly older stage, the hyphae at the base are more
loosely interwoven. This results largely from the increase in the
length of the threads and contributes to the enlarging of the base of
the young basidiocarp. As the plant develops. the hyphal cells in-
crease in size until in the stage represented by Figure 3, they average
4 to 5u in diameter and 10 to 20u in length. The hyphae near the
surface are more closely interwoven.

Pileus Primordium.—In Figure 19 (a higher magnification of a
portion of the apex of Fig. 2) the apex has increased in breadth by a
slight spreading of the hyphae and by interstitial growth of its ele-
ments. There is no differentiation in staining, but the whole structure
has the appearance of very active growth. In a little later stage,
Figure 20, the growth direction of these apical threads is out and up-
ward, with a slight tendency, of the lateral ones, to epinasty. At
the same time profuse branching takes place which supplies new ele-
ments that interlace and ramify among the older ones. Thus, in a
longitudinal section, a weft of closely interwoven hyphae is presented,
except at the periphery where the terete ends of hyphae, advancing
in growth, project (Figs. 20 and 21). This divergent growth of the
hyphae at the apex of the young fruit body marks the origin of the
pileus and differentiates it from the stipe primordium. The origin
of the pileus primordium corresponds very closely to that described
by DeBary in Nyctalis asterophora, N. parasitica (6), and Collybia
dryophila (8, 9).

Further Differentiation of the Pileus.—By continued growth of the
primordial elements the pileus is delimited from the stipe in the fol-
lowing manner: ‘The central hyphae continue to grow toward the
surface, some curving outward more than others. This growth is
accompanied with vigorous branching and interlacing of hyphae,
which add new elements. In this way the inner structure of the pileus
is formed (Figs. 3-6). At the same time the hyphae, by branching
and growing in a radial direction, accompanied by strong epinasty,
curve downward and form the margin of the young pileus (Fig. 22).
By this radial and downward development of the margin of the pileus
an annular groove, Figure 6, is formed on the surface of which is the
hymenophore primordium. Figures 3-6 show the gradual development
from the primordial condition of the pileus to that stage in which it is
well differentiated from the stipe.

The pileus continues to increase in thickness and diameter in a

224 A. W. BLIZZARD

manner similar to that described above and gradually changes to a
broadly convex form with an incurved margin (Figs. 7-10).

Hymenophore Primordium.—Simultaneously with the formation
of the annular groove by epinastic growth of the marginal hyphae,
the hymenophore primordium is differentiated by the rich content
in protoplasm of the hyphae forming the external annular zone in the
furrow. They are crowded, and stained deeply. as shown in longitudi-
nal section (Fig. 6).

The annular region is composed of more or less blunt and cylin-
drical ends of numerous hyphal branches which have their origin both
in the stipe and pileus elements. Their growth direction is obliquely
out and downward. The oldest are on the stipe and by centrifugal
development new elements are added to this area near the margin of
the pileus which continues to curve down over this surface. Figure 22
is a median longitudinal section of the fruit body at this stage of devel-
opment and shows this structure in detail.

At first this annular primordial layer curves out and upward at an
angle of about 45°. As the pileus expands and becomes more convex,
new primordial elements are introduced Ly branching and interstitial
growth in centrifugal succession as above described. This causes it
to curve in the form of an arch (Figs. 7-8).

Since the development is centrifugal it must be borne in mind that,
at the time this layer is in the primordial stage at the margin of the’
pileus, near the stipe it will be further differentiated.

Palisade Layer; Origin and Development of the Lamellae.—The
hyphae of the hymenophore primordium branch in a digitate manner.
By this branching new elements are interpolated in the spaces between
the older hyphae. This process continues gradually until a compact
layer of short hyphae is formed. Simultaneously with this the cells
enlarge, especially the terminal ones, and the surface smooths up into
an even, compact layer (Fig. 23). This is the palisade layer and pre-
cedes the origin and development of the lamellae, as has been described
for a number of endogenous forms. ‘The differentiation of the palisade
layer appears first near the stipe and progresses centrifugally toward
the margin of the pileus, as did the development of the hymenophore
primordium. As the cellular elements of the palisade layer increase
in size, a great pressure is produced within this structure. This
pressure is released to some extent by the palisade layer being thrown
into equally spaced, radial folds beginning near the stipe (Fig. 12).

THE DEVELOPMENT OF SOME SPECIES OF AGARICS 225

Simultaneously with this, subadjacent hyphae along radial areas
corresponding with the gill areas, by elongation, push their way down-
ward and govern the origin of the gill salients. Figures 24 and 25
show this feature in excellent detail. These down-growing salients
of the level palisade layer are the first evidence of the appearance of the
lamellae. Continued growth of these salients produces the lamellae,
as observed in species of Agaricus (1, 2, 4), Coprinus (5), Cortinarius
(10) ,° etc.

The subhymenial hyphae are branched in a corymbose manner,
Figure 27, and supply new elements within the hymenophore layer.
Figure 28 is a higher magnification of a portion of Figure 27 and shows
in detail the corymbose branching.

Growth in width of the lamellae occurs by the further elongation
of the tramal hyphae which branch as above described. These new
elements are interpolated between the older hyphae at the edge of the
gill.

The development of the lamellae is centrifugal as are the structures
preceding their origin. Therefore the oldest portions of the lamellae
are nearest the stipe and proceed in a radial direction to the margin.
Since the margin of the pileus is involute, a tangential section through
that portion of the fruit body parallel with the axis of the stipe will
present an appearance as represented by Figure 16. Below, it shows
a portion of the inrolled pileus edge. This relation of the gills to the
involute margin of the pileus has been adequately described by Atkin-
son for Agaricus rodmant (4).

Structure of Stipe and Pileus.—As the basidiocarp grows the stipe
becomes even in diameter. This results largely from the elongation
of the peripheral hyphae and more abundant branching in the upper
portion ‘of the stipe, together with the enlurgement of the cellular
elements. The process is a gradual one, as Figures 3—9 show.

The pileus at the same time increases in all its dimensions and
becomes more expanded (Fig. 10). The general direction of its ele-
ments is horizontal and radial (Fig. 29). Those on the surface are
more closely interwoven, and this serves to produce a smooth surface.
Figure 30 is a high magnification of a portion of the pileus which shows
this structure very clearly. The hyphal cells have very firm walls
and are exceedingly large, measuring 7 to 8u in diameter.

226 A. W. BLIZZARD

CLITOCYBE ADIRONDACKENSIS PECK

Basidiocarp Primordium.—The undifferentiated basidiocarps of
this species are long, slender bodies, tapering toward the apex. They
are usually curved or bent in various directions. ‘Those studied
measure 60y in diameter and I mm. to 2 mm. in length. The hyphae
are very siender, wavy threads, averaging 1.54 in diameter. They
run in a longitudinal direction (Figs. 31 and 41). The central hyphae
extend to the apex where they converge into a blunt point. The periph-
eral ones end in such a manner as to form a slanting surface from
the base to the apex. There is no differentiation at this time,. the
whole fruit body staining homogeneously.

Pileus Primordium.—When the young primordium of the fruit
body reaches a diameter of about 300 the lateral threads at the apical
end begin to spread laterally (Figs. 32 and 42). The central hyphae
continue their growth upward. The interhyphal spaces are filled by
new elements which are produced by branching of the older hyphae.
‘This structure is the pileus primordium, and, as in Omphalia chryso-
phylla, the divergence of the threads indicates the differentiation of
the pileus fundament at the apex. . :

Further differentiation of the pileus is the result of continued
growth of this primordial tissue. Around the upper lateral surface
of the stipe primordium and on the under side of the young pileus,
the ends of the diverging hyphae stain deeply and mark the origin of
the hymenophore fundament (Fig. 33). The central apical hyphae
continue their growth upward and by profuse branching add materially
to the thickness of the pileus, while the intermediate elements bend
gently outward.

By continued branching and interstitial growth of its elements
the pileus increases in diameter. At the same time the central hyphae,
as compared with those of the periphery, elongate less rapidly. Thus
the intermediate and peripheral threads, growing upward and outward
at an oblique angle of about 45°, cause the pileus to become plane on
its upper surface (Fig. 35). The marginal hyphae at the same time
continue to curve abruptly downward. In this way a shallow and
very narrow annular groove is formed. ;

In later stages in the under portion of the pileus next the stem,
hyponasty replaces epinasty. The form of the pileus consequently
changes from plane to umbilicate, and then to infundibuliform, while

THE DEVELOPMENT OF SOME SPECIES OF AGARICS 227

epinasty continues to have its influence on the thin younger margin
which is incurved or involute.

Structure of the Sttpe.—-At the appearance of the pileus fundament,
the stipe primordium is clearly differentiated as a definite region.
At this stage the apparently homogeneous structure of the stipe pri-
mordium is changed by the loosening up of its texture. There are
strands of tissue running longitudinally through the stipe. These
strands stain very deeply, causing them to stand out conspicuously.
They intertwine with others in anastomosing fashion, thus forming
intervening hyphal spaces.

The stipe elongates by the lengthening of the cells. These ele-
ments, in the stage of development represented by Figure 40, measure
3-30u in length. The increase in width is the result of branching and
interstitial growth of the hyphae, and also by the increase in diameter
of the cells themselves, which average 3.5 to 4u. In more mature
plants increased thickness is chiefly the result of the latter, as Hoffmann
(14) on page 394 suggested.

Hymenophore Primordium.—The organization of the hymenophore
primordium occurs simultaneously with that of the origin of the pileus
margin. Like it, too, the development is centrifugal. The first
differentiation of this tissue is in the angle between the pileus margin
and stipe and on the upper surface of the stem. Because of the active
increase in its elements and their richness in protoplasm the young
hymenophore primordium takes a dense stain.

As the pileus increases in width, its marginal hyphae add to this
annular zone so that its surface is increased radially and upward
(Fig. 43). Its elements multiply by intercalary growth and present a
frazzled appearance, as observed in Coprinus comatus (5), Agaricus
rodmant (4), and some other plants.

Palisade Layer, Origin and Development of the Lamellae.—By con-
tinued branching of the hyphae, the zone of primordial elements
organize a definite layer of parallel threads which becomes more or
less even on the surface since the ends of the hyphae reach the same
level. This results in forming a compact layer of parallel threads
perpendicular to the surface. Figure 44, is a transection through the
upper part of the stipe and shows a portion of this structure immedi-
ately beneath the curved pileus margin. The hyphal elements of
this layer are slender, cylindrical, septate threads, 4.5—6u in diameter
and 35u in length. The terminal cell is longer than the others of the

228 A. W. BLIZZARD

same thread and slightly larger, which tends to give a clavate appear-
ance to the threads. The cells are rich in protoplasm and present an
appearance of active growth.

Hymenophore Primordium.—The lamellae make their first appear-
ance as folds of the level palisade layer. These folds are the rudiments
of the lamellae themselves. They appear on the surface at or near the
apex of the stipe, Figure 36, and by progressive growth extend out
and upward on what is the morphological underside of the pileus. By
downward growth of hyphae subadjacent to these folds, the trama of
the lamellae is formed (Figs. 45-46). These tramal threads are dif-
ferentiated from the other elements of the hymenophore by the fact
that they do not stain so deeply. These threads branch and furnish
new elements by which the lamellae grow in thickness and at the same
time by apical and intercalary growth the lamellae increase in width
(Fig..47).

The tissue of the pileus and stipe subadjacent to the hymenophore
is peculiar because of extraordinary large interhyphal spaces, due to
the extension exerted by the pressure from interstitial growth and
enlargement of the elements of the hymenophore.

The lamellae develop in length in a radial centrifugal direction,
following that of the palisade layer. They are decurrent from the
beginning, since the hymenophore has formed around the upper lateral
surface of the stipe (Figs. 36-40). At the base of the older ones,
other lamellae sometimes branch off, developing in a manner described
for the primary gills (Fig. 37). These form the forked lamellae some-
times present in this species. Secondary lamellae also arise between
the diverging primary gills, filling the spaces between them.

CLITOCYBE CERUSSATA FRIFS

Basidiocarp and Stipe Primordia.—The youngest basidiocarps of
this species which were collected measure .5 mm. in diameter and
2 mm. in length (Fig. 70). They are composed of slender inteilacing
hyphae, measuring 3u in diameter, which form a close interwoven
tissue. Their general direction is longitudinal, converging at the
apex (Fig. 83). This homogeneous structure is the primordium of
basidiocarp and stipe.

By continued growth of this primordial tissue the stipe fundament
is finally differentiated by the formation of the pileus primordium which
is marked off by the divergence of the apical hyphae. As the stipe

THE DEVELOPMENT OF SOME SPECIES OF AGARICS 229

becomes older the hyphae are more loosely interwoven (Fig. 7). Its
further growth is provided for by means of branching and elongation
of its elements. |

Pileus Primordium.—At the time the stipe fundament is delimited
from that of the pileus, the apical hyphae grow upward and spread
out in all directions. In this feature it is similar to that of Chitopilus
noveboracensis, which is described below. The hyphal elements are
long, slender and terete. This is the pileus primordium (Fig. 84).

By continued radial and diverging growth of its elements the pileus
fundament increases in size (Figs. 72, 77). This gives rise to a hemi-
spherical body which is delimited from the stipe by the annular groove
(Figs. 72-74). Further epinastic growth causes the margin to curve
inward toward the stipe (Fig. 76). At this time the plant has assumed
a beautiful and symmetrical form. In Figure 77, the pileus has en-
larged and the margin has become so strongly involute that the edge
turns upward against the gills. The hyphae do not grow out from
the margin of the pileus nearly so strongly as in Clitopilus novebora-
censts.

Hymenophore Primordium.—The hymenophore fundament is
differentiated in the annular groove between the pileus and stipe and
stains deeply. This area develops in a radial manner, following the
centrifugal growth of the pileus, characteristic of the Agaricaceae.
- This area consists of short hyphae perpendicular to the surface of the
annular groove. It becomes more dense by interpolation of new ele-
ments which are formed by digitate branching of the primordial
hyphae.

Palisade Layer; Origin and Development of Primary Lamellae.—As
the hymenophore becomes more compact by intercalary growth, the
cells themselves increase in size. The end of the hyphae reach the
same level and form an even palisade layer, as shown in Figure 78.
A higher magnification is shown in Figure 85. The hyphal elements
that compose this layer are longer than those of the other species
described in this paper, and are comparatively slender.

As the elements increase in number and size, the resulting pressure
is partly relieved by the level palisade layer bulging out into radial
fold-like ridges. These ridges are the gill fundaments. In this
species, as in Clitopilus noveboracensis, they occur first on the stipe
very near the angle between the latter and the lower surface of the
pileus (Fig. 79). Later the gill salients of the primary lamellae appear

230 A. W. BLIZZARD

on the under surface of the pileus as shown in Figure 80. In this
figure, on each side of the salients, a portion of the palisade layer is
shown. Since these structures develop centrifugally the first differen-
tiation occurs on or near the stipe. Consequently in a tangential
section the portion to the right or left would be cut obliquely and show
tissue nearer the margin than that in the center of the section. Thus,
the palisade layer represents a younger portion of the hymenophore,
in which salients have not as yet made their appearance.

The development of the lamellae in width is as has been described
for the previous species. The subadjacent hyphae by elongation,
aid in the extension of the salients in width or keep pace with their
growth. New elements are also added by intercalary growth to the
palisade layer. Figures 85-89 show in detail the development of a
gill from the palisade stage of the hymenophore through the first
evidence of a gill salient to a well-formed lamella.

Origin of Secondary Lamellae.—As was described for the previous
species, the salients of the secondary gills appear between the primary
lamellae on the under surface of the pileus. Those that appear first
occur near the stem (Fig. 81). Their development is exactly as de-
scribed for the primary gills. They serve to occupy the spaces
produced by the divergence of the primary gills as they proceed from
the stipe.

Structure of Pileus and Stipe.—The more mature pileus is expanded
and the hyphae arrange themselves in a radial horizontal direction.
The trama is composed of hyphal threads that ramify and interlace
among themselves. The stipe increases in width by branching and
interstitial growth of the hyphae. In the more mature pileus and
stem, growth is chiefly by the increase in size of the cellular elements.

CLITOPILUS NOVEBORACENSIS PECK

Basidiocarp and Stipe Primordia.—The fruit bodies representing
the primordial stage of the basidiocarp become comparatively large,
.6 mm. in width and 2 mm. in length, before differentiation of the
pileus occurs. They are elongate bodies which taper gradually to a
point at the apex (Fig. 48). The young basidiocarp presents a closely
interwoven structure composed of slender hyphae averaging about 3u
in diameter at the base; toward the apex they are not so stout. The
general direction of the hyphae is parallel with the direction of the
growth of the fruit body (Fig. 64). The whole extent of the apical

THE DEVELOPMENT OF SOME SPECIES OF AGARICS 231

end and a portion of the peripheral hyphae stain deeply, which indi-
cates an area of active growth.

This fundamental tissue is soon differentiated by growth direction
of the apical] threads as stipe primordium. At this time the inner
structure of the stipe fundament is a woof of slender homogeneous
hyphae, while some of the hyphal threads near the surface of the stipe,
growing more rapidly than the other elements, extend outward and
form a loose floccose layer. Figure 49, a median longitudinal section,
shows this layer as a narrow zone which stains more deeply. This
structure is composed of the dead ends of those hyphae which extend
beyond the immediate surface of the stipe and is very ephemeral.

Pileus Primordium.—The origin and development of the pileus
fundament agrees with the preceding species. The elements extend
outward in all directions with a slight tendency to epinasty. In this
species the fundament consists of a peripheral zone of long radiating
hyphae and a dark staining central portion. The hyphae of both
the loose and the more dense regions have the same origin; 7. e., the
elements of both regions are the result of radial diverging growth from
the stipe fundament.

Further development of the pileus is by the continued radial
growth of the hyphae. At the margin by epinastic growth the hyphae
curve downward, forming the annular groove (Figs. 50, 51). In the
stages represented by Figures 52-54, the hyphae branch profusely
and are organized in a very compact structure, except a very thin,
loose surface zone. ‘The pileus margin develops by centrifugal growth.
On the surface of the annular groove in the angle between the pileus
and stipe the hymenophore fundament is organized.

In a later stage of development, represented by Figure 55, the
pileus margin is so strongly involute that the edge is curved upward
against the lamellae. The marginal hyphae span the intervening
space between the pileus margin and the gills. At this stage of growth
these hyphae function as a marginal veil, though this veil is very
different in origin from the marginal veil of those species with endog-
enous origin of the hymenophore. They do not, at any stage of
development which I have examined, interlace with the hyphae of the
stipe, as Hartig suggested for Armillaria mellea (12). Such an inter-
lacing might occur in case the margin curved down against the stipe
below the hymenophore area. In all the specimens examined, how-
ever, the pileus margin curves up toward the hymenophore. Hypo-

242 A. W. BLIZZARD

nastic growth of the older portion of the pileus begins soon, causing it
to expand, thus lifting the margin up and outward far away from con-
tact with the hymenophore. The growth of the pileus in thickness is
primarily the result of the enlargement of its elements accompanied
by branching and intertwining of the threads.

Hymenophore Primordium.—Soon after the origin of the pileus
primordium, the ends of the peripheral hyphae which are perpendicular
to the surface of the annular groove, are rich in protoplasm and stain
deeply. This is the region of the hymenophore primordium. This
region, as in Omphalia chrysophylla and Clitocybe adirondackensts,
develops centrifugally and adds new elements by intercalary growth.

Palisade Layer.—By continued introduction of new elements,
this layer becomes compact and the free ends reach the same level
(Fig. 66). The increase in size of the cellular elements and the ex:
tending downward of the subadjacent hyphae produce regularly
spaced, radial folds in the palisade layer (Figs. 57, 58). These folds
are the salients of the primary lamellae and appear first on the stipe
very near the angle between the latter and the under surface of the
pileus. Thus, the gills are decurrent from their very first appearance.

At this period of development the hymenophore layer on the under
surface of the pileus is in the level palisade stage, near the stipe (Fig.
56). It gradually grades off into the primordial condition at the mar-
gin. Therefore, since the gills follow the same centrifugal succession
as did the structures preceding their origin, the salients continue their
development toward the margin of the pileus. Thus, Figure 58 (a
little later stage than Fig. 57) shows their first appearance on the
under surface of the pileus.

Further growth in width of the lamellae is brought about by
growth of the tramal hyphae in these folds. This growth aids in
pushing the palisade layer outward at the edge of the salient. By
corymbose branching new elements are introduced into the palisade
layer by intercalary growth. The hyphae that grow down into the
lamellae from the trama of the pileus form the trama of the gills.
Figures 67—69 show a serial development of a gill from the origin of
the salient to a lamella fairly well along in growth.

Origin and Development of Secondary Lamellae.—As the primary
gills advance from the stipe to the margin of the pileus, they diverge
from each other. In the spaces so produced on the under surface of
the pileus near the stipe, the secondary lamellae arise. Figure 59, a

THE DEVELOPMENT OF SOME SPECIES OF AGARICS wh, gn

transverse section through the stipe and pileus, shows the primary
lamellae as ‘“‘bars’’ extending between the pileus and the stipe. Be-
tween the ‘“‘bars”’ on the morphological under surface of the pileus,
down-growing salients of the secondary gills are shown. They develop
and progress radially, as do the primary lamellae. Figure 61 is a
slightly oblique transection through the margin of the pileus and upper
portion of the stipe and shows the increase in number of lamellae on
the pileus margin as compared to the number of primary gills on the
stipe.

Further Growth of Pileus and Stipe.—The pileus elements in the
more mature stage have in general a radial, horizontal direction.
The trama of the pileus increases by branching and elongation of its
elements. The size of the stipe increases likewise by branching and
interstitial growth. The lengthening or elongation of the stem, as in
the previous species studied, is the result of the extension in length.
of the cellular elements.

SUMMARY

1. The young basidiocarp and stipe primordium consist of a homo-
geneous weft of slender, terete, interlacing hyphae. The general
growth direction of the elements is parallel with the axis of the young
fruit body. The hyphae converge at the apex. The cellular elements
are comparatively short, cylindrical cells, rich in protoplasmic content.

2. The pileus primordium is differentiated by the divergence of
the apical hyphae which grow upward and laterally. This divergence
serves to mark the origin of the pileus and differentiates it from the
stipe fundament.

Further differentiation of the pileus is the result of continued
growth of the primordial tissue. By profuse branching and intersti-
tial growth of the elements, an intricately interwoven tissue is pro-
duced. At the same time the lateral hyphae by epinastic growth
bend downward, forming the annular groove.

3. The primordium of the hymenophore is organized simultane-
ously with the origin of the pileus margin. The first differentiation
of this tissue is on the surface of the annular groove in the angle be-
tween the pileus and stipe. This annular layer progresses centrif-
ugally.

4. By continual branching of the hyphae, the hymenophore primor-
dium changes to a definite layer of parallel threads, perpendicular to

234 A. W. BLIZZARD

the surface. By the enlargement of the cells of these parallel hyphae,
and the evening up of the hyphal elements, a level palisade layer is
produced.

5. The primary lamellae originate as evenly spaced, radial folds
of the level palisade layer. The first folds that appear are the rudi-
ments of the primary gills. Their further development is produced
by the elongation of the subadjacent hyphae of the pileus which push
their way into the salients and form the trama of the gills. These
tramal hyphae branch and furnish new elements by which the lamellae
grow in thickness and at the same time by apical and intercalary
growth increase in width.

6. The secondary lamellae arise as down growing salients of the
palisade layer on the under surface of the pileus near the stipe between
the primary gills. They develop as do the primary lamellae.

In conclusion I wish to express sincere thanks to Professor Geo.
F. Atkinson under whose direction this study was undertaken for
many helpful suggestions.

DEPARTMENT OF BOTANY, COLLEGE OF ARTS AND SCIENCE,
CORNELL UNIVERSITY

LITERATURE CITED

1. Atkinson, Geo. F. The Development of Agaricus campestris. Bot. Gaz. 42:
241-264, pls. 7-12. 1906.

2. ——. The Development of Agaricus arvensis and A. comtulus. Amer. Journ.
Bot; 1s 3522) His:.7,°2.. 1914:
3. ——. The Development of Amanitopsis vaginata. Ann. Mycol. 12: 269-392,

pls. 17-19. I914.

4. Morphology and Development of Agaricus rodmani. Proc. Amer. Phil.
Soc. 54: 309-343, 8 diagrams, pls. 7-13. 1915.
5. Origin and Development of the Lamellae in Coprinus. Bot. Gaz. 61:

89-130, 8 diagrams, pls. 5-12. I916.
6. DeBary, A. Zur Kenntniss einiger Agaricinen. Bot. Zeit. 17: 385-388, 393-
398, 401-404. pl. 13. 1859.

Te Morphologie und Physiologie der Pilze. Flechten. und Myxomyceten.
Leipzig, 1866.

8. Vergleichende Morphologie und Biologie der Pilze, Mycetozoen, und
Bacterien. 1884.

9. Comparative Morphology and Biology of the Fungi, Mycetozoa, and

Bacteria. Oxford, 1887.

10. Douglas, Gertrude E. A Study of Development in the Genus Cortinarius.
Amer. Journ. Bot. 3: 319-335, pls. 5-13. 1916.

11. Fayod, V. Prodrome d’une histoire naturelle des Agaricinees. Ann. Sci. Nat.
VIII. Bot. 9: 181-411, pls. 6, 7. 18809. ;

THE DEVELOPMENT OF SOME SPECIES OF AGARICS 235

12. Hartig, R. Wichtige Krankheiten der Waldbaume, usw. 12-42, pls. 1, 2. 1874.
13. Hoffmann, H. Die Pollinarien und Spermatien von Agaricus. Bot. Zeit. 14:

137-148, 153-163, pl. 5. 1856.

14. Beitrage zur Entwickelungsgeschichte und Anatomie der Agaricinen.
Bot. Zeit. 18: 389-395, 397-404. pls. 13, 14. 1860.
15. Icones Analyticae Fungorum, Abbildungen und Beschreibungen von

Pilzen mit besonderer Riicksicht auf Anatomie und Entwickelungsgeschichte.
1861. ;

DESCRIPTION OF PLATES VI-XI

The following photomicrographs were made by the author with a Bausch and
Lomb vertical camera and Zeiss lenses, and with a horizontal Zeiss camera.

Prare Vil

Fics. 1-17. Omphalia chrysophylla.

Fics. 1-10. Median longitudinal sections, showing the general habit and
development of the basidiocarp. X 24.

Fic. 1. Basidiocarp primordium composed of a homogeneous weft of slender
threads. The general direction of the hyphae is parallel with the axis of the young
fruit body. X 24.

Fic. 2. The first differentiation of the basidiocarp is the flaring out of the
hyphae at the apical end which is the region of the pileus primordium. The portion
of the basidiocarp below this differentiation is the stipe primordium. X 24.

Fics. 3-5. Sections showing the gradual development of the pileus. The
hyphae grow up and outward with a slight tendency to epinasty. They also show
the gradual growth that takes place in the stipe which causes it to become even.
2a.

Fic. 6. Shows the further development of the pileus and stipe. The former
develops by a radial centrifugal growth. The marginal hyphae curve down and
form an annular groove on whose surface is the hymenophore primordium. The
latter structure is differentiated as a more dense staining area in the angle between
the margin of the pileus and stipe. X 24.

Fic. 7. Represents a further development of these structures. The hymeno-
phore primordium has differentiated into the palisade layer. X 24.

Fics. 8-10. These sections show the continued development of the basidio-
carp. The palisade layer has been replaced by the lamellae. The stipe and pileus
has developed by means of branching of its elements and interstitial growth. X 24.

Fic. 11. A tangential section near the stipe, showing the palisadelayer. X 24.

Fic. 12. A tangential section very near the stipe which shows the down-
growing salients of the palisade layer, the rudiments of the lamellae. > 24.

Fics. 13-17. Tangential sections, showing the development of the lamellae
from the first appearance of the salients to the well-formed lamellae. X 24.

Fic. 16. Isa tangential section through the margin of the pileus, showing the
involute edge beneath the gills. X 24.

236 A. W. BLIZZARD

PLATE VII

Fics. 18-30. Ompbhalia chrysophylla.

Fic. 18. A higher magnification of a portion of the apex of Fig. 1. It shows
the converging of the apical hyphae and homogeneous nature of the whole struc-
ture. X. 100;

Fic. 19. A higher magnification of a portion of the apex of Fig. 2. The
threads are spreading apart slightly and have increased in size. This differentiation
marks the region of the pileus and stipe primordia. X 500.

Fics. 20, 21. A higher magnification of a portion of the apices of Figs. 3-a,
showing the further growth of the primordial hyphae of the pileus. They grow
radially outward and by epinasty curve downward. X 300.

Fic. 22. A higher magnification of the pileus margin of Fig. 8, showing the
annular groove on the surface of which is the hymenophore primordium. This
primordium is composed of the ends of hyphae whose origin is in the pileus and stipe
elements. They are rich in protoplasm and stain deeply. The hyphae of the
pileus margin by strong epinasty curve down by which the annular groove is formed.
Ka230:

Fic. 23. A tangential section of the pileus near the stipe which shows in
detail the structure of the palisade layer. This layer is formed by branching and
. interstitial growth of the primordial hyphae. As the cells themselves increase in
size the layer becomes compact and even. XX 300.

Fic. 24. A tangential section of the pileus showing the beginning of a gill
salient. The pressure within the layer is relieved by this downward folding of the
level palisade layer. At the same time subadjacent hyphae by elongating push
down into this fold forming the trama of the gills. X 300.

Fic. 25. <A tangential section showing a salient a little further developed. In
this section the detail of the structure stands out so definitely that there can be
no possible mistaking as to how the gill salient proceeds in developing. The tramal
hyphae can be easily traced from the pileus elements above down into the palisade
layer itself. In this way new elements are introduced in the periphery of the salient
and also the trama of the gill is produced. XX 300.

Fics. 26, 27. Are tangential sections showing further development of the gill
salients. The tramal hyphae are evident and the corymbose branching, by which
new elements are added to the palisade layer of the gill, is clearly shown. X 300.

Fic. 28. A higher magnification of a portion of the palisade layer of Fig. 27.
The corymbose branching of the tramal hyphae and the intercalary growth of the
elements are well shown. XX 720.

Fic. 29. A median longitudinal section which shows the structure of the pileus
and its relation to the palisade layer between the gills. The hyphae branching in a
corymbose manner supply elements to the palisade layer. XX 230.

Fic. 30. A high magnification of the edge of the pileus surface of Fig. Io.
The very large and stout hyphae are well shown. The hyphae on the right side of
the figure are interwoven and serve to produce a smooth surface. XX 720.

THE DEVELOPMENT OF SOME SPECIES OF AGARICS 237

Prare VIII

Fics. 31-47. Clitocybe adirondackensis.

Fic. 31. A median longitudinal section of a young basidiocarp. It repre-
sents the primordial stage at which time it is composed of a loose weft of wavy,
slender, and homogeneous hyphae. X 32.

Fic. 32. A median longitudinal section which shows the pileus primordium
at the apical end, differentiated from the stipe primordium. The apical threads are
spreading and serve as an arbitrary line of demarcation between the areas of the
two primordia. X 32.

Fic. 33. A median longitudinal section showing further development of the
pileus and stipe. By continued growth of the central primordial hyphae umbonate
pileus is produced. Between the margin of the pileus and stipe an area of densely
staining hyphae is shown. This is the hymenophore primordium which develops
centrifugally as the pileus continues to grow. X 32.

Fics. 34,35. A median longitudinal section of later stages. The pileus has
increased in thickness but the expansion is comparatively little. Since the mar-
ginal hyphae elongate more rapidly than the central ones the pileus becomes plane.
The hymenophore primordium develops at the same time, advancing toward the
margin of the pileus and stains more deeply. X 32.

Fic. 36. An oblique transection through the upper portion of the stipe. The
lower part of the figure shows a portion of the surface of the stipe beneath the
hymenophore. To the left and above this region is a portion of the palisade layer.
The remaining peripheral portion shows the folding of the palisade layer and the
development of the gill salients. X 32.

Fic. 37. Cross-section of the extreme lower portion of the pileus, show-
ing the general habit of the gills and manner of development. XX 32.

Fic. 38. A tangential section through the pileus and near the stipe, show-
ing the decurrency of the lamellae and also the incurving of the pileus margin.
Worse:

Fic. 39. A tangential section midway between the: margin of the pileus and
stipe. It shows the thickness of the pileus and the nature and general direction
of the gills. Some are connected at their base which is chiefly the result of branching.
Be.

Fic. 40. A median longitudinal section, showing the general habit of the
plant. The pileus is plane and the gills extremely decurrent.

Fic. 41. A high magnification of a young basidiocarp, showing in longi-
section the structure of the primordial condition. A homogeneous weft of wavy,
slender threads which converge at the apex. X 230.

Fic. 42. A median longitudinal section, showing the apical hyphae growing
outward. This is the pileus primordium. X 300.

Fic. 43. A median longitudinal section, showing the hymenophore pri-
mordium. This is composed of the ends of hyphae which have their origin in the
stipe and pileus elements. The margin of the pileus curves down over the surface.
Thus, the marginal hyphae add new elements to the hymenophore regions, both of
which develop centrifugally. X 300.

Fic. 44. A cross-section through the basidiocarp near the apex of the stem

238 A. W. BLIZZARD

immediately below the margin of the pileus. This shows the palisade layer which is
formed by the gradual increase of primordial elements by intercalary growth. At
the same time the cellular elements, which result in a compact layer. > 300.

Fic. 45. A cross-section similar to the above only of a slightly older stage.
It shows the first appearance of gill salients, which are the outfolding of the palisade
layer. Subadjacent hyphae grow into this fold, and by elongation force the salient
outward, at the same time branching in a corymbose manner, new elements are added.
xX 300.

Fic. 46. A cross-section of the pileus which shows further growth of the gill
salients. The hyphae that force their way down into the salients from the pileus
elements do not stain so deeply and are easily distinguished. They form the
trama of the lamellae. XX 300.

Fic. 47. A cross-section of a little older stage than the preceding figure.
This shows the apical development of the gill by which the lamellae increase in
thickness.

PLATE IX.

Fics. 48-63. Clitopilus noveboracensts.

Fic. 48. A median longitudinal section, showing the homogeneous weft of
slender, interwoven hyphae. The peripheral hyphae end at varying distances from
the tip, so that the surface slants gradually from the base to the apex. X 20.

Fic. 49. A median longitudinal section, showing the flaring of the hyphae
at the apex which serves as a line of demarcation between the pileus and stipe
primordia. On the surface of the stipe primordium is a very narrow zone of tangled
hyphae which stain more deeply. This is composed of the ends of hyphae which
project farther than those that compose the weft. XX 20.

Fics. 50-54. Median longitudinal sections, showing older stages of develop-
ment. The marginal hyphae by epinastic growth turn downward, forming the
annular groove on whose surface is the hymenophore primordium.

Fic. 55. A median longitudinal section of a more mature plant, showing its
general habit. The margin turns in and upward towards the gills. The marginal
hyphae extend outward as a loose weft and span the space between the pileus margin
and gills. At this stage it has the function of a marginal veil. XX 13.

Fic. 56. A tangential section through the pileus near the stipe, showing the
palisade layer. X 20.

Fic. 57. An oblique transection through the margin of the pileus and upper
part of the stipe. Thecavity within represents the annular groove. On the surface
of the stipe the palisade layer has been thrown into folds. These folds are the gill
salients. Thus the origin of the primary gills is on the stipe. X 20.

Fic. 58. A tangential section of a young pileus, showing the origin of the
primary gills as they extend from the stipe on the lower surface of the pileus toward
the pileus margin. XX 20.

Fic. 59. A slightly oblique transverse section through the pileus margin
and upper part of the stipe, showing the origin of the secondary lamellae between
the primary gills. The primary gills appear as ‘‘bars,’’ connecting the pileus and
stipe. X 20.

Fic. 60. A tangential section through the pileus, showing the decurrency of
the gills; also the incurving of the pileus margin. XX 20.

THE DEVELOPMENT OF SOME SPECIES OF AGARICS 239

Fic. 61. A slightly oblique cross-section through the pileus margin and stipe,
showing: (1) primary gills on the stipe; (2) the upper left hand portion of the
figure shows two ‘‘bars”’ of the primary gills; (3) a secondary gill between the
““Dbars’’; (4) primary and secondary gills on the pileus margin. On the right lower
hand portion of this figure is shown a part of the involuted pileus margin. XX 20.

Fic. 62. <A tangential section through the involuted margin of the pileus,
showing the relation of the involuted margin to the gills. X 20.

Fic. 63. A tangential section through the margin of the pileus just within the
limit of the involuted margin. It shows very clearly how the marginal hyphae have
been pushed against the gills. x 20.

PLATE X

Fics. 64-69. Clitopilus noveboracensis.

Fic. 64. A higher magnification of the apex of Fig. 48, showing the converg-
ing of the hyphae. X 200.

Fic. 65. <A higher magnification of a portion of the pileus primordium, show-
ing the spreading of the hyphae at the apex. X 200.

Fic. 66. <A higher magnification of the palisade layer. Its elements are in-
creased by intercalary growth. This, together with the increase in size of the
hyphae, produces the compact palisade. XX 300.

Fic. 67. A cross-section through the stipe, showing the first fold in the palisade
layer. It isan outward growing salient, the rudiment of a lamella. XX 300.

Fics. 68, 69. Transverse sections through the pileus margin and stipe, showing
further development of the gill salient. The apical growth of the gill is well shown
by which it increases in width. The tramal hyphae can be definitely made out.
xX 300.

Fics. 70-74. Clitocybe cerussata.

Fic. 70. A median longitudinal section of a young basidiocarp, showing the
basidiocarp primordium. It is a homogeneous weft of slender, interlacing hyphae,
whose general direction is longitudinal and converge at the apex. X 20.

Fic. 71. A median longitudinal section, showing the diverging hyphae at the
apex. This serves to separate the pileus and stipe primordia. X 20.

Fics. 72-74, Median longitudinal sections of older stages, showing the centri-
fugal development of the pileus. The stipe at the same time increases in thick-
ness. The hymenophore is differentiated in the surface of the annular groove.
X 20.

Fics. 75, 76. Median longitudinal sections, showing further growth of the
pileus. The margin curves down and toward the stipe. X 20.

Fic. 76. A median longitudinal section of a more mature plant, showing its
general habit. The strongly incurved margin is well shown. XX 20.

PLATE XI

Fics. 78-89. Clitocybe cerussata.

Fic. 78. A tangential section of the pileus, showing the palisade layer near the
margin, “X<.20.

Fic. 79. A slightly oblique transection through the pileus margin and stipe,
showing the origin of the lamellae as folds in the palisade layer on the upper por-
tion of the stipe. _ X 20.

240 A. W. BLIZZARD

Fic. 80. A tangential section through the pileus, showing the origin of the
primary gills on its under surface. On either side of the down-growing salients a
portion of the palisade layer is shown. X 20.

Fic. 81. Transection of the pileus and stipe, showing the origin of the sec-
ondary lamellae between the primary gills which show as “‘bars.’”’ X 20.

Fic. 82. Tangential section of the pileus, showing further development of
the gill salients. The tissue below the salients is a portion of the involuted pileus
marcin.§ G20,

Fic. 83. A higher magnification of the apex shown in Fig. 20. This shows
the converging of the apical threads and the general homogeneous structure.
X 100.

Fic. 84. A longitudinal section of the pileus primordium, showing the general
growth direction of the hyphae. X 300.

Fic. 85. Transection of the stipe, showing the palisade layer. XX 300.

Fic. 86. A transverse section of the stipe, showing the first evidence of a
fold in the palisade layer. X 300.

Fic. 87. Further development of a salient represented by Fig. 86. It
shows the apical growth by which the elements of the palisade layer are increased
by elongation of the hyphae from the trama of the pileus. XX 300. ,

Fic. 88. Represents a little older stage than the preceding one. The tramal
hyphae are shown growing down into the gill and branching; thus new elements are
supplied to the palisade layer. X 300.

Fic. 89. A slightly older stage, showing further development of the gill.
X 300.

AMERICAN JOURNAL OF BOTANY. VOLUME IV PLATE VI.

BLIZZARD : OMPHALIA CHRYSOPHYLLA

~ ; i
| "hy , a { :
* : ;
4 7 ;
n
i * i ;
ae | } a ‘
ae 1 :
iNe
- Zi we)
3 i
: i H
: : a . ° -
‘
‘i
~~ . ,
é fi , , i
; \
“ : .
‘
=. . al. a!
.
; t
f ‘
; :
. |
i
: F
: ‘ ‘ ;
a . . oy { fi
j \
' J : ’
: :
pf '
: .
- :
. 7 x : ;
- ©
. ~ P
ra 1 :
aa ~— ;
15 - p<
. 2 = :
‘ B
war
es 2 5
\ ‘
i
: ,
. ‘2 7 = , :
- : i.
j A
’ le, y
; : : stg 4
i
: we : 4 a
. 1 . : : : :
Py ' 5 4
é = 7
1 . vy m
; i
ry
; a ;
: ,
* }
: 1
;
{
* A ‘
s 4

j »
-
i
@
426,
=
ee me
n v
a |
)
kK
~
Ne
{
. : sy
,
‘
; i 1
1
ay ‘
d
. i Ae
+ 7 Mp
io =
‘
. ol
‘oe
Pd ,
‘ A
‘ > A
.

AMERICAN JOURNAL OF BOTANY. VoLuME IV, PiaTte VII.

BLIZZARD: OMPHALIA CHRYSOPHYLLA

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE VIII.

BLIZZARD: CLITOCYBE ADIRONDACKENSIS

us

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE IX.

BLIZZARD: CLITOFILUS NOVEBORACENSIS

|

&

ce

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE X.

BLIZZARD : CLITOPILUS NOVEBORACENSIS AND CLITOCYBE CERUSSATA

>
any
ag
~ {3
(es
‘
\
1
f
hh »
j
;
i

1
'
‘
rt rE
+ ye
.
L
.
1
.
'
\
> ‘
‘
. =

”

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE XI.

BLIZZARD: CLITOCYBE CERUSSATA

4

THE ORIGIN AND DEVELOPMENT OF THE GALLS PRO-
DUCED. BY. TWO. CEDAR’ RUST (FUNGI

J. L. WEIMER

The question of the origin of the outgrowths caused by Gymno-
sporangium Juniperi-virginianae Schwein. and Gymnosporangium
globosum Farlow on Juniperus virginiana L. has never been settled
satisfactorily. The galls produced by G. Juntperi-virginianae have
been studied by several workers but there still exists considerable
difference of opinion as to the method of their origin. The excres-
cences caused by G. globosum have been studied but little.

} While making observations on these galls incident to the prepara-

tion of another paper the writer became interested in their method of
origin. Observations were made throughout two summers and the
earliest stages of the development of these galls were studied in the
field and later microscopical studies were made. The results of these
observations and studies together with a résumé of the literature on
the subject are given below.

G. JUNIPERI-VIRGINIANAE

Farlow (1880) states that prior to the time of writing it had been
generally accepted that the cedar apples originated in the young cedar
stems but that so far as he could ascertain they were deformed leaves.
Sanford (1888) studied the pathological histology of the galls produced
by this fungus and decided that the galls are modified cedar leaves,
while Wornle (1894) after also studying these galls histologically
concluded that they originated from the stem. Heald (1909) thinks
that the cedar apples originate from the stem in the axis of a leaf.
Kern (1911) places G. Juniperi-virginianae among the foliage inhabit-
ing species and Coons (1912) states that while he has never observed
or produced infection artificially it is evidently a leaf infection. Reed
and Crabill (1915) claim that the cedar apple is nothing but a hyper-
trophy of a cedar leaf infected by the fungus G. Juniperi-virginianae.
Giddings and Berg (1915) picture minute galls situated near the end
of cedar leaves, hence apparently originating from the leaf. Steward

241

242 J. L. WEIMER

(1915), after having studied the histology of these growths, concludes
that they originate as modified axillary buds; the leaf tissue becoming
involved later. |

The writer’s observations go to show that the cedar apples caused
by this species usually break through the upper or inner side of the
leaves, the first evidence of infection being the discoloration of the
whole or a part of a leaf, followed later by a swelling usually from the
upper surface but more rarely from the sides. The young galls grow
rapidly and assume the characteristic shape and color very early in
their development. It was found that when the infected leaves were
removed the galls remained attached (Pl. XII, Fig.1). This led to the
belief that they must be in very close association with the leaf and
perhaps originate from it. Specimens such as are pictured by Gid-
dings and Berg (1915) and Coons (1912), where the galls are located
near or even beyond the center of the leaf, were found in considerable
abundance (PI. XII, Figs.2and 3). Thisstrengthened the theory that
these galls originate from the leaf. If these galls originated in the
stem or as modified axillary buds with separate fibro-vascular systems
it would be reasonable to suspect that in the very young stages at least
the gall would be more firmly attached to the stem than to the leaf.
A single gall has been found by the writer which has the appearance
of having originated from the stem and it may be true that this mode
of origin also exists, although it is certainly not the common method
about Ithaca, New York. The writer has had the privilege of exam-
ining young galls from West Virginia and Wisconsin and the method
of origin herein described was also found in those galls.

Before proceeding with a discussion of the internal anatomy of
these galls a brief description of the structure of the healthy cedar leaf
and stem will be given. The cedar leaf is attached to the stem through-
out a large part of its length, only the apical portion being free. In
cross section the leaf is triangular in outline at the apex but gradually
becomes four-sided toward the base. The epidermis consists of a
single layer of somewhat flattened, elongate cells with the outer wall
covered by a thick layer of cutin. The epidermal layer on the upper
or inner side of the leaf is broken by numerous stomata. Beneath
the epidermal layer is a hypodermis on all the sides except the upper.
For the most part this consists of a single layer of sclerenchymatous
cells. This may however be reinforced at certain places, principally
at the corners and in the region of the resin duct, by additional cells

GALLS PRODUCED BY TWO CEDAR RUST FUNGI 243

of the same character. The central part of the leaf is occupied by a
single fibro-vascular bundle of the collateral type. This is composed
of a small group of scalariform tracheids and a group of phloem cells
about equal in size. Just back of this bundle near the base of the leaf
isaresin duct. The remainder of the tissue of the leaf is made up of
parenchyma cells. The parenchyma cells near the upper or inner
surface below the stomata are globose in shape and are loosely ar-
ranged, forming a tissue similar to mesophyll in appearance. The
outer layers of cells are elongate, the long axis being perpendicular to
the surface forming a palisade tissue. The structure of the very
young stems which bear the young cedar apples is only slightly dif-
ferent from that of the leaf except of course that the fibrovascular
system consists of a medullated central cylinder which is split up into
several collateral bundles by the presence of leaf gaps. The cortical
tissue of the stem and the parenchyma cells of the leaf are so much
alike that it is impossible to distinguish between them. The parts of
the stem not covered by leaves are protected by an epidermal layer
similar to that of the leaf.

One of the first and most conspicuous things which may be observed
in a longitudinal section of a stem bearing a young gall (Pl. XII, Fig. 4)
is the position of the gall as compared with that of the opposite leaf.
It is evident in every case that the gall occupies a position identical
with that of the leaf on the opposite side of the stem. There is no
sign of an axillary structure of any kind. Usually the leaf whose
position the gall occupies and on which it develops becomes distorted
beyond recognition except that there is evident a portion of its tip.
A section through the leaf bundle at the base of the gall shows clearly
that the vascular bundles of the gall arise from this leaf bundle. This
is best studied in galls which have originated some distance from the
axil of the leaf as shown in Plate XIII, Figures 1 and 2. In these figures
it will be seen that the gall has been formed by the production of a
large number of parenchyma cells from the parenchyma of the leaf,
and by the vascular bundles which have arisen from the leéaf-trace
bundle. Examination of serial sections of such a gall precludes the
possibility of the existence of any separate vascular bundle in the leaf
from which the gall bundles might have arisen. In cases where the
gall lies at or near the base of the leaf and from external appearances
might possibly be axial in nature, serial sections show no vascular
supply derived from the stele except the normal small leaf trace or its

244 J. L. WEIMER

modification. Stewart (1915) decided that the gall: bundles are derived
from the central cylinder entirely separate from and above the leaf-
trace bundle. He illustrates this in Figure 1 of his paper, where he
shows at K a section of an axillary bud from which he states the cedar
apple is formed. The writer"has in only one case found a structure
similar to that represented by Stewart. In this case (Pl. XV, Fig. rb)
the structure in question is a section of one side of a terminal bud.
An examination of all the sections in the series reveals the presence of
the embryonic leaves. The young gall (g) beside this bud shows dis-
tinctly the difference in the appearance of a true gall and a bud.
Evidently Stewart has mistaken a normal axillary bud for a young
gall. The writer was permitted to examine some of Stewart’s slides
and this convinced him that Stewart was mistaken in thinking these
structures to be young galls. A careful search of these slides failed to
reveal the presence of mycelium in the buds. Stewart admits that
‘the fungus has not entered the stem at this stage,’ but concludes that
these axillary buds are young galls because structurally these two seem |
to him alike. So far as seen by the writer this worker’s sections show
no cases which, when carefully interpreted, as discussed below, demon-
strate the axial nature of the gall.

The excrescence caused by G. Juniperi-virginianae in its earliest
stages consists simply of a few large parenchyma cells similar to those
of the leaf. Often no distinct epidermal tissue is apparent at this
stage but before the galls enter the winter condition a few layers of
cork cells are laid down. The time at which this exterior covering is
formed varies in different galls. The beginning of such a layer of cork
cells is evident in some very young stages while in other cases galls
nearly mature show almost no sign of its development.

That the fibro-vascular system of the gall originates from that of
the leaf is evident from the study of the very young stages. How this
takes place may be seen in Plate XIII, Figures rand 2. The leaf-trace
bundle first shows an increase in size beneath the enlarged portion of
the leaf. Soon strands of vascular tissue are found leaving the leaf
bundle at almost an angle of ninety degrees and passing into the young
gall. The vascular tissue of the gall develops rapidly and very early
in the development there is present a large amount of conductive tissue
in the gall. This same method of origin of the vascular tissue of the
gall can be traced in those growths which occur near the base of the
leaf. In this case, however, the leaf-trace bundle is very materially

GALLS PRODUCED BY TWO" CEDAR RUST FUNGI 245

affected by the gall and it soon becomes developed to such an abnormal
extent that its identity is nearly or quite lost. Reed and Crabill (1915)
give a good diagrammatic drawing of the bundle of a young gall
originating near the stem. Stewart (1915) thinks that had Reed and
Crabill made transverse instead of longitudinal sections of this infected
leaf they would have found two bundles entering the gall rather than
one. Such sections of numerous galls have been made by the writer
and in no case has more than the one abnormally large bundle been
found. In Plate XIV, Figures 1, 2, and 3 are shown sections from a
series cut from an infected leaf, the gall being formed near the axil.
A section taken a little way above the junction of the leaf and stem is
represented in Figure 1 (line a—d, text figure 1). Here it is evident that
the vascular bundle has been affected since
it has more than doubled in size. A section
taken farther from the stem is illustrated in
Figure 2 (line c-d, text figure 1). This
shows the bundle split into three parts by
the intercalary formation of large cells
filled with resin. These segments of the
vascular system later branch out and be-
come diffused throughout the gall (Fig. 3)
(line e—f, text figure 1). A photograph of
a stage somewhat comparable to Stewart’s
text figure is shown in Plate XIV, Figure.
4, ‘Fhe central cylinder of the stem is "¢,“*#wing of a portion of a
4 hes cedar twig with two leaves
sliowm at @ and passing off from this is the itached. Lines a}. cd. and
greatly enlarged and modified leaf trace ef show the Anpronitate po-
bundle breaking up and passing out in all sitions from which the sec-
directions in the gall. In Plate XIII, Fig- tions illustrated in Plate XIV,
ure 3, is shown a transverse section of ier Cs ashes uence,
; were taken. s—stem, g—gall
ene stem at swith two leaves at 1 and 4.4 7 jeag
l’. From one side of leaf / a gall (g)
is being produced. The vascular system of leaf / is much enlarged
and from it strands of vascular tissue (v) extend into the gall (g).
These figures check the opinion of Reed and Craybill concerning the
single bundle supply. In cases where the gall occurs near the leaf
base the increase in vascular tissue occasioned by its presence enlarges
the leaf trace even through the cortex. At the base of the gall its
vascular tissue frequently takes the form of an irregular hollow cylinder

TEXT-FIG. 1. Diagramma-

246 J. L. WEIMER

simulating that of a branch. To interpret correctly, especially in
longitudinal sections, the enlarged and irregular base of the leaf trace
(a mass of tissue sometimes even near its base partially broken up,
and dividing soon into two masses, the larger upper one simulating a
branch stele) serial sections are clearly absolutely essential. It is
quite probable that Stewart has drawn his conclusions from individual
sections. It is easy to see, further, how in this case a longitudinal
section that is not quite median might lead to erroneous conclusions.

Sanford describes exactly the same condition that the writer has
found in numerous cases. The writer therefore concludes with the
majority of investigators along this line that in most cases at least
and probably in all cases the gall is foliar, and does not represent a
transformed branch.

G. GLOBOSUM

There has been no controversy in regard to the origin of the gall
produced by G. globosum. Heretofore most workers have assumed
from the external appearance of the old galls that they originate in
the stem. Farlow (1880) who first named this fungus states that
unlike G. Juniperi-virginianae it does not break through the central
part of the leaf, but bursts through the stem at the point of attach-
ment of the leaves. Pammel (1905) states that the galls break through
the stem where the leaf is attached. Kern (1911) described the telial
stage of this species, as being caulicolous. Stewart (1915) gives an
account of histological studies made which he interpreted as showing
beyond a doubt that this cedar gall originates from the limb as has
always been supposed.

In order to make more careful observations on this subject a small
cedar tree about four feet high and bearing numerous cedar apples
was selected and all the galls removed early in April (1914) in order
that they might not be confused with other galls appearing later.

This tree was kept under close observation and on July 25 the
first young gall was visible. No aecia were mature at this time.
The young galls seemed to be composed of modified portions of leaf
rather than stem tissue. These galls were tagged and their develop-
ment followed throughout the summers of I914 and 1915. They
grew very slowly and in late autumn were not more than two milli-
meters in diameter. The following spring (1915) these cedar apples
sporulated, thus showing that this fungus, like G. Juniperi-virginianae,
requires nearly two years for the completion of its life cycle.

GALLS PRODUCED BY TWO CEDAR RUST FUNGI 247

On March 19, 1915, several small cedar trees were planted in pots
in the greenhouse and on April 7 several leaves were found on these
trees from the surface of which telial horns were developing (Pl. XV,
Fig. 2). One or more were seen to come from the upper surface of
the infected leaves which were swollen very little or not at all. These
telial horns resembled those of G. globosum in shape and color and the
spore measurements corresponded to those of that species. Inocu-
lations were made on Crataegus leaves with some of these spores and
the characteristic aecia of G. globosum developed; thus showing that
the original determination was correct. Later similar specimens
were found in nature. Often the infected leaves are yellowed through-
out a certain portion of their length and the telial horns develop from
those discolored areas. These tentacles may be found developing
from any part of the upper surface or side of the leaf. Sections of some
of these leaves showed them to be completely permeated with mycelium
which in some cases at least did not extend to the base of the leaf.
Infection must have undoubtedly occurred in the leaf.

Having observed that galls of G. globosum sometimes originate
in the leaf, more careful observations were made to determine if pos-
sible whether this is always true. A great number of galls of this
species were examined both during the autumn and winter of 1914
and 1915 and during the summer of 1915. Hundreds of galls were
examined and in every case the foliar origin was found. These galls,
however, usually develop near the base of the leaf and displace a cer-
tain part of it. As the galls continue to develop the terminal portion
of the leaf remains attached to the gall and may be found here for
some time. A careful study of Plate XV, Figs. 3 and 4 will make
this point clear. A large amount of variation occurs. In some cases
the gall may grow out from the upper surface of the leaf as do the galls
caused by G. Juniperi-virginianae, or they may burst out of the side.
A close inspection of older galls showed in nearly every case the dead
tip of the original leaf still intact (Pl. XV, Figs. 3, 4, 5 and 6).

The gall grows slowly and is perennial, forming spores for several
years. In the early stages these galls are nearly mahogany red in
color as compared with the green color of the minute galls of G. Juni-
peri-virginianae. The red color gradually changes to grayish brown
in the older galls. The shape of these galls is more or less globose
from the beginning and often flattened on the side next to the stem
(Pl. XV, Fig. 4). When the gall becomes older, it displaces the leaf

248 J. L. WEIMER

as stated above and as it continues to develop from year to year it
becomes firmly attached to the twig, appearing to have originated in
the twig (Pl. XV, Fig.5 and PL) Figs):

In case of the above mentioned infected leaves where there was
scarcely any swelling, the infection presumably took place in the
summer of 1913 but was not apparent in the late summer or fall of
1914 and first became obvious in the spring of 1915. That the fungus
had been developing in the leaf for some time seems certain when it is
considered that in nineteen days after the trees were removed to the
greenhouse telial horns had been produced. For some unknown reason
the characteristic stimulation of cellular activity did not occur and
when the mycelium reached the spore-bearing age, spores were pro-
duced.

Other cedar trees brought into the greenhouse early in the spring
of 1914 produced cedar apples during the spring of 1915. These were
scarcely more than telial horns coming directly from the leaf as in the
other cases described. These were probably infected in the fall of
1913 and the mycelium was able to live in the leaf from that time until
the spring of I9I15, or approximately two years before causing any
noticeable effect on the host.

These small galls developing on the leaf at considerable distance
from the stem seldom reach any great size, probably due to their
distance from the stem and a consequent lack of sufficient vascular
tissue development.

A study of a large number of serial sections through the stem and
young gall shows a condition such asis apparent in Plate XVI, Figures
I, 2, and 3. Plate XV, Figure 7, showsa section of a cedar leaf which
had a slight discoloration but almost no swelling. The leaf when sec-
tioned was found to be permeated with mycelium. A corky exterior
layer K is already being developed in the gall shown in Plate XVI,
Figure 1. The resin duct 7 is present and the vascular bundle is the
leaf-trace bundle somewhat enlarged. Figure 2 shows much the same
condition. Figure 3 illustrates a still more advanced stage. In this
section the tip of the old leaf still remains visible at the apex and the
corky exterior covering is well developed. The gall has become closely
attached to the stem similar to the condition found in old galls where
the stem tissue is probably also involved.

GALLS PRODUCED BY TWO CEDAR RUST FUNGI 249

SUMMARY

The galls produced by G. Juntpert-virginianae and G. globosum
on Juniperus virginiana originate as modified leaves.

The vascular systems of the galls are composed of the enlarged
and modified leaf-trace bundles.

ACKNOWLEDGMENTS

Grateful acknowledgment is made to Dr. Donald Reddick for
valuable suggestions and assistance and especially to Dr. Arthur J.
Eames for assistance in the preparation of material, interpretation of
slides and for criticism of manuscript.

DEPARTMENT OF PLANT PATHOLOGY,
CORNELL UNIVERSITY

LITERATURE CITED

Coons, G. H. Some Investigations of the Cedar Rust Fungus, “symnosporangium
Juniperi-virginianae. Nebr. Agr. Exp. Sta. Rept. 25: 215-246. I912. Z

Farlow, W.G. The Gymnosporangia or Cedar Apples of the United States. Boston
Soc. Nat. Hist. Anniv. Mem. 1880: 1-38.

Giddings, N. J. and Berg, A. Apple Rust. W. Va. Agr. Exp. Sta. Tech. Bull. 154:
I-73... 1915.

Heald, F. D. The Life History of the Cedar Rust Fungus, Gymnosporangium
Juniperi-virginianae Schw. Nebr. Agr. Exp. Sta. Rept. 22: 105-133. 1909.

Kern, F. D. A Biologic and Taxonomic Study of the Genus Gymnosporangium.,
INDY. Bot./Gard, Bull.7:.392-494. ‘I91T1.

Pammel, L.H. The Cedar Apple Fungi and Apple Rust in Iowa. Iowa Agr. Exp.
Sta. Buil. 84: 1-36. 1905.

Reed, H. S. and Crabill, C. H. The Cedar Rust Disease of Apples Caused by
Gymnosporangium Juniperi-virginianae Schw. Va. Agr. Exp. Sta. Tech.
Bull. 9: I-106. 1915.

Sanford, Elmer. Microscopical Anatomy of the Common Cedar Apple (Gymno-
sporangium macropus). Annals of Botany 1: 263-268. 1888.

Stewart, Alban. An Anatomical Study of Gymnosporangium Galls. Amer. Journ.
Bot. 2: 402-417. 1915.

Wornle, P. Anatomische Untersuchung der durch Gymnosporangium-Arten her-
vorgerufenen Missbildungen. Forst. Nat. Zeitschr. 3: 129-172. 1894.

EXPLANATION OF PLATES XII-XxVI
PLATE XII

Fic. 1. Young galls caused by Gymnosporangium Junipert-virginianae showing
their axillary position and their relation to the leaf. The two galls at the right
were removed by pulling on the tips of the leaves to which they are attached. Com-
pare the method of origin here with that shown for G. globosum galls in Pl. XV, Fig. 3.

250 J. L. WEIMER

Fic. 2. Mature gall of G. Juniperi-virginianae developed from the upper
surface of the leaf and producing one telial horn.

Fic. 3. Three mature galls of G. Juniperi virginianae with telial horns partly
gelatinized. These galls have evidently developed from the upper side of the
leaves upon which they occur.

Fic. 4. Longitudinal section of stem and leaves of young cedar twig showing
the relation of the gall (g) to the leaf (2) which bears it and to the leaf on the opposite
side of the stem.

PLARE. ChE

Fic. 1. Young gall (g) forming on the leaf (/) at a considerable distance
from the stem (s). The vascular tissue in the young gall is very abundant and arises
from the leaf-trace bundle (f).

Fic. 2. A young gall borne near the tip of the leaf showing the vascular
development asin Fig. 1. The letters correspond to those in Fig. 1. The connection
of the vascular tissue is more readily visible. ;

Fic. 3. A transverse section of a stem (s) with two opposite leaves (/ and 1’).
A gall (g) has developed from the side of leaf / and vascular strands (v) are derived
from the enlarged leaf-trace bundle at f.

PLATE XIV

Fic. 1. Section through a leaf (/) with a basal gall, the section taken as shown
in diagram and transverse to the leaf trace. The vascular bundle (v) is considerably
enlarged. Only one bundle is present, supplying both leaf and gall. This precludes
the possibility of a separate origin of the vascular system of the gall, z. e., of the axial
nature of the latter. (See text figure 1.)

Fic. 2. Section of the same leaf (/) as shown in Fig. r but taken farther from
the stem (s). The vascular bundle has broken into three distinct segments.

Fic. 3. Section from the same leaf as in Figs. 1 and 2 but taken still farther
from the stem. Here the vascular tissue has become much diffused.

Fic. 4. Transverse section of a medium-sized gall (g) and the stem which bears
it (s). The leaf onthe opposite side of the stemisshownat /. The vascular tissue of
the gall originates as one large strand at a which finally breaks up into a fan-like
system of bundles. How this takes place is made clear by a careful study of Figs.
1; 2and.2

PLATE XV

Fic. 1. Longitudinal section of young stem (s) showing terminal bud (0),
young gall (g) and leaves (J). The bud (0) has identically the same appearance as
the young gall shown by. Stewart (1915) in Fig. 1 of his paper. There is no mycelium
in this bud while mycelium is abundant in the gall beside it.

Fic. 2. Telial horns of G. globosum issuing directly from the leaf.

Fics. 3 AND 4. Young galls of G. globosum originating from leaves, the tips
of which are apparent at the top of the galls. The white appearance of the upper
portion of the galls is due to fragments of the leaf tissue.

Fics. 5 AND 6. Mature galls showing the remains of the leaves from which
they originated. The galls shown in Fig. 5 have fruited more than once.

Fic. 7. A transverse section of a leaf which was slightly discolored and very
slightly swollen at the base. The leaf is permeated with mycelium throughout nearly

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE XII.

WEIMER: GALLS PRODUCED BY CEDAR RUST FUNGI

AMERICAN JOURNAL OF BOTANY, VOLUME IV, PLATE XIII.

WEIMER : GALLS PRODUCED BY CEDAR RuST FUNGI

.

¥ >

1

: 2

i
‘
ar 4
4 ! *
a
7 ai
: 4 cs %
as
ia
7 h
'
a
"y :
a

“
vy
F -

. ‘

’

i

¢
‘
’

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE XIV.

WEIMER: GALLS PRODUCED BY CEDAR RUST FUNGI

VOLUME IV, PLATEIXV.

AMERICAN JOURNAL OF BOTANY.

GALLS PRODUCED BY CEDAR RusT FUNGI

WEIMER

AMERICAN JOURNAL OF BOTANY. VoLUME IV, PLATE XVI.

WEIMER: GALLS PRODUCED BY CEDAR RUST FUNGI

aD 3S

rat

ae

&

a)

GALLS PRODUCED BY TWO. CEDAR RUST FUNGI 251

its entire length. A layer of cork has been laid down in the cortical tissue as denoted
by the dark line extending from the base to about the center of the leaf (see a—b
in photograph).

PEATE OV

FIGs. I AND 2. Sections of leaves (J) affected with G. globosum showing
resin ducts (7) and their relation to the stems. The white area beneath the epidermis
(k) in both galls is the corky covering which develops very early in galls caused by this
fungus.

Fic. 3. Section of a gall in a more advanced stage than represented in Figs. 1
and 2. The tip of the leaf is evident at / and the corky layer (K) surrounds the gall
on all free sides. The gall is firmly attached to the stem and it can easily be seen
how the condition shown in Pl. XV, Fig. 5, develops.

Pais
Fa ees SS

rn

0p Michigan e

\

amount (a ie appr

" : Up for halts nee

Yule

yn Pee

sansO la
id

4,

AMERICAN
JOURNAL OF BOTANY

a)
S|

VOL. IV May, I9I7 No. 5

THE PERENNIAL SCAPOSE DRABAS OF NORTH
| AMERICA

EpWIN BLAKE PAYSON

The treatment of the group of Drabas here presented grew out of
an attempt to name some seven members of this group collected in the
summer of 1916 in the mountains of central Idaho by J. F. Macbride
and the author. It soon became evident that available literature
was entirely inadequate and that either specific lines were extremely
hard to define or that a number of species were being treated under
comparatively few names. Careful study of adequate herbarium ma-
terial failed to show intermediate forms and it was noticed that each
form (of which several specimens could usually be found) was restricted
to arather limited range. Differences that at first sight seemed incon-
sequential proved to be constant. The conclusion was obvious that
there were within this group many distinct and easily separable species.
That they should have long passed for a few was not strange when
their diminutive size was considered. Practically all are plants of
arctic or alpine habitats and this alone might account for their very
similar aspect. Under similar conditions the various species have
developed along analogous lines. The author believes that with the
recognition of many instead of few species practically all difficulty in
the determination of these plants will disappear.

In the twenty-six species comprising this group three are circum-
polar and are also found in the main chain of the Rocky Mountains.
One is a local species from the region of Hudson Bay. Twenty-two
are peculiar to western North America and their distribution will
be considered in detail. Seven have been found in the Rocky Moun-
tains proper and, of these seven, two have possibly been developed
farther westward and have migrated east. Fifteen are truly western,

253

“onal Mus®©/

fi lng tit
Z.

25d EDWIN BLAKE PAYSON

since no representatives have been reported from the Rocky Moun-
tains. Of these fifteen, four are peculiar to Idaho and seven are found
farther west in Nevada, California and Oregon. Sixteen of the twenty-
two are not found as far north as the southern boundary of Washington
and perhaps none extend farther south than the southern boundary
of Colorado and Utah. It can thus be seen that the great multipli-
cation of species occurs between the thirty-seventh and forty-sixth
parallels of north latitude and west of the main chain of the Rocky
Mountains. This region is characterized by many more or less isolated
mountain ranges, ranges often separated by many miles of low terri-
tory. Since climatic and edaphic conditions are in general very similar
in most alpine regions and the specific distinctions are not of such a
nature as to be considered adaptive, we can scarcely believe that the
differentiation of species has come about by selection. There is,
however, a close relation between isolation and the multiplication of
species and this factor must be considered seriously in any attempt
to account for the evolution of this group.

Characters that do not vary with ecological conditions and
remain constant in a given species are often quite different in
different families and even genera. Of paramount importance to
the systematist in these Drabas is the character of the pu-
bescence. Classification must always be based primarily upon this
character. The uniformity of the pubescence within a species is
most interesting and the use of a lens or even a compound micro-
scope is imperative. Some species never bear truly stellate hairs,
as for example, D. alpina and D. fladnizensis. One very inter-
esting species bears no other hairs on the leaves except strong marginal
cilia. In some species no cilia are present and the leaf surface is
covered with sessile stellate hairs as in Lesquerella. D. cruciata
and D. asterophora are of especial interest because of the prevalence
of cruciform hairs. Nearly every species might be determined by
the character of the pubescence alone. Other characters of consider-
able importance are the presence or absence of pubescence on the pods,
the length of the style, the shape of the pod, the shape of the leaf and
the form of the fruiting inflorescence.

The material cited under the specific names may be found in the
Rocky Mountain Herbarium at the University of Wyoming and in
the herbarium of the Missouri Botanical Garden at St. Louis. To the
curators of these institutions the author wishes to express his gratitude.

THE PERENNIAL SCAPOSE DRABAS OF NORTH AMERICA 255

Much of the material seen has been examined by Dr. E. Gilg. He
used a number of herbarium names which have unfortunately never
been published and there is some confusion as to his conception of
types. His determinations were very valuable and often served to
corroborate the author’s opinions. The types of the species published
by Macbride and Payson are to be found in the Gray Herbarium at
Harvard; those published by the author alone are in the Rocky Moun-
tain Herbarium, with the exception of D. Mulfordae, which is at St.
Louis. Some names are omitted from this treatment that might be
expected here. D. crassifolia Graham is scapose but belongs rather
to the biennial class than to these densely cespitose plants. The
identity of D. densifolia Nutt. remains obscure. Material labeled
as typical in the Gray Herbarium seems referable to D. oligosperma.
Draba eurycarpa Gray has long been considered related to this group
of plants, but there seems to be some doubt about it being a Draba.

KEY TO THE AMERICAN PERENNIAL SCAPOSE DRABAS
Leaves bearing no stellate or conspicuously branching hairs.
Leaves densely imbricated, forming compact, subspherical
tufts on the ends of the caudex branches, pods

corymbose.
ROM SEM aIOLOMS eae t yi rahe Sisters ie lars dive 6 etal O48 Was 4. D. globosa.
Pods pubescent, more densely cespitose................ 5. D. sphaerula.
Leaves not so densely imbricated, scapes mostly over I cm.
long.
Style very short, stigma sessile or subsessile.
Pods racémose, flowers yellow... 2... 6. eo 1. D.alpina.”,
Pods. corymbose; flowers white... ..-...........2..0-. 2. D. fladnizensis.

Style evident, .5 mm. or more long.
Pods glabrous.

Leaves 5-6:'mm. long, obovate........2..... '...15. D. cylomorpha.
Leaves 2-3 mm. long, pods narrowly elliptical..... 3. D. oretbata.
Pods pubescent.

Leaves,broadly. oblanceolate... .2.%......652.54 16. D. Lemmon.
Leaves linear, midvein prominent.

Leaves glabrous except for strong cilia on the

TMAH OIG area ee can ers OUND a Orme ec aoc. 4 7. D. Nelsoni.
Leaves pubescent, cilia weak................ 6. D. Douglasit.

Leaves bearing some stellate or branching hairs.
Pods glabrous.
Style less than .5 mm. long.
Bodsiimearc(. cnt Or, moreviong) =.) a. ev acenis 13. D. lonchocarpa.
Pods broader, less than 1 cm. long.
Leaves stellate...... SP Wee NLA ere te ae ae 12. D. nivalis.

256 EDWIN BLAKE PAYSON

Leaves with simple or branching hairs........... 7.0).
Style evident (over .5 mm. long).
Pods'globoses fe: ac cs ake ara eon eras nee ee 225).
Pods flattened.
Pods 3-4 mm. long.
Leaves obovate \ shui uence eke tn eee 21
Leaves lineariniils Give So. tie eee Chee 9. D
Pods over 5 mm. long.
Leaves broadly lineare 4s 2k. nua rae 14. D,
Leaves broadly spatulate or oblanceolate.
Leaves densely stellate and silvery.......24. D
Leaves more sparsely pubescent, green.
Pods narrowly lanceolate, not over 2
PANU SWAC Gite ee ey ee ee 20.)
Pods broader (3 mm. or more).
Hairs mostlycruciform™. ..,08- 6. 17.508
Hairs simple or branching.......... TD
Pods pubescent.
Seeds-winged, leavesysilvery. 6. 3: i. 0.058 eee eee 23°)
Seeds not winged.
Pods globose, stellate pubescent........ -25..0
Pods more or less flattened.
Leaves densely imbricated, not over 5 mm. long,
silvery: stellateycilia absent... s. eon. oe 26.)
Leaves not so densely imbricated, mostly over
5 mm. long.
Leaves linear or slightly spatulate.
Fruiting raceme corymbose............. Lie)
Fruiting raceme elongated.
Stellate hairs sessile, near apex of leaf...8. D
Stellate hairs not sessile nor usually
restricted to apex of leaf.
Ciliations not conspicuous, plants
not ‘soboliferousauacne sree ncaed tes 92.)
Ciliations conspicuous, plants so-
boliferous..52 5. coe ee epee: 18. D
Leaves not linear.
Leaves silvery-white, entirely stellate..... 24. D
Leaves greenish, or if silvery, cilia evi-
dent.
Pods narrowly lanceolate............20. D
Pods broader (3 mm. or more).
Style .5 mm. long, plants of
Hudson-Bay 202 35 cate pi ae. FOu sD,
Style I mm. long or more (usually).
Ciliations numerous and con-
SPICUOUS i> SR ees 18.)

. alpina.

. Sphaeroides.

. uncinalis.

. oligosperma.
. laevicapsula.

. argyred.

. cruciata,

. asterophora.
. alpina.

. plerosperma.

. Sphaerocarpa,

. subsessilts.

. vestita.

. olagosperma.

. ncerta.
. Mulfordae.

. argyred.

. cruciata.

. Bellis.

. Mulfordae.

THE PERENNIAL SCAPOSE DRABAS OF NORTH AMERICA 257

Ciliations absent or incon-
SPICUOUSIE ne ec ete. Se 19. D. ventosa.

He. ALPINA Le. Sp; Plt 642-9 .1753:

A circumpolar species probably never found within the United
States.

Specimens Examined.—AvLASKA: Hall Island, July 14, 1899,
Trelease & Saunders, no. 3922; same locality, July 14, 1899, Trelease
& Saunders, no. 3924.

2. D. FLADNIZENSIS Wulf. in Jacq. Misc. Austr. 1:147. 1778. 7

Specimens Examined.—CANADA: Digges Island, Hudson Strait,
Sept. 15, 1884, R. Bell; Okkak, N. E. Labrador, G. Auspach; Table
Top Mt., Gaspe Co., Quebec, Aug. 10, 1881, J. A. Allen; Nottingham
island, Hudson Bay, Aug. 26, 1884, A. Bell.. CoLorRapo: Gray’s
Peak and vicinity, Aug. 6, 1885, H. N. Patterson, no. 3; Sawatch Range,
Brandegee, no. 13, 713. GREENLAND: Prakes Fiord, W. H. Burk,
no. 10, 1891; Disco, Ivannersoil, June 22, 1871, Th. M. Fries. UTAH:
Mt. Barette, July 26, 1905, Rydberg & Carlton, no. 7240.

3. Draba oreibata Macbride & Payson, n. sp.

Cespitose perennial: leaves oblong-linear, obtuse, 4-6 mm. long,
midvein evident, glabrous except for unbranched cilia on margins:
scapes leafless, slender, glabrous, 3-5 cm. long: sepals glabrous, yel-
lowish; petals white, twice as long as the sepals: fruiting inflorescence
elongated: pods flat, glabrous, narrowly elliptical, 8-16 mm. long,
2-3 mm. broad; style slender, nearly I mm. long.

D. oreibata probably finds its closest relative in D. fladnizensis.
From this it may be separated by the shorter, more obtuse leaves, the
elongated instead of the corymbose fruiting inflorescence, the elliptical
instead of lanceolate pod and especially by the slender style. In
D., fladnizensis the stigma is sessile or nearly so.

Specimens Examined.—IDAHO: alpine summit of Lost River Mts.,
west of Clyde, Blaine Co., July 10, 1916, Macbride & Payson, no. 3135
(type in Gray Herbarium, duplicate in Ry. Mt. Herbarium). UvtTau:
southern Utah, May 14, 1874, A. L. Siler; southern Utah (near Osmer),
A=H. Siler, no. 34, southern Utah, 1874, Parry, no. 34.

4. Draba globosa Payson, n. sp.

Densely cespitose perennial; caudex much branched: leaves densely
imbricated, forming globose tufts on the ends of the caudex-branches,
scarcely over 5 mm. long, broadly linear, acute, midvein evident,

258 EDWIN BLAKE PAYSON

glabrous except for short, unbranched marginal cilia: scapes glabrous,
rather stout, scarcely I cm. long: flowers unknown: fruiting inflores-
cence corymbose: pods 4-6, broadly lanceolate, flattened, glabrous,
about 5 mm. long, 3 mm. wide; style stout, scarcely I mm. long: seeds
2-4 in each cell, not winged.

This plant was evidently recognized as distinct by Dr. E. Gilg for
I find some specimens in the Rocky Mt. Herbarium labeled by him
with a name which has never been published. The aspect of this
species and the next suggest D. subsessilis Watson. ‘This resemblance
is however probably not indicative of any real relationship.

Specimens Examined.—UTAH: Fish Lake, Uintah Mts., June 17,
1902, Goodding; Little Cottonwood Canyon, Salt Lake Co., Aug. 3,
1904, Garrett, no. 1555; Alta, Wasatch Mts., Aug. 12, 1879, Jones,
no. 1235. Wyominc: La Plata Mines, Snowy Range, Aug. 29, 1898,
E. Nelson, no. 5246a (type in Ry. Mt.-Herbarium). The type was
found on the same sheet with two specimens of ‘ D. andina”’ and was
given an ‘“a’’ number. Other sheets of this collection in various
herbaria may be found also to bear specimens of D. globosa.

5. Draba sphaerula Macbride & Payson, n. sp.

Densely pulvinate-cespitose perennial; caudex much branched:
leaves about 2 mm. long, densely clustered on the ends of the caudex
branches into small, compactly imbricated spheres, glabrous except
for the unbranched marginal ciliae, broadly linear, acute, midvein
obscure: scapes barely rising above the leaves, pubescent: sepals
pubescent, yellowish; petals yellow, exceeding the sepals but little:
fruiting inflorescence corymbose: pods few (1 or 2), broadly lanceolate,
about 3 mm. long, flattened, pubescent with stellate or branched
pubescence; style evident, less than I mm. long: seeds neither winged
nor margined.

D. sphaerula is quite closely related to D. globosa and further col-
lections may even show the two to merge. It differs from globosa
in the pubescent instead of the glabrous pods, the smaller, more
densely imbricated leaves and the shorter scape. That the very short
scapes are not due to ecological factors was evident when the type
of sphaeruia was collected, for it grew among the tufts of D. Nelsoni
in which the scapes were much elongated.

Collected on an alpine slope near Parker Mt., Custer Co., Idaho,
July 17, 1916, Macbride & Payson, no. 3336 (type in Gray Herbarium,
~ duplicate in Ry. Mt. Herbarium).

THE PERENNIAL SCAPOSE DRABAS OF NORTH AMERICA 259

6 DD DeucisAsi: Gray,’Proc. Amer. Acad-7: 328. - 1867.

Braya oregonensis Gray, Proc.. Amer. Acad. 17: 199. 1882.

Cusickia Gray, l. c.

D. Crockeri Lemmon, Bull. Torrey Club 16: 221. 18809.

Specimens Examined.—CALiFoRNIA: Sierras of California, Lemmon.
OREGON: Dry, stony hills, eastern Oregon, May 24, 1898, Cusick,
no. 1883. UTAH: 1858-59, H. Engelmann. ‘This speciman is re-
ferred here very doubtfully. It is however too immature to make
determination certain.

7. Draba Nelsonii Macbride & Payson, n. sp.

Cespitose alpine perennial, caudex-branches clothed with dead
leaves below; leaves linear, 5-7 mm. long, I mm. or less wide,
acute, midvein prominent, glabrous except for the strong marginal
cilia: scapes slender, glabrous or sparingly pubescent, I-4 cm.
long: sepals greenish, usually glabrous; petals yellow, nearly twice
as long’ as the sepals: fruiting inflorescence elongated : pods 6-15, broadly
lanceolate. 3-6 mm. long, 2-3 mm. broad, simply pubescent, flattened:
style 1 mm. more or less long: seeds 1-4 in each cell, wingless.

Of all the newly characterized species, D. Nelsonit seems to be the
most distinct and most widely distributed. Its affinities are undoubt-
edly with the oligosperma group. In manner of growth, elongated
raceme and rather small pods it is quite similar to D. oligosperma
and might even be mistaken for that species except for the total
absence of stellate pubescence in D. Nelsoni1. It is a great pleasure
for the authors to dedicate this plant to their friend and teacher, Dr.
Aven Nelson, through whose efforts and assistance the expedition
was made possible that led to the discovery of this fine species.

Specimens Examined.—CALIFORNIA: Castle Peak, Nevada Co.,
‘Aug. 3, 1903, Heller; Modoc Co., 1898, Mrs. Bruce; rocky, exposed
ridges, Mt. Stanford, Nevada Co., July 17, 1892, Sonne, no. 14.
IDAHO: exposed alpine summit, Antelope Mts., near Martin, Blaine
Co., July 6, 1916, Macbride & Payson, no. 3077 (type in Gray Her-
barium, duplicate in Ry. Mt. Herbarium); Soldier Mts., near Corral,
Blaine Co., June 26, 1916, Macbride & Payson, no. 2894; exposed
summit near Parker Mt., Custer Co., July 17, 1916, Macbride &
Payson, no. 3253. OREGON: Blue Mts., July, 1886, Cusick, no. 1345.
WASHINGTON: Yakima Region, 1882, Brandegee, no. 373 (in part).

8. D. OLIGOSPERMA Hook. Fl. Bor. Amer. 1: 51. 1833.
D. andina (Nutt.) A. Nels. Bull. Torrey Club 26: 352. 1899.

260 EDWIN BLAKE PAYSON

D. saximontana A. Nels. Bull. Torrey Club 27: 264. 1900.

D. oligosperma was described as having white flowers and it is
certain that in D. andina the flowers are yellow. Since there seems
to be no other difference, however, the two species are merged.

Specimens Examined.—BRITISH COLUMBIA: Carbonate Draw,
July 13, 1904, J. Macmillan, no. 297; Canmore, Ry. Mts., June 29,
1885, Macoun; Silver City, Ry. Mts., Aug. 3, 1885, Macoun; N. Fork
of Old Man’s River, Aug. 10, 1883, Macoun. IDAHO: Sawtooth
National Forest, 1910, C. N. Woods, no. 80; Boise, Wilcox. Mon-
TANA: Monida, Madison Co., June 16, 1899, A. Nelson & E. Nelson,
no. 5423; Bridger Mts., June 15, 1897, Rydberg & Bessey, no. 4181;
Bridger Mts., June 4, 1901, W. W. Jones; Bridger Canyon, May 15,
1901, E. J. Moore; Little Belt Mts., Aug. 10, 1896, Flodman, no. 498;
Bridger Canyon, Bozeman, May 27, 1899, Blankinship; June, 1894,
Mrs. Moore; Mt. Bridger, June 26, 1899, Blankinship; Bridger Mts.,
June 15, 1897, Rydberg & Bessey, no. 4180 (this seems to be nearly
typical oligosperma; its flowers are white). NEvapDA: Ruby Hill,
July 7, 1891, Jones; Bunker Hill, Toityabe Range, July 29, 1913,
Kennedy, no. 4184. UtTan: Logan Peak, Cache Co., July aengme:
Charles Piper Smith, no. 2245. Wyominc: Bush Ranch, Sweetwater
Co., June 10, 1900, A. Nelson; Golden Gate, Yellowstone Park, June
28, 1899, A. Nelson & E. Nelson; Laramie Hills, Albany Co., June 3,
1900, A. Nelson, no. 7019; Laramie Hills, Albany Co., May 30, 1898,
A. Nelson, no. 4323; Laramie Hills; June: 21, 1802, 6B. "CiBa7ens:
no. 65; Kemmerer, June I, 1907, A. Nelson, no. 9027; Laramie Hills,
June, 1893, A. Nelson, no. 3237; T. B. Ranch, Carbon Co., June 20;
1901, Goodding, no. 58; Telephone Mines, Albany Co., Aug. 1, 1900,
A. Nelson, no. 7873; Laramie Hills, May, 1895, A. Nelson, no. 1223;
sandy hilltops, Laramie, May 25, 1910, A. Nelson, no. 9334; Freezeout.
Hills, July 10, 1898, £. Nelson, no. 4487;.La Plata Mines, Aug. 29,
1898, E. Nelson, no. 5246; Gros Ventre Fork, June 8, 1860, Hayden;
West Slope of Wind River Mts., June 6, 1860, Hayden; gravelly hills
in Wind River Valley, May 15, 1860, Hayden; near Mammoth Hot
Springs. Yellowstone Park, June, 1893, Burglehaus; Yellowstone
Park, 1885, Tweedy, no. 567; near South Gap, June, 1873, Parry, no.11.

D. pectinata (S. Wats.) Rydb. Bull. Torrey Club 39: 327. I912.
D. glacialis var. pectinata S. Wats. Proc. Amer. Acad. 23: 260. 1888.
These seem according to specimens labeled by Rydberg in the Gray
Herbarium to be referable to D. oligosperma.

THE PERENNIAL SCAPOSE DRABAS OF NORTH AMERICA 261

9. Draba incerta Payson n. sp.

Somewhat cespitose perennial, caudex branching: leaves linear or
linear-spatulate, 7-10 mm. long, about 2 mm. wide, not rigid, midvein
obscure, greyish with long stellate pubescence and weak marginal
cilia: sepals villous; petals yellow, twice as long as the sepals: fruiting
inflorescence elongated: pods numerous (6-14), flattened, simply
pubescent, broadly lanceolate, 4-6 mm. long: style evident, less than
I mm. long: seeds neither winged nor margined.

Draba incerta ia a rather unsatisfactory species appearing almost
as if it were produced by peculiar ecological conditions. It differs so
strikingly in leaf characters from its nearest relative (D. oligosperma),
however, that one could scarcely consider them identical. I find that
Dr. Gilg has evidently considered this as a distinct species, although
he has included more than one species under the manuscript name
which he gave it. It may be distinguished from D. oligosperma by
the thinner leaves which are not at all rigid and by the absence of the
prominent midrib which characterizes that species. The pubescence
too is longer and more diffuse.

Specimens Examined.—ALBERTA: Tunnel Mt., Banff, May 9,
1902, V. B. Sanson; Sulphur Mt., Banff, June 16, 1901, L. R. Waldron.
WASHINGTON: Yakima Region, 1882, Brandegee, no. 371. WYOMING:
among rocks on the summit, the Thunderer, Yellowstone Park, July
13, 1899, A. Nelson & E. Nelson, no. 5818 (typein Ry. Mt. Herbarium);
Mt. Washburn, Yellowstone Park, Aug. 1885, Tweedy, no. 566.

10. D. BELLII Holm, Repert. Nov. Sp. Fedde 3: 338. 1907.
I have seen one specimen of this species and that is from “‘crevices
of rocks,’’ Mansfield Island, Hudson Bay, Aug. 30, 1884, Dr. R. Bell.

11. Draba vestita Payson n. sp.

Very densely cespitose perennial; caudex much branched; leaves
persistent and densely clothing the branches of the caudex; 5-7 mm.
long, I mm. or more wide, broadly linear, thin and not at all rigid,
midvein evident; pubescence rather long, involved, hairs in large part
unbranched, none really stellate; sepals pubescent; petals apparently
yellow, about twice as long as sepals: fruiting inflorescence corymbose;
pods rather few (4-6), broadly lanceolate, flattened, densely pubescent
with simple or branching hairs, 5-8 cm. long: style about I mm. long:
seeds not winged.

The name D. Gilbertiana has been given to herbarium sheets of

262 EDWIN BLAKE PAYSON

this species by Dr. Gilg. Unfortunately it has been impossible to
discover his type and so, in conformity to the Vienna rules, a new
name has been given to the species.

Specimens Examined.—ALBERTA: Sheep Mt., Waterton Lake,
July 28-31, 1895, Macoun, no. 10278. BRITISH COLUMBIA: Heights
above Carbonate Draw, Beaverfoot Mts., July 13, 1904, R. T. Shaw,
no. 305. CALIFORNIA: Tiukuk Knob, Placer Co., Aug. 12, 1892,
C. F. Sonnee, no. 15; summit of range between Devil’s Cliff and Linker’s
Knob, Aug. 10, 1901, Kennedy & Doten. MONTANA: Bridger Mts.,
June 15, 1897, Rydberg & Bessey, no. 4173; Upper Marias Pass, Aug.
3, 1883, Canby, no. 28 (type in Ry. Mt. Herbarium); Mt. Bridger,
June 26, 1899, Blankinship. OREGON: cliffs of the Wallowa Mts.,
July 31, 1899, Cusick, no. 2307. WASHINGTON: Cascade Mts., 1882,
Brandegee, no. 373. WYOMING: Yellowstone Park, 1884, Tweedy,
no. 204.

12. D. NIVALIS Lilj. Svensk. Vet. Akad. Handl. 1793: 208. 1793.

Specimens Examined.—ALASKA: U. S. Coast Survey, 1871-72,
M. W. Harrington. CoLorapvo: Elk Mts., 1881, Brandegee, no.
13268; Sawatch Range, Brandegee, no. 12714 (the Colorado specimens
seen are not typical). GREENLAND: Distr. Holssenborg, Aug. 2,
1886, L. Ko.derup Rosenvinge. HuDSON STRAIT: Nottingham Island,
Aug. 24,1884, R. Bell. LABRapbor: Dead Islands, Aug. 17, 1882, J. A.
Allen; Okkak, N. E. Labrador, G. Auspach; northern Labrador, 1873,
G. Auspach, no. 404.

13. D. LONCHOCARPA Rydb. Mem. N. Y. Bot. Gard. 1: 181. 1900.
D. nivalis elongata Wats. Proc. Amer. Acad. 23: 258. 1886.
Specimens Examined.—CANADA: Kicking Horse River, Ry. Mts.,

Aug. 13, 1890, Macoun; Kicking Horse Lake, Ry. Mts., Aug. 12, 1890.

Macoun. IpAHO: rock crevices, Parker Mt., Custer Co., July 17,

1916, Macbride & Payson, no. 3240. MONTANA: Boulder Creek,

Aug. 1887, Tweedy, no. 36; Upper Marias Pass, Aug. 3, 1883, Canby,

no. 26; McDonald’s Peak, Mission Range, July 19, 1883, Canby, no. 27.

WASHINGTON: Mt. Paddo, July 12, 1886, Suksdorf, no. 836.

14. Draba laevicapsula Payson n. sp.

Loosely cespitose perennial; caudex branching, leafy branches
occasionally 1-2 cm. long: leaves linear, 7-10 mm. long, narrowed
slightly at base, usually obtuse, not rigid, midvein evident; pubescence
rather loosely stellate, marginal cilia evident, especially toward base

THE PERENNIAL SCAPOSE DRABAS OF NORTH AMERICA 263

of leaves: scapes slender, glabrous, about 5 cm. long, naked or with
one or two small bracts: flowers unknown: fruiting inflorescence race-
mose, comparatively short (about 2 cm.): pods 4-7, narrowly lanceo-
late, flattened, 7-9 mm. long, 2-3 mm. wide: style scarcely I mm.
long: seeds not winged.

Dr. Gilg apparently confused D. laevicapsula with D. incerta but
it seems rather to be associated with D. oligosperma.

Specimens Examined.—CANADA: Rocky Mts., Aug. 1885, Macoun;
IDAHO: summit of Steven’s Peak, Coeur D’Alene Mts., Aug. 5, 1895,
Leiberg, no. 1477 (type in Ry. Mt. Herbarium). Montana: Upper
Warias Pass, Aug. 3,1883, Canby, no. 29.

15. Draba cyclomorpha Payson n. sp.

Cespitose perennial: leaves clustered on the apices of the many
branches of the caudex, rounded obovate, fleshy, midnerve indistinct,
5-6 mm. long, 3-4 mm. broad; pubescence simple, largely confined to
the leaf margins: scapes leafless, sparingly pilose, I-2 cm. long: flowers
unknown: fruiting raceme short and corymbose: pods 3-10, glabrous,
flattened, typically nearly circular but at times oblong, 4-5 mm.
broad: style scarcely I mm. long: seeds not winged.

This species has been confused with D. Lemmonzi to which it is in
fact most nearly related. It differs from that species principally in
the glabrous, broader pod.

Specimens Examined.—OREGON: Alpine Wallowa Mts., Aug. 29,
1900, Cusick, no. 2497 (type in Ry. Mt. Herbarium); alpine summits,
Powder River Mts., Aug., 1886, Cusick, no. 1344.

16. D. LEmmont Wats., Bot. Calif. II. 430. 1880.

Specimens Examined.—CALIFORNIA: summit of Mt. Lyell, Aug.
19, 1878, Lemmon (co-type); Mt. Dana, July, 1902, Hall & Bab-
cock, no. 3606; Mt. Dana, June 28, 1863, Brewer, no. 1735; eastern base
of Mt. Brewer, July 4, 1864, Brewer, no. 2811; Mt. Goddard, July,
1900, Hall & Chandler, no. 668; Little Kern Cr., 1897, Purpus, no.
5118; Mt. Dana, July, 1901, H. M. Evans.

17. Draba asterophora Payson n. sp.

Loosely cespitose perennial with rather long trailing caudex-
branches: leaves about I cm. long. 4—5 mm. wide, obovate to oblanceo-
late, obtuse, thickish, midvein obscure; pubescence rather sparse,
consisting mostly of long stalked, cruciform hairs, simple cilia almost
entirely wanting; scapes slender, glabrous, 3-4 cm. long; fruiting

264 EDWIN BLAKE PAYSON

inflorescence shortened with a tendency to become corymbose; pods
6-10, broadly lanceolate, 6-8 mm. long, 4 mm. broad, flattened,
glabrous; style short (about .5 mm. long) but evident: seeds flattened,
broadly winged.

D. asterophora is evidently allied to D. Lemmoni and D. cyclomorpha
as shown by the similar leaves and fruiting racemes. The ranges of
these three species also are rather close. The stellate hairs and winged
seed make D. asterophora easily separable. But one specimen has been
seen and that is from an altitude of 9,000 ft. on Mt. Rose, Washoe
County, Nevada, Aug. 17, 1905, P.. B. Kennedy, no. 1154) (type
in Ry. Mt. Herbarium).

18. Draba Mulfordae Payson n. sp.

Cespitose perennial: leaves linear or slightly spatulate, obtuse,
7-10 mm. long, I-2 mm. broad, not rigid, midvein obscure often rising
above the ground on leafy shoots or sobols; pubescence of strong
marginal cilia and long stalked stellate or branching hairs; scapes
slender, pubescent, 3-4 cm. long; sepals pubescent, petals white (?),
about three times as long as the sepals; fruiting inflorescence elongated ;
pods 8-12, lanceolate, 6-8 mm. long, flattened, pubescent; style slender,
over I mm. long; seeds not winged.

This plant is rather intermediate between the oligosperma and the
ventosa group. ‘The linear leaves, strong cilia and elongated raceme
ally it to the former and the tendency to produce leafy shoots to the
latter. But one specimen has been seen; it is from Soda Springs,
Idaho, June 21, 1892, and was collected by A. Isabel Mulford (type
in Missouri Botanical Garden Herbarium).

19. D. VENTOSA Gray, Amer. Nat. 82212. 91874"

D. Howellit Watson, Proc. Amer. Acad. 20: 354. 1885.

D. sobolifera Rydb. Bull. Torrey Club, 30: 251. 1903.

This seems to be the most variable species of the group. A number
of varieties might be made separating plants of different localities but
it seems to be impossible to draw specific lines within the group of
specimens cited. It should be noted here that the’style in D. sobolifera
is .5 mm. long instead of 5 mm. as the description reads.

Specimens Examined.—CALIFORNIA: Siskiyou Mts., June 16, 1884,
Howell (type no. of D. Howellit). NEVADA: Schellbourne, July I1,
1891, Jones. OREGON: head of Divine Creek, Steins Mts., June 14,
I90I, Cusick, no. 2569; Steins Mts., June 2, 1885, Howell. UTAH:

_ THE PERENNIAL SCAPOSE DRABAS OF NORTH AMERICA 265

Tate Mine, Marysvale, Aug. 28, 1894, Jones, no. 5936 (type no. of
D. sobolifera Rydb.); Delano Peak, July 26, 1905, Rydberg & Carlton,
no. 7231; near Beaver, June 7, 1913, H. Redeker, no. 50. WYOMING:
High Peak between Snake River and Wind River Valleys, 1873,
Parry (type no. of D. ventosa).

20. Draba cruciata Payson n. sp.

Cespitose perennial; caudex-branches slender; leaves oblanceolate,
usually toothed, thickish, midnerve obscure, 7-10 mm. long, 2-3 mm.
broad, acute or acutish; pubescence stellate, each hair usually bearing
four arms, ciliae wanting or inconspicuous; scapes slender, 5-7 cm.
long: sepals yellow, glabrous or pubescent; petals yellow, about three
times as long as sepals: fruiting raceme elongated: pods narrowly
lanceolate, 7-9 mm. long, 2 mm. broad, glabrous or simply pubescent,
flattened: style slender, 1 mm. long; seeds not winged.

The relationship of this plant is not at all evident. Hall suggests
that it is near D. Lemmoni and if it is, it should be placed with D.
asterophora on account of the branched cruciform pubescence. The
slightly toothed leaves are a most interesting development. CALI-
FORNIA: Vicinity of Mineral King, Tulare Co., July 10, 1904, Hall &
Babcock, no. 5361 (type in Ry. Mt. Herbarium).

27. D. UNCINALIS Rydb. Bull. Torrey Club 30: 251. 1903.
I have seen no specimens of this species.
Type Locality.—Tate Mine, Marysvale, Utah.

22. Draba sphaeroides Payson n. sp.

Loosely cespitose perennial; caudex much branched: leaves clus-
tered on the apices of the caudex-branches, narrowly spatulate, obtuse,
green, 3-5 mm. long; pubescence rather long, ciliate and branching,
but few truly stellate hairs present: scapes naked, slender, sparingly
pubescent with branched or stellate hairs, 1-1.5 cm. long: sepals
glabrous or sparingly pubescent; petals yellow, twice as long as the
sepals: fruiting inflorescence racemose, 1.5-2 cm. long: pods 6-12,
glabrous, ovoid, scarcely flattened, 3-4 mm. long; style slender, about
I mm. long.

Plants with globose pods are more or less anomalous in this genus
but in aspect and all other characters this plant is so obviously a
Draba that no one would think of placing it elsewhere. Its affinities
are doubtless with D. oligosperma. Collected above receding snow at
an altitude of 10,800 ft. on Jarbidge Peak, Nevada, July 8, 1912,
Nelson & Macbride, no. 1981 (type in Ry. Mt. Herbarium).

266 EDWIN BLAKE PAYSON

23. Draba pterosperma Payson n. sp.

Loosely cespitose perennial; caudex branched; leaves mostly in
round tufts either on the apices of the caudex-branches or rising above
the ground on sparingly leafy shoots or sobols, oblong, 3-5 mm. long,
I-2 mm. broad, rounded at the apex, midvein evident; pubescence
silvery, loosely stellate, cilia present: scapes slender, pubescent, 2-6
cm. long: flowers showy; sepals pubescent; petals yellow, 7-8 mm.
long, over twice as long as the sepals: fruiting inflorescence elongated:
pods 6 or 8, broadly lanceolate, 8-9 mm. long, 4-6 mm. broad, flattened
and often unsymmetrical, pubescent with stellate hairs: style slender,
2 mm. or more long: seeds about 4 in each cell, broadly winged.

Because of the winged seeds in this species and in D. asterophora
one would be inclined to consider them closely related but such is
probably not the case. This plant seems to be related to ventosa
and so we must assume that the development of winged seeds has been
accomplished independently by two different groups.

Specimens Examined.—CALIFORNIA: rock crevices, Marble Mt.,
Siskiyou ‘Co.; July 10, 19010, (Geo. D. Butler, no, 1716, (pena ys
Mt. Herbarium); Marble Mt., June, 1901, H. P. Chandler, no. 1654
(Mo. Bot. Gard. Herbarium and perfectly typical).

24. D. ARGYREA Rydb. Bull. Torrey Club 30: 251. 1903.

Type Locahty.—Sawtooth Mts., Idaho, head of Pettit Lake.

The specimens cited below are somewhat doubtfully referred here.
Since, however, no authentic material of D. argyrea has been available,
since our plants agree fairly well with the description and are from
the same vicinity it has been thought best to leave the question un-
decided.

Specimens Examined.—IpAHo: rock crevices, alpine basin in
Sawtooth Mts., above Redfish Lake, Blaine Co., Aug. 9, 1916, Mac-
bride & Payson, no 3677; crevices in granitic rocks, Smoky Mts.,
Blaine Co., Aug. 13, 1916, Macbride & Payson, no. 3734.

25. Draba sphaerocarpa Macbride and Payson, n. sp.

Cespitose perennial; caudex much branched; leaves mostly borne
in tufts on erect, nearly leafless shoots that rise above the caudex-
branches, oblong or obovate, obtuse, 4-7 mm. long, thickish and mid-
vein indistinct; pubescence finely and densely stellate; leaves silvery,
cilia absent: scapes pubescent, rather stout: flowers unknown (prob-
ably yellow): fruiting raceme elongated, developing almost from very

THE PERENNIAL SCAPOSE DRABAS OF NORTH AMERICA 267

base of scape: pods many (8—I2), scarcely compressed or flattened,
ovate, stellately pubescent, 2-5 mm. long: style slender, I mm. or
more long: seeds not winged.

This plant is most closely related to D. argyrea and in leaf char-
acters it is practically identical. It is distinguished from that species
by the small, subspherical pods and the peculiar inflorescence which
develops from near the base of the scape. D. sphaerocarpa was col-
lected at a much lower elevation than were the specimens referred to
D. argyrea.

Type.—IDAHO: dry, granitic washes near the head of Redfish Lake,
Blaine Co., Aug. 9, 1916, Macbride & Payson, no. 3677a (Gray
Herbarium).

26. D. SUBSESSILIS Watson, Proc. Am. Acad. 23: 255. 1888.
Type Locality.—“ On the White Mts. of Mono Co., California,
at 13,000 ft. altitude”’ (W. H. Shockley, July, 1886).
Specimens Examined.—CAa.iForniaA: Mt. Dana, June 28, 1863,
Brewer, no. 1735a; White Mts., Mono Co., Aug., 1885, W. H. Shockley.

UNIVERSITY OF WYOMING,
LARAMIE

THE OSMOTIC CONCENTRATION OF THE TISSUE FLUIDS
OF JAMAICAN MONTANE RAIN-FOREST
VEGETATION!

J. ARTHUR HARRIS AND JOHN V. LAWRENCE

I. INTRODUCTORY REMARKS

Purpose of Investigation.—This paper is one of a series in which
various problems involving the investigation of the osmotic pressure
or osmotic concentration of the fluids of plant tissues are treated.
Specifically it presents an extensive series of determinations of the
freezing-point lowering of the extracted leaf sap of plants from the
Blue Mountains of Jamaica, discusses the differences in these values
in their relation to local differences in the environmental complex, and
briefly compares the series as a whole with others now available.

In another place (Harris, Lawrence and Gortner, 1916) we have
put forward in detail the arguments for the carrying out of such
studies as a regular part of systematic and thoroughgoing phyto-
geographical investigation. It seems unnecessary, therefore, to repeat
these arguments here.

After completing a series of determinations of the osmotic con-
centration of the tissue fluids of a number of species of plants from
the southwestern deserts, in the vicinity of the Desert Laboratory
during the winter and spring months of 1914, and comparing them
(Harris, Lawrence and Gortner, 1915) with a series made in the more
mesophytic habitats in the neighborhood of the Station for Experi-
mental Evolution on Long Island, the next most desirable step seemed
to be. the investigation of the sap properties of the plants of an ex-
tremely hygrophytic region.

Since such field studies could be most conveniently carried out
during the winter months, at a time when we could be absent from

1 Results of investigations carried on at Cinchona, by courtesy of the British
Association for the Advancement of Science and the Jamaican local government,
under the joint auspices of the department of botanical research and the department
of experimental evolution of the Carnegie Institution of Washington, and with the
collaboration of the New York Botanical Garden.

268

OSMOTIC CONCENTRATION OF TISSUE FLUIDS 269

experiments under way at the Station for Experimental Evolution, it
was quite natural to think of the Tropical Laboratory at Cinchona,
established by the New York Botanical Garden and later maintained
by the British Association for the Advancement of Science and the
Jamaican local government, as the most promising locus for such
work. This station presents the advantages of furnishing living
quarters and laboratory space on the edge of a primaeval montane
rain forest within twenty miles of a point where ice, essential for the
preliminary freezing of tissues for the extraction of sap, can be ob-
tained. This was quite successfully packed over the Port Royal
Mountains, through the Yallahs river valley and up to Cinchona on
mule back by negro helpers.

To Professor Bower and the other members of the British Associa-
tion committee in charge of the Tropical Laboratory at Cinchona
and to Mr. Wm. Harris, F.L.S., superintendent of public gardens
and plantations, we are indebted not only for the use of the laboratory
but for other courtesies that added to the success and pleasure of
our work while in Jamaica.

Characteristics of the Region Investigated.—The higher portions of
the Blue Mountains are characterized by a relatively low but uniform
temperature, by a large and well-distributed rainfall, accompanied
by much fog and cloudiness and high relative humidity.

The rainfall upon the northern is far greater than that upon the
southern slopes of the mountains. The averages given by Shreve
(1914) for the upper mountains, in which all our collections were
made, are:

WO ina oiomano eet Maes nc cent Sukh, Sons So ees 105.70 inches
Ne wypebslienvie ia Gao: Sn 4 ee Ad ag a, Ge eb ca la 113.85 inches
Bite ountain Peak 6.0 tego posh og so dahl na sed 130.48 inches

Notwithstanding the heavy rainfall there are neither ponds nor
constant streams above 4,500 feet, but in places there are depressions
on the higher portions of the main ridge of the mountains which are
developed as sphagnum bogs. Below 4,500 feet the water emerges to
feed numerous swift mountain streams. Transient water courses are
found much higher.

While the rainfall is large it is not comparable with the maximum
precipitations known in other tropical plant environments. Further-
more the amount varies greatly from year to year, both in quantity

270 J. ARTHUR HARRIS AND JOHN V. LAWRENCE

and distribution. Thus Shreve (1914), in working through the records
which have been kept at Cinchona for the past thirty-nine years,
finds variation in the total annual precipitation from about 59 to
about 179 inches. In October, the rainfall has varied from about
3 to 43 inches. In February, precipitation has ranged from less
than an inch to nearly 13 inches. At New Haven Gap during the
three months of April, May and June, 1892, there was not a measur-
able amount of rainfall, whereas during the same three months in
1894 there fell 62 inches of water.

Thus the vegetation is by no means free from occasional periods
of drought.

Notwithstanding this fact, moisture is so great in quantity and
so uniform in distribution that it supports a dense evergreen arbores-
cent and herbaceous vegetation, a large proportion of the constituent
species of which are of a pronouncedly hygrophilous character. As
a factor in the development and maintenance of the vegetation, the
distribution as well as the actual quantity of the precipitation is a
factor of great importance. Precipitation is almost exclusively in
the form of light showers of brief duration or gentle and long con-
tinued rain, but never in the torrential downpours so characteristic
of deserts and tropical lowlands. ‘Transient showers of too: brief
duration to be registered as giving a measurable quantity of rainfall
are frequent. Shreve gives a table showing that at Cinchona on an
average from one third to two thirds of the days of the twelve indi-
vidual months of the year have a measurable precipitation.

On the northern slopes fog is prevalent from below 4,500 feet to
the summits of the highest peaks from 10 a. m. to 4 p. m. on a large
proportion of the days during all the months of the year, with the
possible exceptions of July and August. Fog is much less frequent
on the southern exposure of the mountains, but even here it is often
seen on the upper slopes, and a large percentage of the days are cloudy
or partially cloudy. Shreve, after nearly a year’s residence in the
Blue Mountains, describes the condition as follows: ‘‘The typical
course of the day’s weather is: clear from sunrise until 9 to II a. m.,
intermittently or entirely cloudy until nearly sunset, with two or
three hours of fog in the mid-day or early afternoon, the sun setting
clear. Rain usually occurs in the mid-day or early afternoon and
the night is clear.”

As a consequence of the high and well-distributed rainfall and

OSMOTIC CONCENTRATION OF TISSUE FLUIDS 271

the prevalence of fog, atmospheric humidity is high, ranging from
about 80 to about 89 percent in the various months of the year, with
an annual average of about 84 percent.

Temperature is low and remarkably uniform throughout the year.
At a depth of six feet at Cinchona the monthly mean soil temperature
is 16.4° C., with a mean annual range of 1.5°. For air temperatures
the annual mean is 16.0°, the annual mean range 2.9°, and the average
daily range 6.6°.

Our work was of necessity carried out within a radius sufficiently
narrow to permit of the collections being made afoot, and brought
back to the Laboratory for freezing within a few hours. Materials
were drawn from the territory made accessible by the trail from
Cinchona through Morce’s Gap to a point somewhat south of Vinegar
Hill, by that from Morce’s Gap to John Crow Peak, by that from
Cinchona to a point on one of the Green River affluents south of
New Haven Gap, and by that from Cinchona through New Haven
Gap to the lower slopes of Sir John Peter Grant Peak. Collections
were by no means limited to the immediate vicinity of the trails, but
were also drawn from the denser parts of the jungle, which was pretty
thoroughly penetrated in various directions.

While a few determinations are based upon collections made
between 5,500 and 6,000 feet, especially from New Haven Gap and
from the slopes and summit of John Crow Peak, the main bulk of our
constants are based on samples gathered between 4,500 and 5,500
feet. Below 4,500 feet conditions change rapidly. Thus at Resource,
one mile south of Cinchona and 1,300 feet lower (3,700 as compared
with 5,000 feet), the mean rainfall is about 68 as compared with
about 106 inches per annum at the Laboratory. The fogs which are
so characteristic a feature of the northern slopes of the mountains,
and which roll over the ridges from the windward sides, are dissipated
on the lower leeward (southern) slopes. Thus conditions are not
merely warmer but far drier. Here, too, much of the natural vegeta-
tion, which in most of the area studied was in a primaeval condition,
has been replaced or distinctly modified by agricultural operations—
chiefly the planting of Arabian coffee, which thrives and because of
the superiority of the product is commercially profitable in a region
so broken as to be useful for only the more valuable hand-tilled crops.

Materials and Methods.—In order that the constants of the present
study may be comparable with those derived from other regions it

272 J. ARTHUR HARRIS AND JOHN V. LAWRENCE

has seemed desirable to limit the determinations to those based on
terrestrial plants. Epiphytic forms are reserved for treatment, with
comparable forms from other regions, in a special publication.

In a habitat in which erosion is so active, epiphytes are frequently
brought to the ground by the fall of trees. Furthermore, conditions
on the litter-covered forest floor, on large fallen and partially decayed
logs, and on the higher limbs of trees, differ by only imperceptible
degrees. Thus our separation of the epiphytes from the terrestrial
forms has of necessity been somewhat arbitrary.

All of the Bromeliaceae we have omitted from the present treat-
ment.

Of the Orchidaceae we are publishing determinations for the terres-
trial Prescottia stachyoides and Stenorrhynchos speciosum. Epidendrum
verrucosum we have included since we always collected it growing in
soil on rocky banks. Fawcett and Rendle give its occurrence as
“fon trees, rocks and dry banks.” Epidendrum iwmbricatum, which
Fawcett and Rendle cite as occurring on trees and which we found
growing as a typical epiphyte, we have omitted from the present
paper. ‘The parasites have been discussed in an earlier number of this
Journal (Harris and Lawrence, 1916).

The species of the genus Peperomia have caused consideeaine
trouble. They may be either truly epiphytic, rooted in the masses
of leaf mould on fallen logs, or terrestrial in peaty soil. So far as we
were able to observe P. stellata is always terrestrial. We have there-
fore included it, but have reserved all other species of Peperomia
for a special memoir on epiphytic vegetation.

Blakea trinervia and Tradescantia multiflora, which may be either
rooted in the soil or epiphytic, have been included in this paper.

Methods.—The methods employed were those of previous papers of
this series. Considerable difficulty of a purely physical sort was
encountered in the collection of the samples. Much of the work
had to be carried out in the rain or in tangled vegetation dripping
wet from recent rain or fog. It was often necessary, therefore, for one
worker to crouch under a poncho and wipe each leaf dry with absorbent
tissue before it was placed in the collecting tubes for preliminary
freezing (Gortner and Harris, 1914). ;

The frozen tissue was squeezed with the greatest thoroughness
possible in a press with a powerful hand screw to avoid any possibility
of the differential extraction of sap as noted by Dixon and Atkins
(1913) and ourselves (Gortner, Lawrence and Harris, 1916).

OSMOTIC CONCENTRATION OF TISSUE FLUIDS 273

The freezing-point lowering of the filtered sap was determined by
means of ether or carbon bisulphide vaporized by a dried air current
in a vacuum jacketed bulb.

The results are expressed in terms of freezing-point lowering,
A, corrected for undercooling, and in atmospheres pressure P from a
published table (Harris and Gortner, 1914).

Classification of Habitats—In these studies it has been our policy
to adopt in so far as possible the classification of plant habitats drawn
up by specialists in ecology or phytogeography. Such a course makes
for simplicity and lack of confusion in the literature, lends added
value to such habitat studies as have already been made by correlating
with them new kinds of observations, and finally precludes any possi-
bility of bias in the classification of determinations in a way to make
them agree with any preconceived theory.

For the Blue Mountain region it has been possible to follow the
classification presented in the splendid work of our colleague Forrest
Shreve (1014) whose extended experience in the montane region of
Jamaica and whose analyses of the previous scattered literature and
meteorological data have made it unnecessary for us to go back of his
large publication on the region.

For descriptive details presented in a most readable manner and
a wealth of carefully selected illustrations the reader must turn to
Shreve’s book. Here only the most salient and essential points will
be set forth. 3

The fundamental division is that into the two main slopes of the
mountain chain. These are designated as windward and leeward
rather than northern and southern to emphasize the predominant
influence of the moisture-laden trade winds in determining the char-
acteristics of the vegetation. The subdivision of the two main slopes
is made on the basis of topography, into ravines, slopes and ridges.
In carrying out our work we have found it desirable to emphasize
certain of these regions at the expense of others. Such descriptive
details as are essential will be given under the discussions of the
individual habitats.

We have not found it practicable to consider individually all of
the five types of habitats recognized by Shreve.

Because of the morphologically xerophilous character of its scrub
vegetation we desired to investigate rather fully the sap properties
of the ‘‘ruinate”’ of the once cleared southern slopes. This seemed

274 J. ARTHUR HARRIS AND JOHN V. LAWRENCE

to us more important than a consideration of the primaeval forest of
the leeward slopes. As the other extreme in the vegetation of the
southern side of the ridge, the vegetation of the leeward ravine seemed
desirable.

In dealing with the collections from the windward sides of the
mountains we have not found it practicable to follow the treatment
accorded them by Shreve who discusses the ravines and the slopes
separately. The two habitats blend quite imperceptibly into each
other. The distinction between the vegetation of the two has seemed
to us to be primarily one of the loftiness of the trees and the abundance
of the extremely hygrophilous ferns, mosses, and hepatics. While an
investigation of the concentration of the sap in the bryophytes and
filmy ferns that are so characteristic a feature of the more hygrophytic
habitats would be of great interest, we preferred to devote our time
to the study of arborescent and herbaceous seed plants of the type
to be met with in other regions with which comparisons are to be drawn.

For this reason we have treated the collections from the leeward
ravines and leeward slopes together.

Our collections have, therefore, been distributed among the follow-
ing habitats. :

I. Ruinate of the Leeward Slopes:
II. Leeward Ravines.

III. Ridges.

IV. Windward Slopes and Ravines.

The distinction between these habitats is by no means always
sharply marked. Ravines and ridges are merely the extremes of the
topographic series. Between them and the intervening slopes there
is, from the purely topographical side, no sharp line of demarcation.

Furthermore, the habitat distinctions are not based primarily
upon the substratum but upon meteorological conditions. Air move-
ments undoubtedly play a considerable rédle in determining the char-
acter of the vegetation. Thus fog is often blown over the main ridge,
rolling down the leeward slopes for some distance, to be dissipated
below. ‘The vegetation of the ridges which are at the same time gaps
exhibits many of the characteristics of the ravine.

In view of these facts it is altogether improbable that any two
botanists would agree exactly upon the classification into habitats
of a series of 398 collections—the number upon which the present
discussion is based. While in some cases our disposition of a given

OSMOTIC CONCENTRATION OF TISSUE FLUIDS 275

determination may have been somewhat arbitrary, it was not in-
fluenced in any measure by the magnitude of the constant, for the
collections were all classified before the corrected freezing point lower-
ings were calculated. Thus there seems no possibility of personal
equation influencing the results.

II. PRESENTATION OF DATA

I. Rutnate of Leeward Slopes

The slopes which were once cleared for coffee or cinchona planting
but have since been abandoned—in a large part, long ago—are known
locally as ruinate.

The ruinate is characterized, as is of course to be expected, by a
relatively large number of introduced, in some cases widespread,
species.

While the ruinate has been described by writers as a xerophilous
scrub formation, it occupies an area supplied with an abundance of
precipitation, quite as much in fact as the primaeval forest of the
same slopes.

In so far as the conditions are really those of a xerophytic environ-
ment they must be due to (a) edaphic conditions influencing water
absorption, and (b) to the lowness and openness of the stand, per-
mitting free air movements with consequent increased transpiration.

The classification of this vegetation as xerophilous is due, we
believe, to two factors. First, in contrast to the extreme hygrophily
of the ravines of both leeward and windward slopes, the structurally
really mesophytic species of the ruinate have a far more xerophytic
aspect than they would if growing in a region of more moderate
humidity, just as they would pass for decidedly mesophytic types in
deserts like those of southern Arizona. Second, there are a number
‘of truly desert species which have a profound effect upon the physiog-
nomy of the vegetation. Agave is not common but Yucca aloifolia
is frequently seen. Baccharis scoparia is probably the chief form
lending a xerophytic aspect to the vegetation.

What we have just said concerning the ruinate applies to only the
areas in the neighborhood of 5,000 feet where our determinations were
made. Below this level, and especially on the southern face of the
Port Royal mountains, conditions are much drier and the truly desert
species more numerous.

276 J. ARTHUR HARRIS AND JOHN V. LAWRENCE

A habitat in which such introduced forms as Daucus Caroia,
Pastinaca sativa and Plantago lanceolata thrive, and in which occurs
a number of species common to this and one or more of the hygrophy-
tic habitats, can hardly be regarded as truly xerophytic.

The determinations from the ruinate are given in the accompanying
protocol.

Since the data are presented in a uniform way for the four habitats,
an explanation of the form of these lists may be given here.

The plants are first of all divided into ligneous and herbaceous.
Under each of these groups the species are, for convenience of reference,
arranged alphabetically. The values of A and P opposite the species
names are averages whenever more than a single determination for
the species could be secured in the habitat. In such cases the values
are designated by bars, A and P, the individual determinations upon
which these averages are based with their dates of collection are given
beneath the species name and its average constants for the habitat
in question. In cases in which only a single determination could
be secured, the values of A and P are given, with the date of collection,
in place of the average value.

LIGNEOUS PLANTS

Asclepias physocarpa (E. Meyer) Schlecht. Feb. 28, A = 0.86, P = 10.4
Baccharis scoparia (L.) Sw. A = 1.18, P = 14.2
Feb. 6, A = 1.10, P = 13.3; Feb. 18, A = 1.15,P = 13.8; Feb. 24) A> ares,

P = 15.4.
Bidens incisa (Ker.) G. Don Feb.. 7,A°= 0.01,.2 — ae
Bocconta frutescens L. A = 0.91, P = 11.0

Feb..5,,A = 0:82, P = 9.9; Feb: 28,,A°— "0.90, P= 12.0;

Borreria verticillata (L.) Meyer Feb. 5, A = 0:68, P =8.2

Caesalpinia sepiaria Roxb. A = 0.97, P = 11.7
Miat..6,A = 0:05, 2: =) 11155. Nlarv6, Ac—"0. 089 — ince

Cestrum odontospermum Jacq. Mar. 6, A = 0.99, P = 11.9

Citharexylum caudatum L. Mar. 18, A = 2.03, P = 24.4

Coffea arabica L. Mar. 6, A= 1:29, Po arses

Cracca grandiflora (Vahl.) Kuntze Feb: 25) 4 = 0.38556 — pee

Crotalaria Saltiana Andr. Feb. 8, A = 0.82, P =~ 9:9

Dodonaea jamaicensis DC. A = 1.18, P = 14.2
Feb. 5, A = 1.05, P = 12.7; Feb. 7, A = 1.09; P = 13.1; Feb. 26, A = 1.41,

P = 16.9.
Duranta repens L. A = 1.26, P = 15.2

Feb. 28, A = 1.29, P = 15.5; Feb. 28, A = 1.25, P = 15.0; Mar. 6, A = 1.25;
Dee a ;

OSMOTIC CONCENTRATION OF TISSUE FLUIDS Oe |

Echites torosa Jacq. ED i; 5) 441.085) = 1320
Eroteum theoides Sw. (Cleyera theoides (Sw.) Choisy) A = 1.14, P= 13.8
Feb. 14,A = 1.03, P= 12.4; Feb. 28, A = 1.15, P = 13.9; Mar.6,A = 1.19;

P= 14.3; Mar. 6, A-=.1.20, P = 14.5:

Young leaves were also taken Feb. 14 and gave: A = 1.27, P = 15.3.
Eupatorium glandulosum H. B. K. A = 0.82,P = 9.8
Feb. 28, A = 0.82, P = 9.8; Mar. 6, A = 0.81, P = 9.8. __ =
Eupatorium heteroclinium Griseb. KOO. i — 1 202

Heb. 5, A = 17. — A0* Heb. 7, A = 11,03; P = 124.
Eupatorium triste DC. Feb. 5, A = 1.08, P = 13.0
Garrya Fadyenit Hook. A = 2.13, P = 25.6
Heb7 A> = 11-02). -— 23:0% eb. 14, A: = 2:34, P = 2821.
Iresine paniculata (L.) Kuntze Feb. 5,-A = 0.85, P = 10.3
Lantana Camara L. A =0.81,P = 9.8
Feb. 5, A = 0.74, P = 9.0; Feb. 14, A = 0.94, P = 11.3; Mar. 6, A = 0.76,
P= 92,
Lantana reticulata Pers. Feb: 28, A = 0.76, P = 9.2
Lantana stricta Sw. A =0.73,P = 8.8
Reb. 54:4; =, 0,00, P= 8:0; Feb: 14, A.=:0.76,.P ='9.1; Mar. 6, A = 0.77,
P = 9.2.
Mecranium virgatum (Sw.) Triana Reb, 28; A. =0:71, 2 = 8:6
Miconia quadrangularis (Sw.) Naud. Mat."6,4 — 0,01, Fe — 1019
Micromeria obovata Benth. Keb, =.7, 4 —.0,76,.s—= ~~ 9.1
Oreopanax capitatum (Jacq.) Dec. & PI. A = 1.50, P = 19.1

Feb. 14, A = 1.66, P = 19.9; Feb. 28, A = 1.59, P = 19.1; Mar: 18, A = 1.53,
Pe eal OL 3%
Young leaves taken with the sample of March 18 gave: A = 1.14, P = 13.8

Passtflora edulis Sims Mar. 18, A = 1.58, P = 19.0
Phenax hirtus (Sw.) Wedd. Feb. 28, A = 0.76, P = 9.2
Psychotria corymbosa Sw. Feb. 14, A = 0.82, P = 9.9
Quercus sp. Ware 10. —1,. 80, — 9 Fae3
Rapanea ferruginea (R. & P.) Mez Feb. 14, A-= 1.18, P = 14.2
Relbunium hypocarpium (L.) Hemsl. Feb. 5, A:-=0.76,P = 9.2
Rosa laevigata Michx. Feb. 8, A.=-1.50, P: = 18.0
Smilax celastroides Kunth Feb. 14, A = 1.38, P = 16.6
Vounc leaves. gave: A = 1,14, P= 13.7; .
Triumfetta semitriloba Jacq. Feb: 8) = 0.73; P= 18.8
Vaccinium meridionale Sw. A = 1.25, P = 15.1

Heba 7A = iIet7.(P = 14.1; Feb. 18,A = 25,42 — 15. 9y Kebaoae A, = 1.34,

Y ieee) OPN
Vernonia divaricata Sw. A = 1.14; P= 13:7
Peb. la, Ac——l.20, FP 114.6) Heb. 23854 —“n 17a — TA.0-4 Mar. 6, Ac 'i.03;

P= t2-4;
Viburnum alpinum Macf. A = 1.36, P = 16.4

Deb. 5A 1.20) 2.05.5; Neb, 28,4. -142Pe—17,1-° Mar.6, A= 1.37,
P.=16.5;

279 J. ARTHUR HARRIS AND JOHN V. LAWRENCE

Yucca aloifolia L. A = 1.63, P = 16.3
Feb. 6, A:= 1.78, P = 21.4;'-Feb. 28, A-="1.45, P = 7.4; Mar.6,4 = 66>
P = 20.0.

The values given are those for the fully matured leaves. A determination
from the yellowish leaves which were past their period of maximum physio-
logical activity was taken on Feb. 6 and gave: A = 1.55, P = 18.7. In the
collection of Feb. 28 the young leaves gave A = 0.79, P = 9.5. Juice extracted
from the axis of the plants from which the collection of Feb. 6 was made gave
A = 0.96, P = 11.6.

HERBACEOUS PLANTS

Ambrosia peruviana Willd. Feb. 28, A= 1.02;)P) =
Aristea compressa Buch. Feb. 5, A= 0.72) a an
Begonia obliqua L. A = 0.36, P = 4.3

Feb. 8, A = 0.35, P = 4.2; Feb. 14; A = 0.36, P = 4.2; Feb. 2874 =036;
P-=.3,7-, Mar. 6, A= /041,P = 5.0:

Bidens pilosa L. Feb.°-7, A = 0.7407 173-0

Bryophyllum pinnatum (Lam.) Kurz. A = 0.40, P= AT,
Feb. 5, A = 0.38, P = 4.5; Mar. 6, A = 0.41, P = 4.9.

Cionosicys pomtiformis (Macf.) Griseb. Mar. 18, A = 106.32) = 1288

Daucus Carota L. A = 1.16, P = 14.0
Bebe5, A, ="1on, 2 *12it> Mar.6,;,A =—"131. 7 = 15,6:

Epidendrum verrucosum Sw. A =0.55,P = 6.6

Feb. 6, A = 0.58, P = 7.0; Mar. 6, A = 0.55, P = 6.6; Mar: 16,4) 015m,
PP 6:2;
Hedychium flavum Roxb. X Hedychium Gardnerianum Rosc. (?)

Feb. | 6;-Ac= 0:67, 225 —son8
Lycopodium clavatum L. A =0.70,P = 9.5
Feb. 18, A = 0.80, P — 9.6; Feb, 24, A —'0.73, P= 9.4. oy
Lycopodium Fawcetti Lloyd & Underw. A = 0.89, P =30:6
Feb. 18, A = 0.87, P = 10.4; Feb. 24, A = 0.90, P = 10.8.
Maurandia erubescens (Zucc.) A. Gray. A = 0.84, P = 10.1

Feb. 5, A-=.0.86, P = 10.3; Feb. 14, A = 0.80, P = 9.7; Feb. 28, A= 0.80;
P>=10:7;, Mar6,;A)=0:70,/P. =9.5;

Meibornia uncinata (Jacq.) Kuntze (?) Feb. 8A = 0.62, 2 mazes

Pastinaca sativa L. A = 1.27, P = 15.3
Feb, 14, A — 1.35, 116.2; Mar.6, A= 118) Pe ano:

Pilea grandifolia (L.) Blume (?) Feb, 28, A = 0.66, P = 8.0

Plantago lanceolata L. A = 1.12, P = 13.5

Feb.8; A{='0.95, P = 11,59 Feb. 28,4 = [32 = 13.6; Mar. 6, A = 1.29,
P = 15.5.
Verbena bonariensis L. Feb. 8, A = 0.94, P = 11.3

Il. The Leeward Ravines

The ravines, like the slopes, of the leeward side of the mountains
receive a lighter rainfall, much less fog, and reciprocally more hours
of sunshine, than the windward habitats.

OSMOTIC CONCENTRATION. OF TISSUE FLUIDS 279

The ravines of the leeward slopes are physiographically similar
to those of the windward slopes. Both exhibit a forest covering of
irregular canopy of larger trees with rich undergrowth of shrubs.
The conspicuous difference between the two is chiefly found in the
relative scarcity of epiphytes, both Orchidaceae and Bromeliaceae,
and particularly of the most hygrophilous of the pteridophytes and
the practical absence of tree ferns in the leeward ravines.

LIGNEOUS PLANTS a
Acalypha virgata L. A = 0.87, P = 10.5
Feb. 11, A. = 0.78, P= 9.4; Mar. 11, A = 0.85, P = 10.2; Mar. 18, A = 0.98,
P = 11.8. ‘
Acnistus arborescens (L.) Schlecht. Wake lle —n0. 85. —= 6.2
Besleria lutea L. A =0.74,P = 8.8
Feb. 11, A = 0.65, P = 7.8; Feb. 26, A = 0.60,.P = 8.3; Mar. 11, A = 0.85,
P = 10.2; Mar. 18, A = 0.75, P = 9.0.
Bocconta frutescens L. A =0.79,P = 9.5
Feb. it, A = 0:75; 2 = 9.0; Mar. 11, A-= 0.83, P = 10.0.
Young leaves taken in the collection of Feb. 11 gave values only slightly
lower than those from mature organs, i. e., A = 0.72, P = 8.6.
Boehmeria caudata Sw. Mar. 11, A = 0.86, P = 10.3

Brunfelsia jamaicensis Griseb. A =0.79,P = 9.4
Feb. 11, A = 0.78, 7.= 9.3; Mar. 18, A = 0.79, P= 9.5. __ we
Cestrum hirtum Sw. A =0.73,P = 8.8

Hebwll.Av—\O74,.2, — 8.03,.Vlar. 18,A = 0.72, P. = 8:7.

Cinchona Sp. 0.92, P = 11.1

|
I

Rebs Try Al — 0107, = 11.6; Mar. 18, A =-0.87,.P = 10.5;
Chibadium terebinthinaceum (Sw.) DC. A = 0.82,P = 9.9
PepalinvAt—.0:72; 42 — 8.0; Feb. 26,4 =-0,80, P = 10.777 lar. 11, A-=:0.86,
PI" TOA,
Dendropanax arboreum (L.) Dec. & PI. A =1.11, P = 13.3
Heber — tle 13.3. Mar, 18) A = 1.10, = 1373:
Duranta repens L. A = 1.33, P = 16.0
Mebmidtahy— 1200: — 14-4 Mat. g8,;A\— 1.45, 2o— 27.5.
Eupatorium glandulosum H. B. K. A = 0.64,P = 7.7
Nebw26,,4) —' 0:04, P — 7.7>\ Mars rr, A = 0:64, P= 7.6.
Eupatorium riparium Regel A =0.78, P = 9.3
Mar. iA. = 0.77, .P = 9.2;) Mar. 18; A= .0.78, P= 9.4.
Fuchsia corymbiflora R. & P. A =0.67,P = 81
Mat. 1f, A= 10:73, 2 =-8.8; Mar: 11,4 = 0:70, P= 8.5; Mar.18, A = 0.57,
P= 6,0:
Garrya Fadyenit Hook. Reb 265A —'2) 29 Pi 26.8
Gesneria alpina Urban Mar..11,.Av—-0252,,.P =) 26.3
Guarea Swartzit DC. A = 0.90, P = 10.9

Mar. it; A= 0.79, = 6:6; Mar? 18) A5— 1,02,-P) — 12:3- Mar, 18; A = 0.90,
P = 10.8.

280 J. ARTHUR HARRIS AND JOHN V. LAWRENCE

Hedyosmum nutans Sw. Feb: 26, A= 0.73; P=778is
Iresine paniculata (L.) Kuntze : A = 0.96, P = 11.6
Feb. 11, A = 0.89, P = 10.8; Feb. 26, A = 0.97, P = 11.7; Mar. 11, A = 1.03,
P i= 12rd,
Lantana Camara L. A =0.69,P = 8.2
Feb. 11, A = 0.64, P = 7.6; Feb. 26, A = 0.73, P = 8.8; Mar. 11, A = 0.68,
P = 8.2; Mar. 18,.A = 0.60, P = 8.3. °

Phenax hirtus (Sw.) Wedd. A= 0.77,P = 9.2
Feb, 26, A. = 0.74, P =:8.9; Mar. 11, A -=-0.70,, P= 9.5.

Phoebe montana (Sw.) Griseb. Feb. 117A = 1.02032 ore

Pilea Weddellii Fawc. & Rendle Feb; 26, A. ="0.7.,- bans

Piper hispidum Sw. Mar..18, A = (0,67,.22—ames0

Psychotria corymbosa Sw. A =0.68,P = 8.1
Feb. 11, A = 0.66, P = 7.9; Feb. 26, A = 0.69, P = 8.3.

Rapanea ferruginea (R. & P.) Mez Feb. 26, A°= "1,10, et 2
Young leaves gave: A = 1.09, P = 13.1.

Rubus jamaicensis Sw. Mar; 18)-A = 1.330ee— 160

Senecio Swartzit DC. Mar. 18, A = 0.66, P = 7.9

Solandra grandiflora Sw. A = 0.82,P = 9.8
Mar. 11, A = 0.79, P = 9.5; Mar. 18, A = 0.84, P = 10.1:

Tovaria pendula R. & P. Mar. 11,'A' = 0:94, P= 113

Turpinia occidentalis (Sw.) G. Don Mar..18; A = 10042 7— a2

Viburnum villosum Sw. Mar, 18,°A = Tis@P. = 1210

HERBACEOUS PLANTS

Anthurium scandens (Aubl.) Engler A = 0.52, P = 6.3
Feb. 11, A = 0.51, P = 6.1; Mar. 11, A = 0.53, P = 6.4.

Begonia obliqua L. A =0.37,P = 4.5
Feb. 26,.A.=(0,35,-P = 4.2; Mar. 11, A =.0,30,,.P = 4.7:

Cionosicys pomtiformis (Macf.) Griseb. A = 0.69, P = 8.4

Mar. 11, A = 0.66, P = 8.0; Mar. 11, A = 0.66, P = 8.0; Mar. 18, A = 0:76,
POs,

Elaphoglossum latifolium (Sw.) J. Sm. Feb. 15 A = 0.78, P=" 974

Epidendrum verrucosum Sw. A= "0.51, 2. = et
Mar. 141, Av—.0.50), 7 —' 6-057 Matr..16, 74. == 075 bel ore

Liabum umbellatum (L.) Sch. Bip. A= 0.67,P = 8.1
Mar. 11;:A°=90:67, P = 8.0? Mar.18,,A°=— 0.67307. —"8. 1.

Maurandia erubescens (Zucc.) A. Gray A = 0.80, P = 9.7
Feb, 26, A = 0.75, P = 9.0; Mar. 11,.A = 0.80, P = 9.6; Mar. 18, A = 0.86,

P = 10.4.
Pastinaca sativa L. Mar, 18, A:= 2.16, 2-= Ako
Peperomia stellata (Sw.) A. Dietr. A = 0.43, P = 5.2

Feb. 11, A = 0.42, P = 5.0; Feb. 26, A = 0.40, P = 4.8; Mar. 11, A = 0.41,
P = 5.0; Mar. 18, A = 0.50, P = 6.0.
Pilea grandifolia (L.) Blume A = 0.63, P = 7.6
Feb. 11, A = 0.60, P = 7.3; Feb. 26, A = 0.70, P. = 8.4; Feb. 26, A = 0.58,
P= 7.0.

OSMOTIC CONCENTRATION OF TISSUE FLUIDS 281

Senites Zeugites (L.) Nash Feb. 26, A = 0.69,P = 8.4
Stenorrhynchus speciosus (Jacq.) L.C. Rich. Ne 04529 Jou — 0,3
Feb. 11, A = 0.54, P = 6.5; Mar. 11, A = 0.48, P = 5.8; Mar. 18, A = 0.54,

P = 6.5.
Tradescantia multiflora Sw. Ui Maret, Ae 0.20) 0h = 09 Ac7

III. The Ridge Forest

The ridge forest, closely confined to the main ridge of the mountains
and to narrow strips along the crests of the water divides on both
windward and leeward slopes, is far more open than that of the
slopes or ravines. The vegetation is, therefore, not only more exposed
to the influence of light, but is much more wind swept than that of
the more deeply and densely covered slopes and ravines. This habitat
is, therefore, ‘‘relatively xerophilous in the entire make up of its
vegetation.”

The following are the results:

LIGNEOUS PLANTS

Acalypha virgata L. A = 0.92, P = 11.1
epwon AG 10:66" F).—"10:4; Var 9, Ai 0.96, P — 11,62 Marat, A — 0195,
PS 11.4:
Acnistus arborescens (L.) Schlecht. Marsct6;, A) =-0.97; P= "1147
Actinophyllum Sciadophyllum (Sw.) R. C. Schneider Mar. 9, A = 1.24, P = 15.0
~ Alchornea latifolia Sw. Mar. 4, A = 0.89, P = 10.8
Brunfelsia jamaicensis Griseb. A = 0.83, P = 10.0
Mar. 9, A = 0.80, P = 9.6; Mar. 9, A = 0.90, P = 10.8; Mar. 16, A = 0.80,
P= 9.6.
Cestrum hirtum Sw. Wars toms 10.538 = 10,0
Cinchona sp. A = 1.02, P = 12.3
Hepes 20; At— 1:00, 7. = 12.7 Mar 4, A — 0.9o0;7, — 11:6. Mar. 16, A.—-1405,
Pee —Bl226,
Citharexylum caudatum L. A = 1.92, P = 23.1
Feb. 9, A = 1.95, P = 23.4; Mar. 9, A = 2.05, P = 24.6; Mar. 9, A = 1.77,
P28. 5;
Clethra occidentalis (L.) Steud. A =0.73,P = 8.8
Pebio, A= 10.77, P = 9:3; Mar. 16,,A-='0:687 = 8.2:
Clusia havetioides (Griseb.) Tr. & PI. Feb. 13;/A.=:0:70/.P 5=)-9.5
Cyrilia racemiflora L. Feb. 17, A:= 1.18; PP’ ="14.2

To avoid increasing unduly the number of habitats this determination based
on material from John Crow Peak has been included in the Ridge Series.
Dendropanax sp. Feb. 10, Ay —. 1-00; 2 =" 12:0
Dendropanax nutans (Sw.) Dec. & PI. A = 0.93, P = 11.2
Mar. 9, A = 0.98, P = 11.8; Mar. 16, A = 0.87, P = 10.5.

282 J. ARTHUR HA”RIS AND JOHN V. LAWRENCE

Eugenia virgultosa (Sw.) DC. (?) Feb.. 9, A = 0.72) 2 ae-7

Eupatorium glandulosum H.B.K. A=0.72,P = 8.7
Feb. 18, A = 0.76, P ='9.2; Feb. 20,.A ‘= 0.71, P = 8.5; Maro 34. —somo-

P= 8.4,

Eupatorium parviflorum Sw. Mar. 9, A = 0.85, P = 10.2

Eupatorium triste DC. A = 1.24, P = 14.9
Mar. 9, A°=:1.26;.P=*15.1; Mar. 16, A = 1.21, P ="14.6;

Gesneria alpina Urban A = 0.58, P = 7.0
Rebio; As= 0,00; =37.2- VianaQ Ne" 0 56 0ae olor

Guarea Swarizit DC. Mar. 9, A = 1.07, P = 42:8

Gymnanthes elliptica Sw. Mar.-16, A ="5:00; 2 = 12,6

Hedyosmum arborescens Sw. A = 0.73,P = 8.8
Mar. 16; Ay 0.73)" P = 8.83) Mar. 16,(A—0,73..P —2350.

Mecranium purpurascens (Sw.) Triana A =0.77,P = 9.3
Mar. 4, A = 0.77, P = 9.3; Mar. 4, A = 0.77, P.= 9.2.

Meitenia globosa (Sw.) Griseb. Mar. 16, A = 0:87, 2. — 10.5

Miconia quadrangularis (Sw.) Naud. A = 1.00, P = 12.1

Feb. 9, A = 0.87, P = 10.5; Feb. 20, A = 0.94, P = 11.3; Mar. 4, A = f.05,
P = 12.7; Mar. 9, A = 1.11, P = 13.4; Mar. 16, A = 0.08; P = 31-8 Mar
16, A -= 1.07,5P =-12.0:

_ Miconia theaezans (Bonpl.) Cogn. - A = 0.88, P = 10.6
Feb..9, A.= 0.84, P = 10.1; Feb. 11, A\= 0.84, P = 10.1; Marom. A 0.67
Jeg Fa ree 7

Myroxylon nitidum (Hell.) Kuntze A = 1.35, P = 16.2

Mar. 9, A = 1.40,P = 16.8; Mar.16,A = 1.14, P = 13.7; Mar. 16,A = i151,
P = 18.1.

Ocotea jamaicensis Mez (?) Mar. 4, A = 1.08, P = 13.0

Palicourea alpina (Sw.) DC. A =0.69,P = 8.3
Feb. 18, A = 0.55, P= 6.6: Mar. 16, A = 0.83, P = 10.0.

Pilea Weddellit Fawc. & Rendle Mar: 9, A =10:67;ePs— ene

Psychotria corymbosa Sw. A =0.76,P = 9.1

Feb. 9, A = 0.70, P =.8.4; Mar. 4, A = 0.75, P = 9.1;. Mar: 16; A = 0182,
P = 9.9.

Psychotria Harrisiana Urban Mar: 16, A:-='0.83, P)— 10.6
Rapanea ferruginea (R. & .P.) Mez A = 1.02, P = 12.3
Feb. 9, A = 0.96, P = 11.6; Mar. 16, A = 1.07, P = 12.9.
In the collection of Feb. 9, young leaves gave: A = 0.89, P = 10.7.
Rhododendron (cultivated) A = 1.04, P = 12.6
Feb. 20, A = 1.01, P.= 12.2; Feb. 20,,A = 1.07, P= 12:0:
Solanum punctulatum Dunal. Mar. 13, A = 1.21, P = 14.5
Vaccinium meridionale Sw. Mar. 16, A = 1.32, P = 15.9
Young leaves gave: A = 1.18, P = 14.2.
Wallenia calypirata Urban A = 0.84, P =10.1

Feb. 9; A’= 0:77, .P = 0:3; IMare16,°\) —- 0.01, 7. — 10.0:
Young leaves were also taken on Feb. 9 and gave: A = 0.70, P = 8.5.

OSMOTIC CONCENTRATION OF TISSUE FLUIDS 283

i
HERBACEOUS PLANiS

Anthurium scandens (Aubl.) Engler A = 0.63, P = 7.5
Mar-9, A = 0.61, P = 7.3; Mar. 16, A = 0.64, P = 7.6.

Begonia obliqua L. Mar.’ 9; A = 0:22; 2. = 4.0

Fragaria insularis Rydb. Mar, 9,0 11.15, 22 113.9

Liabum umbellatum (L.) Sch. Bip. A=0.71,P = 8.5
Mar. 9, A = 0.60;-7 = 8:3;. Mar: 16, A. =-0:72,.P = 8.7. |

Peperomia stellata (Sw.) A. Dietr. Wats O,-Ai—= 0:45, 2 5.4

Pilea grandifolia (L.) Blume A= 0.64,P = 7.7

Feb; 9, A.—0:61,.P = 7.3% Feb. 18, A = 0:63, P = 7.6; Mar. 9;-A.= 0.67,
P=8.1.

Plantago lanceolata L. Feb. 24; Av= 1a5,P. = 13.8

Senites Zeugites (L.) Nash Mags “0, ="0.68;-P =< 8:2

IV. Windward Ravines and Slopes

The windward slopes and ravines, exposed as they are to the
direct influence of the moisture-laden trade winds, exhibit in the
highest degree the features of climate and vegetation which find their
simplest expression in the term Rain Forest. The mere statement of
the rainfall in inches per year conveys no adequate impression of the
actual environment to which the species constituting this vegetation
are exposed. The roots of the plants are not merely supplied with
water by the heavy and well-distributed rainfall, much of which is
stored for long periods in the litter of the forest floor, but the foliage
is for much of the time immersed in the floating fog. Thus insolation
is much reduced. Even at times when rain is not falling and when
the plants are not enveloped in fog, high atmospheric moisture is
maintained for long periods of time by evaporation from the litter
on the ground and from the moist foliage. Here are large trees with
trunks and branches burdened with thin-leaved, succulent-leaved and
tank epiphytes, with mats of hepatics and garlands of mosses and
filmy ferns, shading a nearly bare forest floor or in other places over-
topping a tangled shrubby and herbaceous undergrowth. Any ade-
quate description of this forest would not only outrun the space here
available but in view of Shreve’s carefully penned description and
well chosen and admirably executed plates is quite superfluous. One
feature plates cannot depict. This is the reeking wetness of the
foliage. This can only be fully appreciated by one who has had the
aesthetic pleasure and the physical discomfort of collecting in these
forests during or immediately subsequent to the gentle rains, which drip
from the glossy foliage, percolate through thé sponge-like beds of

284 J. ARTHUR HARRIS AND JOHN V. LAWRENCE

mosses and hepatics and replenish the tank leaves of the bromeliads,
if they are not already overflowing, or in the fog which rolls like clouds
of smoke among the trees, covering the leaves like dew.

LIGNEOUS PLANTS

Actinophyllum Sciadophyllum (Sw.) R. C. Schneider A = 1.30, P = 15.7
Feb..4, A = 1.12, P= 13.5, Mar. 2, A = 1.48) P= 172650 a
Besleria lutea L. A = 0.58, P = 7.0

Feb. 4, A =.0.48, P = 5.7; Feb. 4, A-= 0.60, P = 7.2; Feb. 13):475—"0152;
P = 6.3; Feb. 22, A = 0.59, P = 7.1; Mar. 2, A = 0.64, P = 7:75 Mar. 12,
A = 0.67, P = 8.0.
Blakea trinervia L. A =0.58,P = 6.9
Feb, 13, A = 0.42, P = 5.0; Mar.-2, A = 0.55,.P = 6.7;. Mar. 2, A = 10%G7.
P= 8.1; Mar: 13, A= 0:66, P= 7.0:

Cestrum hirtum Sw. Mar. 13, A = 0.72, P =~ 8.7
Clibadium terebinthinaceum (Sw.) DC. A =0.60,P = 7.3
Feb. 13, A = 0.48, P = 5.8; Feb. 22, A = 0.65, P = 7.9; Mar. 3, A = 0.67,
Pi Sal,
Clusia havetioides (Griseb.) Tr. & Pl. A= 0.74, P= 89
Feb..20, A ='0.76, -P -— 0.13" Keb. 20,4 = 0:72) Pi—" 6;
Cyathea furfuracea Baker A = 0.78, P= 9.5
Feb. 24, A = 0.81, P = 9.8; Feb. 24, A = 0.76, P = 9.2.
Datura suaveolens H. & B. Feb. 13, A = 0.47257
Dendropanax nutans (Sw.) Dec. & PI. Mar. 2,A = 106, — "1225
Young leaves from the same tree gave slightly lower values: A = 0.91,
P = 10.9,
Eupatorium glandulosum H. B. K. A =0.64,P = 7.7
Feb. 22, A = 0.54, P = 6.5; Mar. 2, A = 0.62, P = 7.4; Mar. 4, A = 0.76;
P=o9.1.
Eupatorium parviflorum Sw. Mar. 2, A = 1.07, 2 —2@2e.
Eupatorium riparium Regel A = 0.58, P= 7
Mar. 2, A = 0.55, P = 6.7; Mar. 13, A = 0.61, P = 7.4.

Gesneria alpina Urban A =0.51,P = 6.2
Mar. 2,.A = "0-50,,P'=:6.1: Mar. 13, A:="0.52,-P— 6.2:
Guarea Swartzit DC. A = 0.83, P = 10.0

Feb. 13, A = 0:73, P = 8:8; Feb. 22;A = 0.91,.P = 11:0; Mare13)A.—40-64-
P= "TOsd:
Hedyosmum arborescens Sw. A =0.65,P = 7.9

Feb. 4, A = 0.56, P = 6.8; Feb. 4, A = 0.49, P = 5:9; Feb. (113, A= o>
P = 6.7; Feb. 20, A = 0.64, P = 7.7; Mar. 2, A = 0.66, P = 8.0; Mar. 2,
A = 0.70, P = 8.4; Mar. 4, A = 0.65, P = 7.8; Mar. 13, A = 0.82, P =9.9;
Mar. 13, A ='0:82) P = oji9;
Marcgravia Browne (Tr. & Pl.) Krug. & Urban. A = 0.78, P = 9.4
Feb. 4, A-= 0.62, P = 7.5; Feb. 13, A = 0.65, PB = 7.8; Feb. 22, A oraz
P = 8.1; Mar. 2, A = 0:85, P = 10.2; Mar. 4, A. = 0.87, P =)10.5; Nianwaae,
K=O. O24 fs — aloe,

OSMOTIC CONCENTRATION OF TISSUE FLUIDS 285

The foregoing determinations on which the average for the species is based
are from the leaves of the aerial branches, extending from the trunks. Two
determinations on the ‘‘juvenile’’ leaves of the creeping stems were secured.

These are:
Keb. 13, A = 0.53, P = 6.4; Mar. 4, A — 0.71, P = 8.5. » A
Meriania purpurea Sw. A = 0.87, P = 10.5

Feb. 22; A.='0.77, P = 9.3; Mar. 2, A = 0.90, P = 10.8; Mar. 13, A = 0.95,

P= T1.5.
Miconta quadrangularis (Sw.) Naud. Nate 13) A. —-0.07, = 11.7
Miconia theaezans (Bonpl.) Cogn. A = 0.90, P = 10.8
Peb.4 A-=-0:76; P= '9.1;° Mar. 13,4 = 0.97, P= 11.7;- Mar. 13, A-= 0.96,

P = 11.5.

Palicourea alpina (Sw.) DC. A = 0.69,P = 8.3
Feb. 13, A = 0.52, P’= 6.3; Feb. 24, A = 0.78, P = 9.3; Mar. 2, A = 0.63,

P = 7.6; Mar. 4, A = 0.75, P = 9.0; Mar. 13,4 =0.77,P =9.2. _
Pilea Weddell Fawc. & Rendle A =0.62,P = 7.4

Nebw22, A= 0:57, PF) =-6:8; Feb. 24, A = 0.67, P = 8.1; Mar. 4, A = 0.61,
P= 733. oe =
Piper hispidum Sw. A=0.50,P = 6.1
Pebei3, A — 0.43, 7 = 5.2; Feb..22, A = 0.45, P = 5.4; Mar. 13, A = 0.62,
P= 7.6.

Podocarpus Urbanzi Pilger eb. 24,410.02, P= 11.2
Young leaves gave: A = 0.81, P = 9.7. io _
Psychotria corymbosa Sw. A = 0.76, P = 9.2

ebe20, A’= 0.75; P = 9.0; Mar. 4, A = 0.67, P = 8.1; Mar. 13, A =-0.86,
P = 10.4.

Schradera tnvolucrata (Sw.) Schum. Mar.13,,A%=—"i.24, P = 15.0
Solanum punctulatum Dunal Mat 13,-A = \'.04,- P = 12.8
Tovaria pendula R. & P. A= 0.70, P = 8.5
blepsl 3.4 Ne— 10108, 2-68.25) heb. 22, A-— 0,72, P= 8.7.
Vaccinium meridionale Sw. A = 1.33, P = 16.1
Menor AC e-a0.ct. Pot eS. Mar, 134, A —91.36, P= 16.3...
HERBACEOUS PLANTS
Anthurium scandens (Aubl.) Engler A = 0.52, P = 6.3

HeDs 13 Ai 0,50, P= 6.0; Mat. 4, A:= 0.52, P = 6.2? Mar. 13, A= 0:55,
== O20;

Begonia glabra Aubl. A = 0.30, P = a55
ebet «A )—10.20)b.— 3:4*- Keb: 13, A =.0,30/ P= 3.6,
Begonia obliqua L. A= One. P= 3.9

Peb20, AW —70:31,.P = 3.7; Feb. 24, A = 0.330 = 4.0; Mar.2, A-— 0:35,
i Ae Nate, AG .0.32,.1° = 3:8: Mar, 113, 4 —=.0:35, Pale Mar: 13,
M0131, F207.

Elaphoglossum chartaceum Baker Mar. 13047 —= 6,960). = 11.5
Fragaria insularis Rydb. MarGATa AL 31,00, 72) = 13.1
Gesneria mimuloides (Griseb.) Urban A=0.44,P = 5.2

War2, Al 0.A2, neo 50.8 Var 4 Ae O45, P= Sia

286 J. ARTHUR HARRIS AND JOHN V. LAWRENCE

Liabum umbellatum (L.) Sch. Bip. A =0.58, P = 7.0
Mar. 2, A = 0.53, P = 6.4; Mar. 4, A = 0.60, P = 7.2;> Mar. 13,4 = 062.
P = 7.5. a oe
Lobelia assurgens L. A =0.73,P = 8.7
Feb. 13, A = 0.66, P = 8.0;. Mar. 2, A = 0:76, P = 9.1; Mar. 2, A = 0176)
P= Oo

Panicum palmifolium Poir. A = 0.80, P = 9.6
Feb. 4, A = 0.76, .P = 9.2; Feb. 22, A = 0.80, P = 9.7; Mar: 2, A — 0:82,

P = 10.0.
Peperomia stellata (Sw.) A. Dietr. A = 0.42, P = 5.1

Mar. 2, A = 0.42, P = 5.1; Mar. 4, A = 0.41, P = 4.9; Mar. 13, A = 0.43,
P= 5.2:
Pilea Mar. 4
Pilea grandifolia (L.) Blume
Feb. 13, A = 0.57, P = 6.8; Feb. 22, A = 0.59, P =
Pilea nigrescens Urban A =0.57,P = 6.9
Feb. 20, A = 0.56, P = 6.7; Feb. 22, A= 0.51, P = 6.1; Feb, 2404) — 6.55.
P = 6.6; Mar. 2, A = 0.58, P = 6.9; Mar. 4, A = 0.61, P = 7.4; Mar. 13;
A = 0.61, P = 7.4.
Prescottia stachyodes (Sw.) Lindl. A = 0.81, P = 9.7
> Feb. 24, A = 0.84, P = 10.1; Mar. 4, A = 0.72, P = 8.6; Mar, 13,.A5—10387-
= 10.5.
Senites Zeugites (L.) Nash A = 0.62, P = 7A
Feb. 24, A = 0.59, P = 7.1; Mar. 4, A = 0.64, P = 7.8.

|
ee
=

III. DiscussIon OF RESULTS

In analyzing these data we shall consider three main problems:

A. The relationship between growth form and osmotic concentra-
tion.

B. The differentiation of the habitats of the Blue Mountains in
osmotic concentration.

C. The relative value of the osmotic concentration of the fluids of
the plants of the Blue Mountain rain forest as compared with other
phytogeographically different areas which have been investigated by
similar methods.

The only method by which these problems may be investigated is
the statistical one, the comparison by means of averages of different
sections of the data. |

The averages of species means (or of species determinations, when
only one for a habitat is available) are given for herbaceous and
ligneous plants separately, and for all plants, for each of the four
habitats in Table I.

OSMOTIC CONCENTRATION OF TISSUE FLUIDS

Fundamental Averages for Blue Mountain Rain Forest

TABLE [

287

Habitats and Constants

I. Ruinate of leeward slopes:
Freezing-point lowering
Osmotic concentration

II. Leeward ravines:
Freezing-point lowering
Osmotic concentration

III. The ridge forest:
Freezing-point lowering
Osmotic concentration

IV. Windward ravines and slopes:
Freezing-point lowering
Osmotic concentration

I-1V. All species:
Freezing-point lowering
Osmotic concentration

oe 8 © © eo

ie we ee

ee ee ee of

3, 1¢! Ge) her enh

Ligneous Plants

Number

40

136

Ligneous and
Herbaceous Plants

Herbaceous Plants

Mean oe Mean

£089) |) 17 .812
13.05 9-77

.QOI 13 .628
10.83 7.59

.958 8 Lo
11.54 8.63

“805 | 15. | ..627
9-73 7:52

952) 53 -700
11.44 8.80

Number

ae

45

44

43

189

Mean
1.007
12.07

1922
9.89

914
ete Or

743
8.96

881

10.59

These are the fundamental constants upon which much of the
following discussion must be based.
Comparison of Ligneous and Herbaceous Growth Forms.—The justi-
fication for the division of the determinations into those for herbaceous

and those for ligneous plants is clearly brought out by Table I.

For

each habitat studied the freezing point lowering is on the average
lower for the herbaceous than for the ligneous plants.
differences in terms of atmospheres are given in Table IT.

TABLE [I

The actual

Comparison of Osmotic Concentration of Herbaceous and Ligneous Growth Forms

Ruinate of : Windward
Growth Form the Leeward Leeward Ridge Slopes and All

Slopes Ravines Forest Ravince Habitats
EMUBS PE CICS Peay FN. Sine ie ase bas 12.07 9.89 11.01 8.96 10.59
BigMCOUS SPECIES. Is ss: 13.05 10.83 11.54 9.73 11.44
Herbaceous species............ 9:77 7.59 8.63 7.52 8.80
POET OMG CDR her ie be. 0 B28 R.2k 2.91 221 2.64
Percentage difference.......... 25.03 29.92 25.22 22071 23.08

Thus the difference in the concentration of the sap of ligneous and
herbaceous plants is from about 23 to about 30 percent of the higher

value, that for ligneous forms.

Comparison of Habitats in the Blue Mountain Region.—Turning

288 J. ARTHUR HARRIS AND JOHN V. LAWRENCE

now to the comparisons of the local habitats among themselves we
note the following points which must be taken into consideration in
the analysis of the data.

The comparisons between the windward and the leeward exposures
on the basis of the now available data may be expected to give a
minimum rather than a maximum measure of the differences between
them. This is true for three reasons. First, we have made the com-
parison between the plants of the windward slopes and windward
ravines taken together and two of the sub-habitats of the leeward slopes.
Thus if there be measurable differences between the sap properties
of the windward ravines and the windward slopes, the combination
of the two will tend to minimize the differences which might have been
obtained had it been practicable to deal separately with the properties
of the saps of the windward slopes and ravines. Second, we have
arbitrarily excluded a great number of forms which are apparently
the most hygrophilous and are possibly characterized by an even
lower osmotic concentration than are the species for which determina-
tions are given in these pages. Had it been possible to free the mats
or festoons of certain of the cryptogamic epiphytes from the super-
ficial water with which they are so constantly saturated, without
modifying the concentration of their tissue solutions by drying, we
believe that a series of determinations falling almost if not entirely
in the lower range of variation in osmotic concentration as shown by
the available determinations might have been obtained. Third, to
render the results from the Blue Mountain habitats as nearly as
possible comparable with others which have been or are being in-
vestigated we have excluded the Bromeliaceae, the Orchidaceae, with
the exception of truly terrestrial forms, and some other phanerogamic
epiphytes. There is, as far as we are aware, no @ priori reason to con-
sider that these forms would be characterized by low osmotic con-
centrations. While the detailed discussion of these ecologically most
interesting forms is reserved for a comparative study to be published
later, it may be said in passing that the concentration of these forms
has been found to be usually far lower than that of other species of
the vegetation.

These facts while they must detract somewhat from our constants
as an exact description of the region in question, make differences
secured under these limitations much more significant.

In considering differences in sap concentration in relation to local

OSMOTIC CONCENTRATION OF TISSUE FLUIDS 289

habitats in the rain-forest region the comparison of each of the four
habitats with the three others may be made in detail in a series of
four tables.

In view of the differentiation between herbaceous and ligneous
plants demonstrated above, the comparison must first be made for
each class separately.

In these tables each of which is devoted to the values showing
the absolute and relative magnitudes of the constants of a given
habitat, the comparisons are made in two ways. First, the actual
differences in mean osmotic concentration, P, between any habitat
and the three other habitats have been determined. ‘These are the
values with signs. Second, the ratio of the mean osmotic concentra-
tion of the sap of every habitat to that of each other habitat with
which it is to be compared has been determined. ‘These are the values
given in black-faced type.

The first method has the obvious advantage that differences are
expressed in the concrete terms of osmotic concentration. Relative
values, as employed in the second method, are on the other hand
more convenient for comparison. The exact method of drawing the
comparisons will be clear from an explanation of the individual
tables.

The first column of Table III, in which the values obtained in the

TABLE III
Ruinate of the Leeward Slopes Compared with Other Habitats
Ruinate of the | 7
Growth Form 1 | Windward |
Tesward Stopes Leeward Ridge Forest Slopes and All Habitats
Ravines Beavines
Herbaceous... . O77 +-2.18 +1.14 | +2.25 70:07
1.00 1.29 143 1.30 | Trl
igneous... .... . 13.05 +2.22 +1.51 +3.32 | =-+1.61
1 I.00 | 1.20 T.13 | 1.34. | I.14
All species..... 12.07 +2.18 | +1.06 +3.11 | +1.48
a 1.00 _ mpi aN 1.22 es I.10 1.35 I.14

ruinate of the leeward slopes are compared with those of each of the
other habitats, gives the growth forms on which the comparisons are
based. It has been practicable to recognize only two of these in the
Blue Mountain region, the herbs and arborescent, frutescent and
suffrutescent plants. The second column gives the actual mean
values in atmospheres of the plants of the ruinate. The third to

290 J. ARTHUR HARRIS AND JOHN V. LAWRENCE

sixth columns contain the actual differences between the mean values
for the plants of the ruinate and of the three other habitats and of the
region as a whole. These are obtained by subtracting the values for
each of the habitats compared from the values for the ruinate as
given in the second column. The same method is followed in drawing
up the three other comparison tables to be discussed below.

The signs of the differences are positive throughout. Thus the
concentrations prevailing in the plants of the ruinate, which has been
recognized by Shreve and others as the most xerophilous of the Blue
Mountain habitats, are higher for both herbaceous and ligneous
plants and for all species of plants than those in any other habitat.
They are over two atmospheres higher than those found in the plants
of the neighboring leeward ravines, over one atmosphere higher than
those of the ridge forest and from over two to more than three atmos-
pheres higher than those demonstrated on the windward side of
the range.

The relative values, obtained by dividing the mean concentration
of the plants of the ruinate by those of each of the other habitats,
show that the concentration of the sap of the plants of the most
xerophytic of the habitats is from about 20 to 30 percent more con-
centrated than that of the leeward ravines, about 10-13 percent more
concentrated than that of the ridge forest, and from 30 to 35 percent
more concentrated than that of the plants of the windward habitats.

Table IV, giving the relationship between the sap properties of

TABLE IV
Ridge Forest Compared with Other Habitats
Growth Form Ridge Forest : Windward
Ruinate of Leeward Slopes All Habitats
Leeward Slopes Ravines and Ravines
Herbaceous.... 8.63 —I1.14 +1.04 +1.11 —0.17
1.00 0.88 1.14 I.15 0.98
Ligneous....... 11.54 —I1.51 +0.71 ~-+-1.81 -+0.10
1.00 0.88 1.15 1.19 I.OI
All species..... 11.01 —1.06 +1.12 +2.05 +0.42
1.00 0.91 Trt 1.23 1.04

the plants of the ridge forest and those of the other habitats, shows
that the plants of this habitat have a concentration lower than that
of the comparable growth forms of the ruinate but higher than that
of either the leeward ravines or the windward ravines and slopes. The
amount of the difference is as great as 2 atmospheres in one case only.

OSMOTIC CONCENTRATION OF TISSUE FLUIDS 291

The relative differences are not large. In only a single comparison
does the ratio indicate a difference of as much as 23 percent.

Table V shows that the concentration in the plants of the leeward
ravines is lower than in those of the ruinate or of the ridge forest

TABLE V
| Leeward Ravines Compared with Other Habitats
Leeward
Growth Form : : Windward
ane Ruinate of Ridge Forest Slopes All Habitats
Leeward Slopes at deRacines
Herbaceous.... 7.59 —2.18 —1.04 +0.07 —1.21
1.00 0.78 0.88 1.01 0.86
MASHEOUS. ..4 cs 10.83 —2.22 —0.71 +1.10 —0.61
1.00 0.83 0.94 T.1L 0.95
All species..... 9.89 —2.18 —I.12 +0.93 —0.70
1.00 (| 0.82 0.90 1.10 0.93

but higher than that of the windward slopes and ravines. The dif-
ferences between the concentrations in the leeward ravines and on the
ridges on the one hand and between the leeward ravines and the
windward ravines on the other are not large.

The final comparison is that of the windward ravines and slopes
with the other habitats. This is made in Table VI. The differences
show that the plants of the most hygrophytic habitat of the region

TABLE VI
Windward Windward Slopes and Ravines Compared with Other Habitats
Growth Form Slopes and ;
Ravinge Lessard Stones mee Ridge Forest All Habitats
Herbaceous.... 7.52 —2.25 —0.07 —I.11 —1.28
1.00 0.77 0.99 0.87 0.85
Ligneous....... 9.73 — 3.32 —1.10 —1.81 —1.71
: I.00 0.75 0.90 0.84 0.85
All species..... 8.96 —3.11 —0.93 —2.05 — 1.63
1.00 0.74 0.91 0.81 0.85

under investigation are characterized by a lower osmotic concentra-
tion than those of any other habitat. To this rule there is not a
single exception. The values range from 74 to 99 percent of that
of other habitats.

Comparison of Blue Mountain Rain Forest with Other Regions.—
Our work in Jamaica was undertaken primarily to secure determina-
tions from an extremely hygrophytic habitat for comparison with the

292 J. ARTHUR HARRIS AND JOHN V. LAWRENCE

xerophytic region about the Desert Laboratory at Tucson and the
more mesophytic vegetation in the neighborhood of the Station for
Experimental Evolution on Long Island.

Since carrying out the Jamaican determinations we have been able
to make very substantial beginnings on the investigation of several
other habitats, for example the forests of the upper Santa Catalina
mountains and the various transition stations to the desert floor in
southern Arizona, the Everglades, the Pinelands, and the hammocks
of sub-tropical Florida, rich in West Indian species. A detailed
comparison of the montane rain forest with other regions may profit-
ably be reserved until the completion of these studies. In the mean-
time it is worth while to indicate to phytogeographers and ecologists
the relative position of the Blue Mountain habitats in the series
concerning which published data are available.

Consider first the values for the rain-forest plants as compared
with those obtained in more mesophytic regions. ‘Two such series are
available, that of Ohlweiler (12) based on trees and shrubs growing
at the Missouri Botanical Garden, and that of Harris, Lawrence and
Gortner (’15) for Long Island habitats. °

Ohlweiler’s St. Louis series suffers from two disadvantages as
regarded from the standpoint of this paper. First, it is based upon
a series of species brought together from various natural habitats
and cultivated in a botanical garden. All the species were, however,
capable of growth in the open under the conditions prevailing at St.
Louis. Second, sap was extracted without antecedent freezing of the
leaf tissue., As a result the freezing-point lowerings recorded are
probably too low.

Ohlweiler’s series comprises trees and shrubs only. Comparing
with the general average for ligneous plants from the Blue Mountains
the results are:

Means
SU; BOuisseries: 4) hemes ieee arte ace ws ee 14.96
Blue Mountain'sertes. \o. Yere oct ate ee, Cre tn, caterer 11.44

The trees and shrubs growing in the Botanical Garden at St. Louis
show, therefore, a concentration of their leaf sap of from 2 to 5 atmos-
pheres higher than do those of the various Blue Mountain habitats,
and over 3 atmospheres more than the average for the Blue Mountain
region as a whole.

OSMOTIC CONCENTRATION OF TISSUE FLUIDS 293

The averages for the Long Island series? have been calculated for
individual habitats. The averages for both trees and shrubs and for
herbaceous plants may be compared with the individual Blue Moun-
tain habitats. The means of the accompanying tables, VII-VIII,

TABLE VII
Comparison of Ligneous Plants

Jamaican Habitats Deae Island Habitats
| TIL CET Ae eo 13.05 | 13.34 | Beaches, coastal sand dunes and
marshes

ie getorest. 20 Gea Tie54

14.64 | Dryer woods and open fields
Ibeeward ravines. 2. o. i 56 6 10.83
Windward habitats........| 9.73 | 14.07 | Permanently moist localities
MUMS ae co ith oe erica eat 11.44 | 14.40 | All habitats

TABLE VIII

Comparison of Herbaceous Plants

Jamaican Habitats | Long Island Habitats
inmates of 7. es a 9.77 | 13.62 | Beaches, coastal: sand dunes and
marshes
td @estoreStn af oo. ak os 8.63
10.04 | Dryer woods and open fields
Meeward ravines; ........ 5. 7.59
Windward:-habitats........ 7.52 9.27 | Permanently moist localities
PAUHAITATS on es OI ee 8.80 | 10.41 | All habitats

show that with the exception of the herbaceous plants of the ruinate
there is no habitat of the Blue Mountain region which exhibits an
osmotic concentration of the leaf sap of the species constituting its
flora as high as the lowest mean found in the Cold Spring Harbor series.
The herbaceous plants of the ruinate—the most xerophytic of the Blue
Mountain habitats—show a concentration slightly higher than those
of the Long Island habitats which are constantly moist, 7. e., fresh
water bogs, lake shores and springy hillsides.

* The values given for Cold Spring Harbor are preliminary averages of deter-
minations, not of species means, made in 1914 by Harris, Lawrence and Gortner.
They will be replaced later by averages based on far larger series of determinations
made in 1915 by Lawrence and Harris, and on subsequent determinations by Harris.

294 J. ARTHUR HARRiS AND JOHN V. LAWRENCE

In view of the fact that the Long Island series here used is to be
much increased, further discussion of the observed differences may
be postponed until the more extensive data are worked up.

A conspicuous difference in the osmotic concentration of rain-
forest and desert vegetation is of course to be expected after the
demonstration of the differentiation of the sap properties of the plants
of this and more mesophytic regions. Two fairly satisfactory sets
of determinations for deserts are now available. The magnitude of
the differences between the rain forest and these will give some indica-
tion of the range of variation to be found in the mean osmotic con-
centration of the fluids of the species of different vegetations.

A comparison with the Arizona desert series of determinations
made at the time of hibernal and vernal vegetative activity® is made
in the accompanying tables, [IX—X. In these, averages‘ are given for

TABLEVIX
Comparison for Ligneous Perennials

pore Habitats Arizona Habitats
Roaimate 22 acta cos akeanse ahs $205. )) 22.01 Rocky slopes
Re SetOres taken eee eee 11.54 | 21.04 Canyons
Leeward ray ineseaesee ae tans TODO 3 a.\se17.20 Arroyos
Windward habitats! ye..0. 2.8 0.73
30.34 Bajada slopes
45.20 Salt spots
Aisha bita tes eet eee he e 11.44 | 24.97 All habitats :
TABLE X
Comparison for Herbaceous Plants
Jamaican Habitats Arizona Habitats
RUINAte peek ae eee ae O77) arson Rocky slopes
Ridgettorest::: 5 wage sa eenuase tics SOs Tense Canyons
Doeeward ravines... i004. es eos 7550s 2.00 Arroyos
Windward habitats. 3.4.5 2 aac 7.52
20.53 Bajada slopes
22°57 Salt spots
Pilbhabitats arcana. nee 8.80 | 15.15 All habitats

’ Studies on the summer vegetation have been made and will eventually be
published.

4 The averages for the southwestern deserts are based on species determinations,
not on means of determinations as in the Jamaica series. The difference in method
is of no significance for present purposes.

OSMOTIC CONCENTRATION OF TISSUE FLUIDS 295

each of the sub-habitats for both ligneous and herbaceous forms. In
our origina! paper (Harris, Lawrence and Gortner, 1916) the determina-
tions for the ligneous plants are further subdivided into trees and
shrubs as one class and dwarf shrubs, half shrubs and woody twiners
as the other. Such distinctions have not been so easily made in the
rain forest. The two groups of desert ligneous perennials have,
therefore, been combined to render them more comparable with the
Jamaica ligneous perennials.

The Arizona herbaceous plants were originally divided into the
two very distinct groups, winter annuals and perennial herbs. These
have also been combined to render them more nearly comparable with
the herbaceou plants of the Blue Mountain region.

Of course no one of the desert habitats is at all similar to those
of the Blue Mountains. Those which are least of all comparable,
the bajadas and the salt spots, have been set off from the others.

The tables show at a glance that the concentrations of the desert
are from fifty to nearly two hundred percent higher for individual
habitats in the Arizona deserts than in the Jamaica Blue Mountains.

The differences between the two regions are strikingly exemplified
by a comparison of the herbaceous plants of the desert with the
ligneous plants of the rain forest. The minimum osmotic concentra-
tion in desert herbaceous plants (12.99 atmospheres in the arroyos)
is practically as high as the maximum concentration for ligneous plants
in the Blue Mountains (13.05 atmospheres in the ruinate). The
mean concentration for herbaceous plants in the desert is 15.15
atmospheres as compared with 11.44 atmospheres, the mean con-
centration of ligneous plants in the Blue Mountains.

While logically a comparison of the rain-forest vegetation of the
Blue Mountains with the desert vegetation of the coastal deserts has
no greater significance than that with the vegetation of the Arizona
deserts it will, because of the relatively short distance separating
the two Jamaican habitats, have a greater interest for most readers.

The comparison with the coastal desert of the southern shore of
Jamaica (Harris and Lawrence, 1917) must be limited to ligneous
perennials. The average of the 31 species means for arborescent and
suffrutescent plants of the coastal desert, omitting only the herbaceous
Sesuvium, Bromelia, Bryophyllum and the Cacti, is 30.05 atmospheres,
as compared with 11.44 atmospheres for the montane habitats!

Very high concentrations are also found in the mangrove swamps

296 J. ARTHUR HARRIS AND JOHN V. LAWRENCE

on the southern shore of Jamaica (Harris and Lawrence, 19170).
Thus Rhizophora Mangle shows concentrations ranging from 29.2 to
30.9 atmospheres, Laguncularia racemosa shows concentrations ranging
from 24.6 to 34.8 atmospheres and Avicennia nitida yields values from
41.5 to 54.4 atmospheres.

IV. RECAPITULATION

The Blue Mountains of Jamaica, intercepting as they do the
trade winds in their sweep across the Caribbean Sea, exhibit a con-
spicuous differentiation in the flora and especially in the vegetation
of the windward (northern) and the leeward (southern) sides of the
range.

The windward slopes, and especially the windward ravines, exhibit
all those features of vegetation and of structure of the constituent
species which are called to the mind of the botanist by the term Rain
Forest. In the higher mountains the leeward ravines share many of
the characteristics of the windward ravines and slopes, but the leeward
slopes, and especially the scrub formation known as ruinate, are far
more xerophytic in their botanical characteristics.

The subalpine ridges, while lacking some of the most characteristic
and typical xerophytic species of the ruinate, are nevertheless clearly
far more xerophytic than either the windward slopes or ravines or the
leeward ravines.

These differences have long been known to botanists, and have
recently been splendidly described and illustrated by Shreve.

The purpose of the investigations described in this paper, which
is one of a series on the sap properties of the plant species of diverse
vegetations, is to present the results of an extensive series of cryoscopic
determinations of osmotic concentration of leaf sap in the species of
the Blue Mountains, to compare these habitats among themselves on
the basis of the average osmotic concentration of their leaf tissue
fluids, and to compare the region as a whole with other areas, meso-
phytic and xerophytic, which have been investigated in a similar
manner. ‘

The results of the present study confirm the conclusions concerning
the existence of a higher osmotic concentration in the tissue fluids of
the leaves of ligneous than in those of the tissue fluids of herbaceous
plants, drawn from the investigation of the deserts of southern Arizona.
The difference between the concentration of the sap of the two groups

OSMOTIC CONCENTRATION OF TISSUE FLUIDS 297

of growth forms is clearly marked in the series of determinations from
each of the Blue Mountain habitats. The differences are not, how-
ever, so large as those demonstrated in the desert series.

The four sub-habitats, recognized in the Blue Mountains, show
distinct differences in the osmotic concentration of their tissue fluids.

The ruinate, which has been regarded by ecologists as the most
xerophytic of the habitats, shows a distinctly higher osmotic con-
centration of the leaf tissue fluids of its constituent species than
any other habitat. The plants of the ridge forest show a higher
osmotic concentration than do those of the leeward ravines and the
windward ravines and slopes, but lower than that of the plants of the
ruinate. The leeward ravines are characterized by plants with lower
osmotic concentration than the vegetation of the ruinate and of the
ridge forest, but higher than that of the windward ravines and slopes.
Finally, the windward habitats, which are the most hygrophilous of
the region, are characterized by a sap concentration lower than that
of any other habitat.

The osmotic concentration in the sap of the plants of the Blue
Mountains is the lowest of that of any region as yet extensively
investigated. The ligneous forms show an average concentration
of about I1.44 atmospheres as compared with 14.96 atmospheres in
Ohlweiler’s St. Louis series and 14.40 for our own preliminary series
from Long Island habitats. The average concentration for herbaceous
plants in the Blue Mountains is about 8.80 atmospheres as compared
with 10.41 atmospheres from our preliminary Long Island series.

Comparisons with desert regions show much more striking dif-
ferences. Thus the herbaceous plants of the rain forest show an
average concentration of 8.80 atmospheres as compared with 15.15
atmospheres in the herbaceous plants of the winter flora of the deserts
around Tucson. The ligneous plants of the rain forest have a con-
centration of only about 11.44 atmospheres as contrasted with 24.97
atmospheres in the series of ligneous plants investigated in our south-
western deserts. The Jamaican coastal deserts show slightly higher
concentration even than those of the Arizona series.

While these general averages are the simplest expression of the
differences between these regions, they are by no means an adequate
description. ‘They conceal the differences which obtain in each of
the areas investigated. For a more adequate conception of the
conditions, the reader must turn to the more detailed comparisons

298 J. ARTHUR HARRIS AND JOHN V. LAWRENCE

which are made possible by the data presented earlier in these pages, .
and in the original papers to which reference has been made.

Further comparisons will be made when the data from other field
work are properly arranged for discussion.

CARNEGIE INSTITUTION OF WASHINGTON

V.” LITERATURE CITED

Dixon, H. H. and W. R. G. Atkins. Osmotic Pressures in Plants. I. Methods of
Extracting Sap from Plant Organs. Sci. Proc. Roy. Dublin Soc. N. Ser. 13:
422. 1913. Alsoin notes from Bot. School Trin. Coll. Dublin, 2: 154. I914.

Gortner, R. A. and J. Arthur Harris. Notes on the Technique of the Determinations
of the Depression of the Freezing Point. Plant World 17: 49. 1914.

Gortner, R. A., John V. Lawrence and J. Arthur Harris. The Extraction of Sap
from Plant Tissues by Pressure. Biochem. Bull. 5: 139. pl. r. 1916.

Harris, J. Arthur and R. A. Gortner. Note on the Calculation of the Osmotic
Pressure of Expressed Vegetable Saps from the Depression of the Freezing
Point, with a Table for the Values of P for A = .oo1° to A = 2.999°. Amer.
Journ: Bote 1:75. roid:

Harris, J. Arthur and John V. Lawrence. On the Osmotic Pressure of the Tissue
Fluids of Jamaican Loranthaceae Parasitic on Various Hosts. Amer. Journ..
Bot. 3: 438. Dzag. 1-2. 1916.

Harris, J. Arthur and John V. Lawrence. Cryoscopic Determinations on the Tissue
Fluids of the Plants of the Jamaican Coastal Deserts. Bot. Gaz. In press.

Harris, J. Arthur and John V. Lawrence. 1917). The Osmotic Concentration of
the Sap of the Leaves of Mangrove Trees. Biol. Bull. 32: 202. 1917.

Harris, J. Arthur, John V. Lawrence and R. A. Gortner. On the Osmotic Pressure
of the Juices of Desert Plants. Science, N. ser. 41: 656. I9I5.

Harris, J. Arthur and John V. Lawrence. With the co-operation of R. A. Gortner.
The Cryoscopic Constants of Expressed Vegetable Saps as Related to Local
Conditions in the Arizona Deserts. Physiol. Res. 2: 1. 1916.

Johnson, D. S. The Cinchona Botanical Station. Pop. Sci. Monthly 85: 521.
1914; 86: 33. IQ15.

Ohlweiler, W. W. The Relation between the Density of Cell Saps and the Freezing
Point of Leaves. Ann. Rept. Mo. Bot. Gard. 23: 101. 1912.

Shreve, F. A Montane Rain Forest: A Contribution to the Physiological Plant
Geography of Jamaica. Pub. Carn. Inst. Washington I9g9. 1914.

THE VIABILITY OF RADISH SEEDS (RAPHANUS SATIVUS L.)
AS AFFECTED BY HIGH TEMPERATURES AND
WATER CONTENT

H. D. WAGGONER

The first important work on the effect of high temperatures upon
the viability of seeds was done by Edwards and Colin in 1834. From
that date up to the present time considerable interest has been shown
along this line and a large and valuable literature has accumulated.
A large majority of the earlier workers in this field were interested in
the maximum temperatures that seeds are able to withstand, and
paid but little attention to the cause of the loss of vitality when seeds
were heated at temperatures above their maximum. Recently several
investigators have studied the effect of high temperatures upon proto-
plasm, and important contributions have appeared that throw much
light upon the real cause of death in seeds.

It seemed advisable, owing to the lack of time, to confine the
present paper to a detailed study of the effect of high temperatures
upon the germinating power of seeds and to reserve the consideration
of the direct effect of heat upon the living protoplasm of seeds for a
subsequent study.

A close study of the results obtained by earlier workers shows a
wide difference in the maximum temperatures that the same or
nearly related seeds can endure without injury. It further reveals
the interesting fact that an intimate relation exists between the method
employed in treating the seeds and the temperature at which serious
injury or death occurs. Briefly, the methods heretofore employed
are as follows: (1) The seeds were heated in water or in an atmosphere
saturated: with moisture. By this mode of treatment it is apparent
that the seeds absorbed more or less water during the heating process.
Authors using this method invariably report a low lethal temperature
for the seeds used. (2) The seeds were heated in small closed con-
tainers. By this method the seeds during the process of heating gave
up moisture to a greater or less extent until an equilibrium between
the imbibition energy of the seeds and the vapor tension of the inclosed
air was established. The results obtained by this method varied

299

300 H. D. WAGGONER

according to the amount of water present in the seeds when heated.
(3) The seeds were heated in ovens. In this case it is evident that
the air in the ovens was of low relative humidity and consequently
the seeds lost moisture during the heating. Seed treated in this
manner endure exceedingly high temperatures without apparent injury.

It seemed desirable to make a detailed study of the resistance of
seeds of different water contents to high temperatures under carefully
controlled conditions. Accordingly, experiments were carried on in
the laboratories of plant physiology under the direction of Professor
Chas. F. Hottes, to whom the writer is greatly indebted for searching
criticism and helpful suggestion.

In taking up this subject anew two closely related lines of experi-
mentation were outlined. The one was concerned with the effect of
high temperature upon series of samples of seeds of increasing water
content. The other was to determine the cause for the wide difference
in the resistance of seeds treated by the three methods indicated
above. <A survey of the more pertinent literature will serve to bring
these two points definitely before us.

As early as 1859, Heiden reported that grains of barley, when
exposed for one hour to dry air at 90° C. germinated, while similar
grains heated in water at 60° C. for the same period of time were killed.
In 1865, Fiedler (Sachs), working with seeds of pea, rye, flax, barley
and corn, showed that swollen seeds were killed at from 50° C. to
60° C., while those containing less moisture withstood 70° C. or more.
The seeds were treated in closed test tubes immersed in water main-
tained at the desired temperature. Von Hodhnel (1877) improved the
method as used by Fiedler in that he covered the seeds in the test
tubes with fine metal filings to facilitate the transfer of heat. He
worked with seeds of a number of different species and reported that
most of them when sufficiently dried were able to endure an exposure
at 110° C. for sixty minutes. Some heated at 125° C.. for fifteen
minutes were found viable. Just (1875, 1877) reports that clover
seeds heated in a saturated atmosphere at 50° C. for forty-eight hours,
or at 75° C. for one hour, lost their viability. However, if the seeds
were well dried they could endure a temperature of 120° C. Other
kinds of seeds gave similar results. Detmer (1880) records that the
viability of seeds is not lost in boiling water, provided they are not in
a swollen state. He further states that the less moisture seeds contain
at the time of treatment the greater is their resistance to high tempera-

THE VIABILITY OF RADISH SEEDS 301

tures. Other investigators of this early period obtained similar results.

More recently, Jodin (1899) found that seeds of peas and cress
when dried at 60° C. for twenty-four hours can be heated at 98° C.
in the dry air of sealed tubes for six hours and still retain their viability ;
similar seeds heated in humid air at 40° C. lost their viability in twenty
hours. In subsequent experiments he found that the resistance of
the seeds of pea and cress was increased in a marked degree when
calcium chloride was introduced into the tube with the seeds. He
states that seeds may be exposed to dry air at 65° C. for prolonged
periods without loss of vitality, but adds that this may be done only
if one heats them in an open dish to permit a rapid loss of water from
the seeds. If the air becomes saturated with water vapor, the seeds
can endure only a comparatively low temperature without injury.
Dixon (1902) dried various kinds of seeds over sulphuric acid and
later in an oven at 95° C. for several days without destroying their
viability. He treated the samples in closed test tubes and found that
when sufficiently dry the seeds could withstand temperatures far
above that of boiling water (110° C.-120° C.) without injury. Neu-
berger (1914) and others of recent date, in their studies on the com-
parative resistance of moist and dry seeds to high temperatures, in
a general way, confirm the results of the earlier authors.

Pouchet (1866), working with seeds of Medicago obtained from
sheep-wool brought from South America, found that they germinated
after being exposed to boiling water for four hours. He subsequently
experimented with other Medicago seeds and found that only those
germinated which, after a prolonged treatment in boiling water, did
not swell. Nobbe (1876) confirmed Pouchet’s results. Dixon (1901)
and Schneider-Orelli (1909, I910) attributed the high resistance of
these seeds to the fact that many of the seed coats are impermeable
to water.

From the above review, it is apparent that considerable work has
been done on the resistance of seeds to high temperatures. The
experiments have been carried on with seeds containing widely different
and undetermined amounts of water. As far as the writer has been
able to determine no one has attempted to study, in series, the resist-
ance of the same kind of seeds containing definite and known quan-
tities of water at the time of heating. It is only through a quantitative
study of this kind that a definite knowledge of the various factors
involved can be obtained, and the variation in results harmonized.

302 H. D. WAGGONER

The problems before us then are: (1) to determine the definite relation
between the water content of seeds exposed to high temperatures
and their viability; and (2) to explain the different degrees of resistance
of seeds exposed to high temperatures when treated by different
methods.

MATERIALS AND METHODS

The seeds used in these experiments were Icicle, Black Spanish
Winter and Crystal Forcing radish obtained from Henry A. Dreer,
Philadelphia. These were selected with the view of comparing the
relative resistance of varieties adapted to widely different -cultural
conditions. The seeds were of good quality and throughout the
experiments tested at approximately 90 percent. They were stored
in glass-stoppered bottles and kept in a room where the humidity of
the air varied but little; the moisture (4 percent) in the seeds remained
practically constant throughout the experiments. A tin measure was
used to obtain random samples from the bottles and no selection was
made other than that the seeds with ruptured coats were discarded.

Three methods were used in heating the seeds. (1) The samples
were placed in Florence flasks of 100 c.c. capacity, the bottoms of which
were sufficiently large to allow each seed to come in direct contact with
the glass. The flasks were placed in a wire cage and entirely sub-
merged in a bath containing about thirty-five liters of water previously
heated to the desired temperature. This large quantity of water,
kept in circulation by a Kohler stirrer, rendered it possible to maintain
a temperature constant to within one half of a degree Centigrade.
Increase in pressure during the heating was guarded against by pro-
viding the containers with corks through which capillary tubes thirty
centimeters long were passed. ‘Temperatures above the boiling point
of water were obtained by adding the requisite amount of calcium
chloride to the water of the bath. In these latter experiments pro-
vision was made for replacing the water evaporated from the bath
while the experiments were in progress. (2) The samples in shallow
open pans were placed in a double-walled copper oven filled with
glycerine. The temperature was controlled to within one half of one.
degree Centigrade. (3) The samples were enclosed in muslin bags
and immersed directly in the water of a bath previously heated to ~
the desired temperature.

The water content of the stored seeds was determined by finding
the loss in weight of a random sample brought to constant weight in

THE VIABILITY OF RADISH SEEDS 303

an oven maintained at 104° C. In the experiments requiring a reduc-
tion of the water content of air-dry seeds, the amount of water present
was diminished by heating, first in an oven at 60° C. and later in one
at 100° C. Samples were treated with the desired temperature when,
upon weighing, it was found that the water content of the seeds had
been reduced to the desired amount. To increase the water content
the air-dry seeds were exposed to a saturated atmosphere until they
had absorbed the amount desired in a particular experiment. This
method proved satisfactory when only a slight increase in water
content of the seeds was desired. When a considerable increase was
necessary, the air-dry seeds were soaked in tap water at 20° C. until
they had absorbed the required amount. The seeds, superficially
dried between towels, were placed in flasks and allowed to stand for
a time to permit the water to penetrate uniformly before heating.
A momentary immersion of the dry seeds in 95 percent alcohol was
found beneficial in that it allowed a quick and uniform wetting of the
coats when the seeds were placed in water.

The seeds were germinated on plaster of Paris blocks twelve
centimeters square and three centimeters thick. The surfaces of these
blocks were crosschanneled so that there were one hundred inter-
sections suitable for the hundred seeds used in each test. The blocks
containing the seeds were placed in fiber tubs and water added to a
depth of two centimeters. The tubs were covered and kept in a
dark room at approximately 23° C. A daily record of the number of
germinated seeds was made, and this continued for fourteen days.

EXPERIMENTS AND DISCUSSION

All the earlier investigations in this field were carried on with
very dissimilar seeds of widely different, and in no case definitely
determined, water content. In no instance was an attempt made
to find the effect of high temperatures acting for definite periods of
time upon series of like seeds of different but known water contents.

Radish seeds (Icicle, Black Spanish Winter, and Crystal Forcing),
with an initial water content of from 4 percent to 71 percent as indi-
cated in Tables I, II, and III, were placed in the flasks already
described and heated for thirty minutes at temperatures (50° C. to
125° C.) indicated at the head of the tables. After heating for. this
period they were placed on the blocks for germination. The results

304 H. D. WAGGONER —

THE EFFECT OF HIGH TEMPERATURES UPON RADISH SEEDS OF KNOWN INITIAL WATER CONTENT

Table I. Icicle Radish

Temperatures Employed...... 50° | 55° | 60° | 65° | 70° | 75° | 809 | 85° | g0° | 95° |z00°} 705°) 126° 235° |r20° 123°|125°| Check
Percent of germination:

With 71% water present|80%| 45) 0 87
66 50% bé 6 83% 50 O 88
“c 45% “ “c 84% 69 re) 87
Ceol ae (87%, 72) 14) © 89

30% 74} 25| O go
66 237, (a (iH 74 38 I Oo 90
aan oie ie 78| 66| 38] II] oO 88
Reed Vike Amun ys 85| 74) 69] 63] 26] 18] 6) o 89
Pe OS toa - 88} 82| 85] 74] 54] 36] 15) Io; o 90
a Ah - 86} 91} 90] 91} 82} 79] 76} 4o| oO} O 89
ee gee . | 89| 88) 87| 90] 87} 80] 72] 57/ 5] oO 88
Pe te pia | 90} 89| 88] 82} 73] 67] 32] 4] o 89
VS OU oy. . | 89| 88] 85] 76| 62} 32} 0} Oo} 89
eA pane ca | go| 86] 78] 64] 28} 14} 0] 90

Table II. Black Spanish Winter Radish
Temperatures Employed...... BOo 1551002 | O52 pot War Oe 8oo 1850 | go® | 95° | 100° Check
Percent of germination:

With 71% water present 67%] 43) 0 88
66 50% bo (a9 69% 52 Ol go
1 ASO. “ |80%] 53} 0 go
pe eee me CEC ARGG avila 87
c 30% ‘6 ‘c 62 46 re) . 87
ee reas . 73| 67/ 10) O| O go
wae chy anes < 76 7S1 27 2 Teet | Ol) sO 88
TMA One ‘i 85] 86) 78] 49] 37] 33] 26) 0] o 90
a Ot ss ye 87) 88) 86) 67) 55] 50] 38) 7| o| Oo 88
ete ko are es 90! 90] 87! 88] 82] 77] 66] 38; oO 89

Table III. Dreer’s Crystal Forcing Radish
Temperatures Employed ...... 50°. 155° | Go? || 65° | zo 75°) 80° 8521 go0° |'g5°! | Ta02 Check
Percent of germination:

With 71% water present|87%| 60} 0 93
6 50% c ‘c 89% 66 O 92
«c 45% “6 cc 92 Z| 66 O | Ye)
oven \ 9870172) ule go
a 30% 66 a 75 24: re) 89
eG $3| 74) 26) 01.70 rere)
SOC g. “i 85] 81! 66} Ig] To; oO] O| O 89
ee TA ones. i‘ 91} 84] 80] 68) 47| 39] 13) 3] ©o 92
ovo ie 91} 93) 91| 83] 74] 64) 56) 28; Oo; Oo 89
at aA ae - QI! 90} 91\ 94] 88} 86] 74] 25] Oo 88

THE VIABILITY OF RADISH SEEDS 305

indicated in the tables are based upon the treatment of approximately
sixty thousand radish seeds. The resistance of the three varieties
proved to be so similar that a separate discussion of each is unnecessary.

An examination of these tables shows that there is a definite
relation between the initial water content of seeds heated at high
temperatures and their viability. Seeds of an initial water content
of 71 percent, 50 percent and 45 percent are killed at 60° C. As the
water content decreases in the successive series we find the percent of
germination to increase and the death point of the sample to be
markedly raised. For example as the water content is decreased from
45 percent to 30 percent the lethal temperature shifts from 60° C. to
65° C. Air-dry seeds of approximately 4 percent water content
give a normal germination after treating at 75° C. and are killed
between 95° C. and 100° C. On the other hand, samples carefully
dried until only .4 percent of water is present at the time of treatment,
give a normal germination at 100° C. and are killed between 123° C.
and 125° C. We find then as the water content increases from .4
percent to 45 percent that the maximum temperature at which a
normal percent of germination takes place drops from 100° C. to
below 50° C. and that the lethal temperature drops from between
123°-125° C. to 55°-60° C. The resistance of radish seeds exposed
to high temperatures decreases as their initial water content increases.
Furthermore at temperatures high enough to be injurious, the viability
of radish seeds of a definite initial water content decreases as the tem-
perature to which they are exposed is raised.

In recording the percent of daily germination not included in
the above tables, a greater or less degree of retardation in the germina-
tion of seeds, treated at temperatures a few degrees below their
maximum, occurred so constantly that it was thought advisable to
make a definite quantitative study of the same. In Table IV are
recorded the results of a series of experiments on the rate of germina-
tion as affected by initial water content and high temperatures.

Samples of five hundred radish seeds containing 4 percent, 9 per-
cent, I4 percent, and 18 percent of water, respectively, were heated
for thirty minutes in 100 c.c. flasks submerged in a bath held at 80° C.
and subsequently allowed to germinate on plaster of Paris blocks.
The data (Table I) show that at this temperature seeds containing
the above water contents suffer in direct proportion to the quantity
of water present. The observations were carried on over a period of

306 H. D. WAGGONER

seven days and the results obtained recorded as indicated in Table IV.
The daily percent of germination of the five hundred untreated seeds
is shown in the last column. In the other columns is recorded the
germination of seeds of an initial water content of 4 percent, 9 percent,
I4 percent, and 18 percent respectively, when heated at 80° C. for

thirty minutes.

TABLE IV
The Retardation in Germination Caused by Varying Water Content at Temperature of
80° for 30 minutes

Check

Water Content-of Seeds when Heated | 4% of 14% | 18% Un-

| | treated
Percent of germination the 1st day. ..% va.) -.. 6% 16 O O 30
a e sa Pe 2 GL Os teed eee eee 21% 5.4 O O 45
‘ . A . aa ere ee 125 5 ary. Th. O 14
ol om oe Or ae reine gt 3 Gs deta Sbee aie O 3
: ‘ et ieee: Pe AS lees plan ke) 6 O 2
cs a i ie POU. ep ener ate 1% 26 2e2 fe) I
_ . 7 Ula (Ae ake oko e ae, | o% O 4 O O
Total percent oficermination,. 3... 15.6 eee 839,, | 60.8) |) 2720 O 95

The maximum daily percent of germination in the check (45 per-
cent) occurred on the second day, while in the treated seeds the highest
number of seeds germinated on the third or fourth day. Seventy-five
percent of the untreated seeds germinated in the first two days as
compared with twenty-seven percent for those of 4 percent moisture,
six percent for those of 9 percent moisture and zero percent for those
of 14 percent moisture. The resulting injury due to the increased
water content is shown by the rapid decrease in total germination of
the samples, namely: 95 percent, 83 percent, 60.8 percent, 27 percent
and o percent respectively. The retardation in the germination of
radish seeds becomes greater as the injury due to the treatment
becomes more marked.

To correlate the results found in the literature with my own, it
became necessary to repeat, under control conditions, the different
methods heretofore used. In the preceding section (Tables I, II, IIT)
the intimate relation between initial water content and viability at
high temperatures was definitely determined, and the suggestion
made that the wide difference in the lethal temperature of seeds of
similar kinds as reported by different investigators was to be sought
in the initial water content of the seeds or in the method employed
that would allow an increase or decrease of this water content during
the heating process.

THE VIABILITY OF RADISH SEEDS 207

“Three series of similar samples of radish seeds containing an initial
water content of 19 percent were heated at temperatures from 45° C.
to 105° C. for a period of thirty minutes (Table V). The samples
of Series I were heated directly in water, those of Series 2 in 100 c.c.
flasks immersed in water, and those of Series 3 in an oven. The seeds
were placed upon the plaster blocks for germination. The results in
percent of germination are recorded in Table V.

TABLE V
The Effect of Different Methods of Heating upon the Germination of Seeds

Temperatures Employed 450) FSO 50 002 "O50 | 70° | 75° | 80° | 85° | go° | g5° |100°| 105°
Percent of germination of seeds (Sane.
When heated in water. Se- | | |
PISS S0 a ae ae 88% | 80/58} 2] 0} O| iar
When heated in flasks. Se- age
(ES. 2. oy eae Ak ee a | 89|87/81|66|18| 9| oO] oO

When heated in oven. Se- | lames ciuee
(FILES rae RAR ma a ea 89 | 88) 84| 76; 60] 6] oO
@hecks, untreated... 2... ... 3 88% | 90! 89! 88! 89/ 91 1,90; 90) 89. 881 90) 90| 89

The results as given in this table show very clearly that radish
seeds of similar water content when heated as indicated in Series I, 2,
and 3, respectively, exhibit very different degrees of resistance. The
samples heated directly in water suffered a loss at 50° C. and were
killed at 65° C.; those heated in flasks suffered a similar loss (approxi-
mately 8 percent) at 60° C. and were killed at 80° C.: those heated in
the oven suffered a loss of 85° C. (5 percent) and were killed at 105° C.
Further it is to be noted that in Series 1 the effects of the temperature
are distinctly manifested at 50° C. and that from this temperature
the viability decreases very rapidly, extending through a range of
only 15° C. In Series 2 similar effects are noted 10° C. higher, namely
60° C., and the viability decreases less rapidly, namely extending
through a range of 20° C. The most marked effect is shown in Series
3. The effects of the treatment here lie 35° C. higher than in Series
1 and 25° C. higher than in Series 2. The decrease in viability is
slow at first and then very rapid, falling from 60 percent germination
Ae Oh Opto onnen cettiat 100. C.

The data recorded in Table V, Series 2, are practically a repetition
of those found in Tables I, II, and III, for an approximately similar
initial water content. Starting with an initial water content of 19
percent in Series 1, Table V, we find the injury essentially equivalent

308 H. D. WAGGONER

‘to that of seeds with an initial water content of approximately 71
percent (Tables I, II, III). In Series 3, Table V, with the same
initial water content, we find the injury essentially the same as that
for seeds of 4 percent initial water content (Tables I, II, III). This
is exceedingly suggestive.

Three series of weighed samples of radish seeds with an initial
water content of I9 percent were heated at 65° C., 80° C. and 95° on
for periods of 3, 6, I0, 15, 22, and 30 minutes respectively. The
series were repeated for each of the three methods already indicated
in Table V. At the end of each period of heating the samples were
removed, weighed, and the loss or gain recorded as indicated in
Tables VI, VII, and VIII. The data from these tables were used in

TABLE VI
The Increase in Weight of Seeds when Heated in Water

Periods of Exposure....... Seontnces | 3 Min. 6 Min. to Min, 15 Min. |, 22 Min. | 30 Min.
Percent of 1 increase at 65° C....| 13% 21 | 29.4 a7, 45 53.2
i 2 80° CHRP peo A Meee 45 57.1 60.3 61
= si is EOS €C save 2 BoA 130 | 51 59.8 60.1 60.4
TABLE VII

The Loss of Weight in Seeds when Heated in Flasks

15 Min. | 22 Min. | 30 Min.
2.9
I

Periods of Exposure............... 3 Min. | 6 Min. 10 Min,

Percent ofiloss at.65 °C. oa se. = 2% B 2.9 3 3
FOND Er SS et ce 4% ot ere | 5 5.1 5.
OS Oe hae ee 7, 8 70 7.9 8 8
TABLE VIII

The Loss in Weight of Seeds when Heated in an Oven

. | 22 Min. | 30 Min.

Percent of loss:at 65° Ce aic as 9.8 10.5
Se eo One ieee a sae TOA 13
os Yh a esO& eles n ee 14.4 15.9

constructing the graphs in Fig. 1. The ordinates express the percent
of gain or loss in weight of the samples (above or below 19 percent)
according as the graph extends above or below the line AB. The
abscissae represent the periods of time during which the seeds were
exposed to the respective temperatures. In order to obtain the actual
water content of the seeds, one must add or subtract the indicated

ob

THE VIABILITY OF RADISH SEEDS 309

Pics f

1
- T ic} T
H |
it HHH
c T
see —
1
I I
i T i u
1 coe sfelett
1 T
see T !
7
a ot }
| | I T
I
: !
7 is
i
Gi
ia
E
H rH H+ =
T EE Ht rH Ht
ce EEE EEE
+ n o (EL Se Tee ry 7
it t T T E a
T 1 im perros tr 7
T TT T saenee anumiors
t t 1 H HES Eeslceae
: l i T Tt
i " i.
T T ot Bit ; HH y
ima Poo
me :
1 rH
T t CE -
A i f
i u i = i
| u | i
+
tT | TH H
Y T T
im) T l mm {
4 HH i T T
imi
t i
i
T T
T
1
|
1 t

Graphs showing the changes in water content obtained by heating
radish seeds.

310 H. D. WAGGONER

gain or loss (Fig. 1) to or from the initial water content (19 percent)
of the seeds used.

These graphs show the striking changes in water content obtained.
by heating radish seeds of I9 percent initial water content by the three
different methods. (A subsequent experiment will show that the
gain in weight of the samples heated by method 1 does not express
accurately the amount of water absorbed.) The seeds heated directly
in water increased in weight very rapidly; those heated in the flasks
decreased in weight for about six minutes and then remained constant
during the remainder of the heating; those heated in the oven grad-
ually decreased in weight throughout the entire thirty minute period
of treatment. At the end of thirty minutes the seeds heated at 95° C.
were found to have a water content for method 1, of 79 percent,
method 2, II percent, and method 3, 3 percent. A similar variation,
though less extensive, is seen in seeds heated at 80° C. or 65° C.
respectively (Fig. 1). It has been shown that the resistance of seeds
exposed to high temperatures is inversely proportional to the initial
water content at the time of heating. The results obtained in the
above experiments (Fig. 1 and Table V) are in entire accordance with
this statement.

Data are not at hand upon which to base a complete explanation
of the resistance of radish seeds to high temperatures. Seeds heated
in air lose water by evaporation which is a cooling process. Moreover,
air and water differ materially in their ability to transmit heat. Hence
we cannot say with absolute certainty that we have subjected the
seeds used in the corresponding tests to exactly the same temperatures
for the same periods of time. Data from seeds heated at a given tem-
perature without suffering any change in the water content during
heating would be desirable. However the results as shown in Table
V between water- and oven-heated seeds cannot be explained on the
basis of cooling. These questions together with related ones will
form the basis for a future investigation.

The rate of gain or loss in water content of the seeds when treated
by method 1, 2 or 3 is of interest, since, owing to a change in their
water content, the seeds show a low (Series 1) or a high (Series 3)
resistance. The rapid absorption of water in the seeds of Series I
concomitant with the high temperature is responsible for the marked
injury as indicated in Tables I, II, III, and V. On the other hand the
rapid loss of water in the first six minutes of treatment in Series 2

THE VIABILITY OF RADISH SEEDS 311

and the subsequent maintenance of approximately the same water
content for the whole period of heating—due to equilibrium of vapor
tension—brings the resistance markedly higher. In the seeds of
Series 3 the water vapor was constantly carried off and consequently

the seeds lost water throughout the process.

Attention should be called to the fact that radish seeds heated
directly in water lose in dry substance. It is necessary to know the
extent of this loss in order to interpret more accurately the results
found in Table VI. This experiment was repeated and the loss of
dry matter determined and recorded as shown in Table IX.

TABLE IX
The Loss in Dry Weight of Seeds when Heated in Water at 65° C., 80° C., and 95° C.
Respectively
Periods! of HxXpOSure.........0...0-s0 3 Min. 6 Min. to Min. | 15 Min. | 22 Min. | 30 Min:
Percent ot lossat 65°-C.:....... e217, 4 oy I 1.4 | 2
BAe OO Ne tee 45) HOWE 82 Tet Zor BO Sea
i Ber OS, AC shale cond 785 1.6 2:0 Eis Guts |. 1759

These results show that there is a gradual loss in dry weight when
radish seeds are heated directly in water at the temperatures indicated.
Moreover the higher the temperature of the water the more rapid is
the loss in weight. The seeds heated at 65° C., 80° C., and 95° C.
for thirty minutes lost 2 percent, 5.4 percent, and 7.9 percent respec-
tively. It follows from this that the actual amount of water taken
up by radish seeds heated directly in water is considerably more
than the increase in weight as indicated in Table VI.

CONCLUSIONS

The resistance of seeds of Raphinus sativus L. exposed to high
temperatures is inversely proportional to the initial water content of
the seeds at the time of heating.

At temperatures high enough to be injurious the viability of radish
seeds of a given initial water content decreases as the temperature
to which they are subjected is raised.

The general resistance of Icicle, Black Spanish Winter and Crystal
Forcing radish seeds exposed to high temperatures is very similar.

Radish seeds injured by high water content and high temperatures
are retarded in their germination. This retardation becomes more
marked as the temperature or water content or both is increased.

312 H. D. WAGGONER

Radish seeds of the same initial water content show very great
differences in resistance when heated at the same temperature but
by different methods, namely: in water, in dry corked flasks, or in
open dishes in ovens.

The amount of water absorbed or given off by radish seeds during
treatment is the chief factor determining the resistance of the seeds
heated at the same temperature by the different methods.

When radish seeds are heated directly in water they suffer a
gradual loss of dry substance. This loss becomes greater as the
temperature of the water is increased.

UNIVERSITY OF ILLINOIS,

URBANA
LITERATURE

Buglia,G. Ueber die Hitzegerinnung von fliissigen und festen organischen Kolloiden.
Zeitschr. Chem. Indust. Kolloide. 5: 291. 1909.

Crocker, Wm. and Groves, J. F. A Method of Prophesying the Life Duration of
weeds) eProct Nat. AcadSeiv1:71526" 7ons:

Crocker, Wm. Mechanics of Dormancy in Seeds. Amer. Journ. Bot. 3: 99. 1916.

Detmer, W. Vergleichende Physiologie des Keimungsprocesses der Samen. 1880.

Dixon, H. H. Resistance of Seeds to High Temperatures. Report British Associa-
tion for Advancement of Science. (London) 805. 1902. Abstr. in Ann.
Bot. 16: 590. 1902; also in Nature (London) 64: 256. 1901.

Dixon, H. H. The Germination of Seeds after Exposure to High Temperatures.
Notes from Bot. School Trin. Coll. Dublin, 176. 1902.

Edwards and Colin. De l’influence de la température sur la germination. Mem.
Acad. Sci. Nat. II Bot. I: 264. 1834. Cited by Nobbe, F. Handbuch der
Samenkunde 228. Also, Detmer, W. Vergleichende Physiologie des Kei-
mungsprocesses der Samen, 402. .

Ewart, A. J. Longevity of Seeds. Proc. Roy. Soc. Victoria, n. ser. 21: 1. 1908.

Goodspeed, T. H. The Temperature Coefficient of the Duration of Life of Barley
Grains. - “Bot .Gaz. 5124220; - .1OLr:

Heiden, E. Ueber das Keimen der Gerste. Dissertation 34. 1859. Cited by
Detmer, W. in Vergleichende Physiologie des Keimungsprocesses der Samen,
403. 1880.

v. Hohnel, G. Welche Warmegrade trockene Samen ertragen, ohne ihre Keim-
fahigkeit einzubiissen. Wissenschaftlichpraktische Untersuchungen auf dem
‘Gebiete des Pflanzenbaus. 2: 77. 1877.

Jodin, V. Sur la résistance des graines aux températures élevées. Compt. Rend.
Acad. Sci. (Paris) 129: 893. 1899.

Just, L. Ueber die Wirkungen héherer Temperaturen auf die Keimfahigkeit der
Samen von Trifolium pratense. Bot. Zeit. 52. 1875.

Just, L. Ueber die Einwirkung héhern Temperaturen auf die Erhaltung der Keim-
fahigkeit der Samen. Beitr. Biol. Pflanz. 2: 311. 1877.

Lepeschkin, W. W. Zur Kenntnis der Einwirkung supramaximaler Temperaturen
auf die Pflanze. Bericht. Deutsch. Bot. Ges. 30: 703. 1902.

THE VIABILITY OF RADISH SEEDS 313

Neuberger, Fr. Das Verhalten der Samen von Papilionaceen gegen hohere Tempera-
turen. Kiserletiigyl Kozlemenyck, (Budapest) 17: 121. 1914. Abstr. Bot.
Centralbl. 126: 665.

Nobbe, F. Die physikalische Bedingungen des Keimprocesses. Handbuch der
Samenkunde. 226, 1876.

Pouchet, A. Expériences comparées sur la résistance vitale de certaines embryone
vegetaux. Compt. Rend. Acad. Sci. (Paris) 63: 939. 1886.

Sachs, J. (Fiedler). Beschadigung und Tédtung durch zu hohe Temperatur.
Handbuch der Experimental Physiologie der Pflanzen. (Handbuch der
physiologischen Botanik.-Hofmeister) 63. 1865.

Schneider-Orelli, O. Sur la résistance de grains de Legumeuses aux températures
élevées. Bibliotheque universelle de Geneva. 28: 480. 1909.

Schneider-Orelli, O. Versuch tiber die Widerstandsfahigkeit gewisser Medicago
Samen gegen hohe Temperaturen. Flora 100: 305. 1910.

bz

fe

a aval

Devore 1 TO. > Ant Beer

HANG RT #, ves ‘Newcomer, Editor-in. Chiee
reais Spee aan . | University of Michigan:

See rab ae el
i

te StuaRT Clee Business Manager. ie, a OURS ASE

Ke i ae oe i a a Brooklyn Botanic Garden US a ce ae
Da eS _ Inente W; ‘BArey, | ee ee Py Ah ae
a A bes (Ped Institution Oe x eet ee ae
ei ve oe nee HL H.. Sancuecr i ha ee 6 a

i “hy ee A sabi ae Js Mechigan a lia ie ae i k

ge gh

ee A

less aipaae Sts “The: pages ‘of the (eats abe open on members: of ‘the i
Arcos or eandisates iii for aug ceae oa as
: et cases ‘be submitted to he Editor-in-Chief,

i NE _ Papers, are limited to 24 pages in. tength, ee of PO oe
nent i dione pPEse a be: cate tee at cost Fate: to miasihore oe te sisea

; ee | Page)

arene of Boumny, Tveaklya Botanic prten Brobkiyins N. ve

a’ “ a | : eee - Illustrations, other’ than zine, etchings, and tabular matter fh
be restricted to) approximately ten. n percent of the. Sus de

aa AMOURE, (approximately $3. 00 a page ‘for tabular matters ad ag.sente oe ine
PR ant r halfstones), AL eats ae eye \
et y Separates onad be lewd’ whee proof i is .returdied, | “Fifty cop itl witha SS
ef will be supplied free; cover. and additional copies at aa \

|.” Remittances should be made payable to An re
"| eents must be added to all checks not drawn on New York icy tae ie

ert oy ae ie oy : ~ date of mailing. “The Publishers will, supply missing numbers ak th
SipegAg Hes have been lost in the mails. :
~ Cotrespondence. ‘concerning ‘editotial matters, should be

| " : e Newcombe, 9 Geddes Heights, Ann Arbor, Mich, a
Ge gare Business correspondence, including notice of change ‘of address, an ‘diret
HOVE, Coat - concerning reprints, » should be addressed to: American. Journal of. Botan + Bre
A eh aes Poe Garden, renee yt, yee or 4 North. Queen treet, Lancaster,

Aegonian |nstity PP
& %

yULZ- 1917 +
Wy

AMERICAN
JOURNAL OF BOTANY

VoL. IV JUNE, 1917 No. 6

tab LAXONOMY OF THE AGARICACEAE*

WILLIAM A. MURRILL

No classification is perfect; nature recognizes no very marked
divisions. As Professor Massee, of Kew, used to say, ‘‘Why make a
fence? Some rooster is sure to get on top of it with his head on one
side and his tail on the other.’”’ And often the higher realms of
perfection are of little practical use. It is vastly more important to
help twenty students to a better knowledge of a group of plants than
to tickle the fancy and win the praise of one who no longer needs help.
Any one leaving the beaten track "is subject to criticism, when he
should get only sympathy.

Genera are not now considered sacred entities, but simply groups of
species that are more closely related to each other than they are to
other groups. New species are usually forced on one; and they
entail a lot of extra work requiring initiative, insight, and independent
thinking. Examine the shelves of almost any taxonomist and you
will find them cluttered up with species he knows are new and is too
indolent to describe. The very men who criticize most loudly the
multiplication of species would probably lack the courage, ability,
and patience to publish new ones.

Nomenclature and the rules of nomenclature are only a means to
an end, like the rules of a whist game. No matter how meritorious
one’s work may be, tradition and custom are as hard as granite rock
and the older workers can hardly be expected to change the names to
which they have become accustomed. As our knowledge grows, forms
of widely distributed species may merge into one and names may be
reduced to synonymy. In such cases, it is very unwise and unscientific

* Invitation paper read before the Botanical Society of America and affiliated

societies at New York, December 28, Ig16.
[The Journal for May (4: 253-314) was issued May 31, 1917.]

315

316 WILLIAM A. MURRILL

to uphold an error because it happens to be one’s own. The injection
of the personal element only impedes progress; the mind should be
kept open and unbiased, ready for any new light. What crimes have
been committed, both in Europe and in this country, in the name of
Science! Crimes of ignorance, of carelessness, of egotism, of petty
selfishness, of border antagonisms and national hates! But that is
human nature, and, after all, the scientists have more than their share
of piety.
MORPHOLOGY AND TAXONOMY

Not so very long ago, most American botanists were taxonomists.
Now the pendulum has swung the other way. The temptation is very
strong for professors in small colleges to limit themselves to a few
forms and to go rather thoroughly into their structure and life cycle
rather than to attempt to know and classify many forms, which
requires a large herbarium and a large library. National and state
aid, also, has been largely in the direction of physiology, morphology,
and pathology, which is only natural and proper; but it means that
taxonomy and the study of plants not strictly economic must be em-
phasized in institutions properly endowed and equipped for this
purpose and that all botanists who enjoy the advantages of such in-
stitutions should use them to the fullest extent if the proper balance
in the study and teaching of botany is to be maintained.

The old quarrel between the taxonomist and the morphologist is
based partly on a lack of sympathy due to ignorance; and one way to
restore friendly relations is to increase the breadth of view on both
sides. The taxonomic laboratory should really be an ‘‘assembly
room”’ for all kinds of information about plants, derived from the
geologist, the phytogeographer, the morphologist, and the physiologist.
No fact should be overlooked, no source of information should be
ignored.

The object of the taxonomist should be not only to arrange his
specimens in an orderly way in the herbarium, but to gain the greatest
possible knowledge concerning the species; using all the specimens at
hand, all notes from various collections, and all that has been written
about the species and its distribution, as well as its habitat, abundance,
variation, and economic bearing. Such a study is not easy, but in-
volves the highest mental processes. The weighing of all the facts
and observations regarding species and their systematic position often
taxes the best judgment. It is easy to collect specimens as a fad; it

THE TAXONOMY OF THE AGARICACEAE 317

is not difficult to know a great deal about the objects of nature; but
few are capable of pursuing the study of a group until it is rounded out
into a perfect and orderly whole, so that others may easily follow and
understand. When such a work is done, it is the simplest thing in
the world to criticize both the methods and results. It is the object
of all devoted nature lovers to add as much as possible to the sum
total of natural knowledge. No part of nature is too humble to
demand our best endeavors when it is considered in relation to the
numberless objects that make up the great universe, and it often
happens that what seems very small and unimportant may hold the
secret key to something that we look upon as vastly important.

So much for the taxonomist. Let us next see what the isolated
worker or professor in a small college can do toward broadening his
botanical outlook.

1. He can make sure that he knows by reading or otherwise some:
thing of the relatives and relationships of every form he studies mor-
phologically, cytologically, physiologically, or pathologically. This
would be a distinct advance. For example, a student investigating
the cytology of the scrub pine could learn something of the various
species of pines and their near relatives, and a student tracing the life
history of a certain species of Gymnosporangium could learn to recog-
nize another species of the same genus in case he happened upon it
accidentally. It is just as disgraceful for a morphologist not to know
the taxonomy of a type he is investigating as it is for a taxonomist not
to know why sap rises or the significance of reduction in chromosomes.

2. He can begin, if he never completes it, a local flora of his region,
including all groups of plants. Some groups offer excellent oppor-
tunities for field study even during the winter, while many of the lower
forms are less difficult than the flowering plants. By preserving
specimens, taking good notes, and securing photographs or drawings,
this work may be made really valuable. There is hardly a locality in
America that does not need careful botanical work of this kind. Here
in New York City, we have only made a beginning. With taxonomic
botanists becoming scarcer every year, it will be a long, long time before
we have any adequate knowledge of the flora of this great country.
At present, our books only emphasize our ignorance and the gaps in
our herbaria remind one of a child shedding his milk teeth. The
morphologist could do most of his taxonomic work while out for exer-
cise or on vacation, and it would only give him zest for his more special
problems.

318 WILLIAM A. MURRILL

3. He can have a compact and easily accessible synoptic collection
representing the principal groups of plants, with a few common and
characteristic species in each group, so that he and his friends or pupils
may obtain some idea of the extent and variety of the plant kingdom.
When one takes a tramp or travels, he sees plants, and often these
are about the only living things he does see. Now, it would seem
hardly fair to one’s intelligence to disregard all this wealth of inter-
esting material and be content to go through life both ignorant of it
and uninterested in it.

A speaking acquaintance with plants is obtained by observing
their form, color, habits, and relationships, rather than by dissecting
them. To begin with the microscope is to begin at the wrong end.
Let us not be too proud to know the names of the common flowers
that bloom at our feet. It is Nature’s way with the child; and it is
the logical method for the learner in any subject.

COLLECTING THE FLESHY FUNGI

The fleshy fungi, on account of their perishable nature, present
many difficulties to the collector, and I have found it almost im-
possible to secure good specimens without going after them myself.
Whenever and wherever fleshy fungi are collected, the following ideal
for the herbarium specimen should be kept in mind:

1. Ample typical specimens in all stages, well dried and well
preserved without pressing. Specimens in fluid have little value.

2. Full descriptive field notes, especially of perishable characters.

3. A colored sketch as accurate and detailed as possible and a
photograph if practicable. Color notes with a color guide at hand
may be used with the photograph.

In collecting, it is possible to attempt the entire group of fungi but
this is seldom successful. One gets accustomed to looking for certain
forms in certain places and it is quite difficult to train the eye to several
different sets of conditions. When one collects leaf-spot fungi, he
goes more or less into the open and carries a plant-press or vasculum;
but when one seeks fleshy fungi, he usually goes into the woods and
carries a basket. From the standpoint of collecting pure and simple,
the fungi may be divided into three classes: (1) those occurring on
leaves, in which the host is pressed just as in flowering plants, (2) those
occurring in various places, but requiring no special care in drying for
preservation. This class includes the molds, many of the ascomycetes,

THE TAXONOMY OF THE AGARICACEAE 319

all the lichens, practically all the polypores, most of the gasteromycetes,
all the Thelephoraceae, most of the Hydnaceae, and many of the
smaller and tougher gill-fungi, (3) those requiring the sun or artificial
heat for drying, this class including practically all the fleshy gill-fungi,
many of the larger fleshy cup-fungi, a few of the gasteromycetes, some
of the coral-fungi, hedgehog fungi, and polypores, and all the Boleta-
ceae, this last family being the most difficult for the collector.

It is important that fleshy specimens be allowed to dry as naturally
as possible, even if artificial heat is used, since they often assume
characteristic shapes in drying. The heat should never be strong
enough to cook the specimens, and they should not be pressed in order
to mount them in packets. A fresh specimen badly infested with
insects may be treated with naphthalene or chloroform in order to
prevent the destruction of the specimen before it can be dried. In
‘the mountains of Austria and Italy, where the air is unusually dry,
I found artificial heat rarely necessary. In the wilds of Maine and
Cuba, I used a special drying oven over a camp stove; at the Lake
Placid Club in the Adirondacks, a sunny, steam-heated room; in the
Catskills, a large open fireplace; in Oregon, a “‘biplane’’ made of window
screens suspended over a wood stove; in Virginia, a garret over the
kitchen stove; on the Vanderbilt estate in North Carolina, the ordinary
sunshine; and in Jamaica and Mexico, a drying oven over two tin oil
lamps, which were often kept going all night.

It sometimes happens that a botanist may know a distant region
better than the one in which he lives. When one goes away for a
definite object, he can devote his time to that object, while at home his
attention is often absorbed with numerous other interests. In my
own case, while I have been able to get a fair knowledge of the species
occurring in the immediate vicinity by getting up at an early hour in
the morning, I have given rather special attention to northern New
York, Maine, Virginia, North Carolina, Washington, California,
Mexico, Jamaica, Cuba, and many parts of Europe, the fungi brought
back from these regions amounting to over 30,000 herbarium specimens.

The mycological herbarium at the New York Botanical Garden
contains about 200,000 specimens, half of which were obtained from
Mr. J. B. Ellis, who at the time of his death had described more new
species than all other American mycologists together. The value of
such a collection can hardly be overestimated. From a purely botan-
ical standpoint, it is highly important that original and representative

320 WILLIAM A. MURRILL

specimens of all groups of plants be thus preserved for purposes of
reference and comparison; and, since questions of origin, distribution,
and variation always enter into studies of classification, it is desirable
to have these collections as complete as possible. From the standpoint
of applied botany, the vast number of destructive plant diseases caused
by fungi relate this subject very intimately with horticulture, agricul-
ture, forestry, and allied sciences.

Aside from the use of the collection by systematic botanists, plant
pathologists, and foresters, there is a large and increasing interest
manifested in fungi by the plant-loving public, drawn by fondness for
the queer and unknown, or attracted by bright colors and peculiar
forms, or by their extensive use as food. To all these, the collection
affords the keenest pleasure and offers opportunities for further knowl-
edge and enjoyment.

This collection of fungi is to be the basis of nine volumes of North
American Flora. As the various groups are worked over and new
species published, the number of type specimens in the herbarium will
be greatly increased. Students, collectors, and investigators through-
out the country will continue to send in specimens for determination
and comparison, and will come here to consult not only the ori-
ginals, but the array of additional specimens that show the. variation
and the geographical distribution of given species and groups of
species. .

It is hoped that important contributions may in time be made to
questions of geographical distribution on the basis of various collections
from distinct regions. For the purpose of recording the distribution
of species conveniently and quickly, a chart has been prepared,
copies of which may be properly marked and pasted on the inside of
the species covers, to show at a glance just where a particular species is
known to occur.

If one wishes to distinguish specimens from different regions in the
herbarium, he may use gummed paper markers of different colors on
the species covers, or simply indicate the regions by numbers or letters,
as shown in the following table:

I.. North, America... .\.2.Na?2 White WL alias: oe a ene. eee In. .Orange

Ii. ‘Tropical America... Ta..Red VII. China and Japan!) "Cj; 2 Yellow
Mil South America =. 52. .5an-Blue NATE Miataya tae nee Ma..Brown
IV. Europe and Siberia... .Es..Gray DX2 Australia; ee Au.. Pink

IVR ATTICA eae eee es ee Af. .Black X selislandsso a aa Is. ..Green

THE TAXONOMY OF THE AGARICACEAE 321

THE MECHANICAL SIDE OF THE HERBARIUM

The mechanical side of the fungous herbarium presents fully as
many difficulties as the preparation of specimens in the field. Some
of the specimens are small and flat, while others are large and bulky;
some are tough or woody, others are fragile; some may be poisoned
once for all, while others require constant attention to prevent their
destruction by insects.

The model fungous herbarium contains all the specimens of a group
in a single series and is at the same time neat and easily consulted.
Let us begin with the ordinary herbarium sheet to which is attached
envelopes of various sizes containing the specimens. It is never de-
sirable to leave the specimens exposed on the sheet as in flowering
plants, although this old method had obvious advantages. In case of
small specimens that might be lost, they are best enclosed in pill-boxes
or small elongated paper boxes, or in open cradles with cardboard
bottom and sides of cypress or cork strips attached with glue. Such
containers should always be placed within the envelopes before
mounting. A very convenient paper box is made with a loose cover
so as to avoid delay in opening it. The more specimens that are
fastened to sheets, the less trouble there will be.

For specimens too large to fasten in this way to sheets, boxes of
various sizes will have to be used and these should be either glued to
good cardboard, only one species to a card, or placed in a light wooden
tray that fits the pigeon-hole. In order to prevent the great waste of
time incident to examining a large amount of material and in opening
boxes or packets, a set of sample specimens may be mounted in
uniform boxes fastened to cardboard and covered with transparent
tops made of gelatin or celluloid. With a set of these samples,
hundreds of specimens may be consulted and compared in a few
minutes. Such an arrangement is peculiarly adapted to universities
and small herbaria where distribution is not so much an object as is
the determination of specimens. There still remains the odd lot of
boxes too few to mount on cardboard which can only be placed in a
wooden tray and listed on the outside of the tray. Such individual
representatives of different species cannot be distributed at once
through the herbarium, but must wait until additional material allows
the use of the cardboard.

The preservation of fungi against insects has always been a dif-
ficult problem for the curator. Many methods have been tried in

322 WILLIAM A. MURRILL

various herbaria without complete success. Carbon bisulfid has been
mainly used in this country, but the results are not satisfactory. Cor-
rosive sublimate, so extensively employed for flowering plants, is not
only valueless but decidedly harmful to many of the higher fungi,
since it alters or destroys their surface characters and often changes
their substance to a marked degree. It is much better to lose some
specimens than to have the whole collection thus altered. In the
case of large woody specimens, also, it is very difficult to secure suf-
ficient penetration to preserve the interior portions.

The substance I have used with great success is naphthalene flake,
of the best quality. Experiments conducted here have shown that
adult insects are killed in a few hours when placed in a box with this
substance, and it is probable that those emerging from the pupa stage
succumb in less time. Specimens are treated when first obtained, and
those peculiarly susceptible are kept in an atmosphere of naphthalene
more or less all the time. In going through the collections, when a
packet or box is found containing insects, a spoonful or more of naph-
thalene is added and the incident closed. Possibly there are insects
not yet acquired or some that do not thrive in this region that are not
amenable to this treatment, but it has been more satisfactory here so
far than any other method I have seen tried.

All fungi found upon leaves are treated with corrosive sublimate.
This is done chiefly to preserve the leaves intact, the fungi being so
small that, with few exceptions, insects would hardly do them serious
damage. All other fungi, particularly the conspicuous forms known as
mushrooms, bracket fungi, etc., are placed in boxes with naphthalene
flake for several weeks or longer, according to the season, before dis-
tributing them in the herbarium. Groups peculiarly liable to attack
are examined once or twice a year and fresh naphthalene added when
necessary. After a box collection has been once cleared of pests, it is
not so difficult to keep them out, with a fair amount of precaution and
vigilance.

THE ARRANGEMENT OF ILLUSTRATIONS

The ideal herbarium contains specimens arranged in a single series,
with all notes and illustrations classified with the specimens. The
maintenance of more than one series is both a complication and an
aggravation. However, throughout most of Europe, illustrations
are kept in a separate series, just as exsiccati usually are; and I might
remark that both exsiccati and illustrations should be in duplicate so

THE TAXONOMY OF THE AGARICACEAE 323

that one set may be distributed with the specimens and the other set
kept on file for ready reference. At Kew, a splendid set of portfolios
has recently been made to hold the large number of colored illustrations
made by Cooke, Massee, and others. In the Fries herbarium at
Upsala, the drawings are mounted on cardboard and kept in a separate
case. The advantage of having colored illustrations readily available
when fresh specimens are brought in for comparison appears at once,
since characters are then used for determination which disappear when
the specimens are dried. Herbarium specimens are rarely consulted
for comparison except with dried material. It may also be desirable
to know whether there exists in the collection a good illustration of the
specimen in question so that steps may be taken to fill the gap as new
collections are brought in. It has been decided to adopt the following
arrangement with our collection of fleshy fungi:

I. Keep a set of colored drawings and photographs convenient for
ready reference.

2. Keep all other drawings, such as those of sections, spores, etc.,
with the specimers in the herbarium; and prepare a duplicate set of
photographs and colored drawings for the herbarium whenever
practicable.

~Water-color paintings should be kept in a perfectly dry, dark
place. Naphthalene, camphor, and carbon bisulfid are not particu-
larly harmful to water-colors, but sulfur dioxid, hydrogen sulfid, and
fumes of ammonia or acids should be carefully guarded against. The
colors used should be the best and most permanent on the market, and
each color should be actually tested by the artist if possible before it
is used.

o

THE NEED OF AN AMERICAN ILLUSTRATED WORK

While on the subject of color, I wish to remark that one of the
greatest needs of mycology in this country is a comprehensive illus-
trated work on the larger fungi. The various countries of Europe are
well supplied with such works, some of them quite old and very elabor-
ate. Had it not been for these books, the work of many mycologists
would have been practically lost or left in such a state as to be more
or less useless. America has nothing to compare with any of the il-
lustrated works on fungi in Europe. The)need of such a work is fully
realized by all; but it would require not only a well-equipped her-
barium,and library, but also a considerable amount of money, probably

324 WILLIAM A. MURRILL

$50,000, to carry out such an undertaking. Artists would have to
paint the plants in the fresh condition where they grow and this would
necessitate reaching various parts of the country during the growing
season, although a large beginning might be made in any given locality.
The plates should be prepared and reproduced in the best possible
manner, and accompanied by accurate and comprehensive descriptive
text. Sucha work would be useful to the forester who wishes to protect
his trees from wood-destroying fungi, to the collector of edible mush-
rooms who wishes to use them for food and to guard against poisonous
species, to the student in college or university, and to the general
nature-lover in whatever part of the country he might live. There is
nothing that would give a greater impetus to the study of fungi in all
parts of North America than the publication of such a great illustrated
work.
THE CLASSIFICATION OF THE GILL-FUNGI

Coming now to a discussion of the taxonomy of the Agaricaceae in
a more limited sense, the history of mycology in Europe shows that
some of the older men, like Schaeffer and Bulliard, devoted their
attention to describing and illustrating species and thought very
little about genera; while others, like Persoon, Roussel, Gray, and
Fries, attempted to improve upon the rather primitive divisions of
Linnaeus. Then came the general adoption of the Friesian classi-
fication through the publication and wide distribution of the Systema
and Saccardo’s Sylloge, followed by demands for improvement from
Quélet and Patouillard in France, Karsten in Finland, Hennings in
Germany, and Underwood and Earle in America.

When one travels from country to country and from one herbarium
to another, he gets accustomed to changing his nomenclature as he
does his language and his money. The claim of “existing usage”’
put forward by some has no value whatever unless it refers to the
names used in Saccardo’s work, which is merely a convenient, though
disorderly, compilation of species as they are published, arranged
according to a system in vogue when the work was started many
decades ago. Karsten, a pupil of Fries, questioned the latter’s classi-
fication because based on too few characters. Patouillard and Fayod
considered microscopic characters of great importance, while Maire
goes so far as to include cytological characters.

What we need in America -is a classification that is impartial,
practical, and modern, including all the improvements possible and

THE TAXONOMY OF THE AGARICACEAE 325

based on the study of American rather than European material. We
want no “pounds, shillings, and pence’’ in the form of cumbersome
subgenera, subsections, and subspecies. The taxonomist must know
but not recognize varieties, which are essential to the gardener, the
physiologist, and the plant breeder. Let the grouping be as natural
as possible, but artificial when conducive to clearness. The absence
of sex in the gill-fungi gives one considerable liberty and our knowledge
is still far from complete. A system of classification representing the
genetic relationships of the higher fungi is hardly yet in sight. If the
species could be collected and grown together under cultivation, the
glad day might be hastened, but they cannot. Every house has a
garret; so has every family a genus or two which catch everything not
wanted elsewhere. Do not be too particular with the misfits, but do
not throw them out of the window; they may fit in when you move to
the next house. |

In seeking suitable characters for classification, one must use what
comes to hand, and the same characters may not be available for
different groups. The best and most constant primary character for
the gill-fungi seems to be the color of the spores. Earle tried to use
the “partial veil’’ but failed on account of its evanescent and variable
character. The form and surface markings of the spores may be
quite characteristic, as in Entoloma and Inocybe, but other good
characters should be associated with them. I believe that Patouillard
goes too far with microscopic characters, and, moreover, that these
should be used in keys as little as possible, in order to save time and
trouble. A key character need not necessarily be the most important
generic character, but only the most convenient.

Recent researches on color in the flowering plants have shown
this character to hang on a very slender thread sometimes in that
group, but what would we do in the fleshy fungi without the recog-
nition of color? Poisonous properties alone would hardly seem to be a
sufficient basis for the separation of species, and still there might be
a good practical reason why they should be recognized in certain
instances. I have in mind the variations in the poisonous properties
of Venenarius muscarius, V. pantherinus, and Chlorophyllum molybdites,
and the separation of Clitocybe sudorifica, a poisonous species or
variety, from Clitocybe dealbata, generally considered harmless but
not differing morphologically from C. sudorifica. Ordinarily, physi-
ological properties would seem to have no taxonomic standing, but

326 WILLIAM A. MURRILL

they might be suggestive and lead to more careful morphologic
research.

The many changes made in generic and specific names are to be
deplored, but they are unavoidable. As already intimated, the systems
of classification in vogue in Europe were not in harmony and were
based on different conceptions from ours, so that they had to be
worked over and adapted to our needs. The American code of
nomenclature adopted for North American Flora over a decade ago
has been found to work remarkably well and we see no reason to change
it, even if such a thing were possible, for the set of compromise rules
recently formulated which will never be consistently followed anywhere
in the world. People ask me why I take up Melanoleuca for dear old
Tricholoma, not knowing that Bentham used Tricholoma for a genus
of flowering plants as early as 1820. ‘They say it is a shame to discard
Amanita and use Venenarius for our most poisonous mushrooms, little
dreaming that in the long ago Amanita and Agaricus meant the same
thing and we could not keep them both. It is not my fault that the
old fellows did their work so poorly and with such a delightful dis-
regard of priority rights.

Neither is it my fault that American material has been so poorly
determined by European mycologists. They have no more interest
in America than we have in the Fiji Islands or in Timbuctoo, and
when they receive our specimens they are very apt to be reminded of
a similar European species and be satisfied with that. Then, there is
the great difficulty in studying dried specimens of fleshy fungi unless
one has seen them in the fresh state. Specimens lose something in
drying that can never be replaced. That is why I have often sat up
half the night over the drying oven when the hunting was good in one
of those far-off, wild, and virgin forests ‘‘somewhere”’ in North America
or Europe.

I wish now to bring to your attention the system of classification
I am using for the gill-fungi. Much time might be devoted to the
grouping, the characters, and the descriptive terms employed, but a
prolonged discussion of these details would only weary you. I prefer
rather to outline briefly the main groups of this family and to illustrate
them with colored slides of some of the more common and interesting
species.*

NEW YORK BOTANICAL GARDEN.

* At the conclusion of the paper, lantern slides were used to illustrate the classi-
fication of the gill-fungi.

©OBSERVA EIONS ON, PORES E TREE -RUSIS

JAMES R. WEIR AND ERNEST E. HUBERT

The undertaking of the checking by cultural methods of various
forest-tree rusts occurring in the Northwest has established several
host relationships previously held doubtful. The recent works of
Fraser! and Ludwig? have aided in the clearing of some of the problems
concerned. Fraser’s results with species of Uredinopsis on ferns and
the final conclusion to the effect that the five species of Uredinopsis
used in his experiments have their aecial stage on Abies balsamea is an
important contribution toward a clearer understanding of the inter-
esting group of rusts occurring on ferns. The five species with which
Fraser worked are Uredinopsis struthiopteridis St6rmer, U. osmundae
Magn., U. atkinsoni Magn., U. mirabilis Magn., and U. phegopteridis
Arth., and a study of the microscopical characters reveals no great
differences between them. Fraser* in his last article came to the
conclusion that all of the five species with which he had been working
were identical with the exception of U. mirabilis and considered this
one different on account of the fact that positive results with aecio-
spores from Abies balsamea were secured on Onoclea sensibilis only.
In a recent communication received from Fraser, March 27, 1916, he
states that Arthur examined all the field collections of Peridermium
balsameum Pk. and cultures and came to the conclusion that there are
no morphological differences in the aecial stages produced on Abies
balsamea by inoculations of the five species of Uredinopsis. <A close
comparison of the spore measurements and lengths of beaks of the five
species as published by Arthur* show no great differences in size from
what might be expected as a result of the influences of the various

1 Fraser, W.P. (a) Cultures of Heteroecious Rusts. Mycologia, 4:175. 1912.

(6) Further Cultures of Heteroecious Rusts. Mycologia, 5: 233. 1913.

(c) Notes on Uredinopsis mirabilis and Other Rusts. Mycologia 6:25. 1914.

2 Ludwig, C. A. Notes on Some North American Rusts with Caeoma-like Sori.
Phytopathology 5: 273, I915.

3 Fraser, W. P. Notes on Uredinopsis mirabilis and Other Rusts. Mycologia

Os.252 191d:
2 Arthur,-}eo2, Wredinales, -N- Amer: Blora’y:.005:, 1007.

327

328 JAMES R. WEIR AND ERNEST E. HUBERT

fern hosts. The variations in the spore markings are negligible. In
view of the results secured by Fraser and the determinations by
Arthur, it is suggested that these five species be combined under the
name of one species of Uredinopsis. The aecial stages of all five
species have been found to be identical with Peridermium balsameum
and the close similarity of P. pseudo-balsameum (D. & H.) Arth. with
P. balsameum has led us to consider them here as one species, namely,
Peridermium balsameum. ‘The differences in the description of the
two species do not seem to be of sufficient importance to continue their
separation. The description given by Arthur and Kern® give a
slightly larger spore for P. balsameum and no mention is made of the
color of the spores of P. pseudo-balsameum which are colorless as in the
other species. The peridia of both are fairly long (0.75-1 mm.) and
with the colorless spores furnish excellent means of identification. All
stages of this rust should therefore be referred to one species of Ure-
dinopsis. The aecial stage was not only found in abundance this
season on Abies grandis but also on A. lasitocarpa. This fungus has
been collected in the Northwest in 1896 under the name of P. pseudo-
balsameum (D. & H.) Arth. Hedgcock® reports P. pseudo-balsameum
on Abies grandis, A. lastocarpa, and A. nobilis in 1912. In a recent
article by Schmitz’ a claim is made to the first collection of the fungus
west of the Mississippi Valley. A glance at the literature® will show
that Peridermium balsameum was collected on Abies grandis on the
slope of Mt. Paddo, Wash., by W. N. Suksdorf in October, 1903.
P. balsameum was collected on Abies grandis in California in 1896.
Many other collections of this fungus have been made since then and
the collections of this laboratory at Missoula show considerable
material (eight collections) from the northwestern states collected
during the years I9II to 1916.

UREDINOPSIS PTERIDIS

During the past two seasons a very interesting Peridermium has
been collected on the needles of Abies grandis (fig. 1). This fungus is

5 Arthur, J. C., and Kern, F. D. Uredinales. .N. Amer. Flora-7:115.) 1007:

6 Hedgcock, G. G. Notes on Some Western Uredineae Which Attack Forest
Trees. Mycologia 4: 141. 1012:

’Schmitz, H. Preliminary Note on the Occurrence of Peridermium balsameum
in Washington. Phytopathology 6: 369. 1916.

8 Arthur, J. C., and Kern, F. D. North American Species of Peridermium.
Bull. Torrey Club 33: 403. 1906.

OBSERVATIONS ON FOREST TREE RUSTS 329

conspicuous by its appearance on the second year needles of its host,
by its white aeciospores and unusually long peridia. Most of the
other needle rusts occurring on conifers occupy the needles of the
season and this fact is accounted for by the overwintering of the telial

LETTE E TT
ae

eee
4

|
6

eee
on
ee
<n
~
foment
ene
enna
—
inne
soni
ane
ee
even,
sme,
—
—
anerensenrpsensmeattt
ene
cee,
—_—
pony
cameo
—_—
—_
po
EE

S

L111

a

HEE T

SYSTEM ©

Fic. 1. Uredinopsis pteridis. I stage on Abies grandis showing appearance

upon second year needles.

stage. The infection of these Peridermia is produced in the spring on

the youngest needles and the fungi mature the same year. This

Peridermium has its alternate stages upon Pteridium aquilinum

bubescens Underw. (fig. 2) and the conclusion is that the telial stage
does not winter over as is the common habit of such rusts. The needles

330 JAMES R. WEIR AND ERNEST E. HUBERT

of the fir must become infected during the same summer or fall in
which the telia mature on the fern. Such being the case the advent
of lower temperatures prevents the fruiting on the needles until
favorable conditions are again present, which is in the early spring of the
following year. This is borne out by the collection of the aecial stage
fully matured on Abies grandis as early as April 12, 1916. ‘This is the
earliest collection of any needle rust made in this locality. Other
collections were made on April 14, May 2, and June 17, of the same
year.

On June I9, 1916, sowings of aeciospores of the above fungus on
Abies grandis were made on two plants of Pteridium aquilinum pu-
bescens. The plants were raised in the greenhouse at Missoula, Mont.,
from rhizomes dug up in the field on September 4, 1915, and the inocu-
lations were made by the use of celluloid cylinders and cotton plugs.
On July 25, 1916, a medium infection of uredinia was found on one of
the plants while the other bore no results. The control plants re-
mained normal. A large number (15) of collections of the fungus on
Pteridium aquilinum pubescens (fig. 2) made throughout Idaho,
Washington, and Oregon was always in immediate association with
the rust on the needles of Abies. In one particular instance at Lucern
Lake, Wash., August 23, 1916, a lake flat was grown up to young
Abies grandis and the braken fern. The foliage of the latter was
completely parasitized by Uredinopsis pteridis D. & H. while the
needles of the fir were seriously infected with the aecial form of the
fungus. No other forest tree rust was present in the vicinity. After
a close comparison of the microscopical characters of the above
produced uredinial stage with authentic material of Uredinopsis
mirabilis, U. osmundae, U. struthiopteridis, U. atkinsoniu, U. phegop-
teridis, and U. pteridis, it was found to coincide with the latter. A
careful study of the published descriptions? of the species of Uredinopsis
in connection with the above culture showed that no great differences
existed between U. copelandi Sydow and U. pteridis other than the
hosts. A slight difference in the size of the spores is to be noted.
U. pteridis has a spore measurement of 11-18 by 30-58 uw and that of
U. copelandi is 14-18 by 31—40 yu, a difference in length of about 19 uy.
This variation is no greater than is found usually occurring in spores
of a single species. In comparing the five species of Uredinopsis
which have been found to produce an identical aecial stage on Abies

® Arthur, J. C.. Uredinales. _N. Amer. Flora 7: 115. . 1907.

OBSERVATIONS ON FOREST TREE RUSTS 331

balsamea, it is found that the spore measurements vary to as great an
extent_as 19. It was also found that the spores of the above five
species bore a similarity in the length and shape of the beaks, these

=

| ony
7 8

6

I

eis
cabana
pres
a eG)
TEx
oe
oy
cae
Sane

4

||

Poets

HTT

METRIC hee fe

WUT

Fic. 2. Uredinopsis pteridis. 11 stage of Pteridium aquilinum pubescens. Note
the coiled spore masses.

varying in length from 3-10 u to 12-26 » with almost similar spore
markings. The other two remaining species of Uredinopsis, U.
pleridis and U. copelandi, both have spores with short stout beaks

832 JAMES R. WEIR AND ERNEST E. HUBERT

measuring 3-74 long. The spore markings of these two are very
similar and are very pronounced when compared to the other five
species of the genus. Judging from this it seems that no great reliance
in respect to identity of species can be placed upon spore measure-
ments alone. |

It is suggested, in view of the above comparison, although cultures
are necessary for final determination, that U. copelandi be considered
identical with U. pteridis and placed under the latter species. A
technical description of the aecial stage of U. pteridis follows:

o. Pycnia not found.

I. Aecia from a limited mycelium appearing on second year
needles, hypophyllous, not crowded, forming rows on either side of the
midrib, cylindrical, 0.2 to 0.4 mm. across and 1.5 to 2.6 mm. high;

peridium colorless, rather delicate, rupturing at apex with fringed_

margins; cells overlapping, majority rhomboid, (10) 15.0-26.6 by
32.0-43.7 wu, inner walls coarsely and closely verucose, not striate,

8.3 to 10.0 w thick including tubercles, slightly thicker at one end of.

cell, outer wall, 6.5 to 7.5 uw thick, smooth; aeciospores mostly globoid,
occasionally broadly ellipsoid (50) 13.3-20 by 18.3-24.1 w, standard
(19 by 22 yw), wall colorless, 2 to 2.5 uw thick, coarsely and closely veru-
cose, contents colorless.

On living needles of Abies grandis and A. lastocarpa from early
spring to late fall depending upon elevation.

THE OVERWINTERING OF RUSTS

It has long been a puzzle as to why the Pucciniastrum occurring
on species of Epilobium other than E. angustifolium (L.) Scop, has
not been found to have its alternate stage upon species of Abies.
Successful inoculations of Abies sp. with teliosporic material of P.
pbustulatum on E. angustifolium have been made in Europe and in
America. Fraser!® secured results in 1912 upon Abies balsamea and
has collected the aecia in the field. Check results have been secured
by the writers!! in 1916. All experiments were properly controlled.

The uredinial stage of a Pucciniastrum has often been collected
in the northwestern states upon EF. adenocaulon Haus. but never the
telial stage. Examination of available exsiccati material fails to

10 Fraser, W. P. Cultures of Heteroecious Rusts. Mycologia 4: 175. 1912.

1 Weir, J. R., and Hubert, E. E. A Successful Inoculation of Abies lastocarpa
with Pucciniastrum pustulatum. Phytopathology 6: 373. 1916.

OBSERVATIONS ON FOREST TREE RUSTS 333

disclose any stage of P. pustulatum upon E. adenocaulon other than
the uredinial stage (N. A. U. Nos. 77 and 1087, Fungi Col. Nos. 2575,
2782, 3180, 3773, and 4334, Fungi Dakotensis No. 371, and Jackson’s
Col. No. 1488). Many collections of the II stage have been made by
the writers in months of the year which appear very much out of
season for this stage of the rust as the following dates show: March 20,
1916, April 12, 1916, May 4, 1916, June 3 and 15, 1916, July 9, and
Weeiors, August 4- and 720, 1015, September 2, 10,.and 28, 1916,
October 11, 21, and 28, 1916, November 14, 1916. ‘This indicates a
continuation of the uredinial stage throughout the entire year. Ex-
amination of all local and acquired collections fails to show where a
single collection of the telial stage has been made. On October 18,
1916, three rosettes of Epilobium adenocaulon were secured in the field
and potted in the greenhouse. Two of the rosettes bore the uredinial
stage of Pucciniastrum pustulatum on such portions of the leaves as
were protected by the outer rosette leaves. All the leaves of the
infected rosettes were cut off, care being taken to remove all rusted
areas and to cut back the leaves as close to their bases as possible. The
only two sources of infection remaining open to the oncoming leaves
were the very few urediniospores and the possible mycelium in the
portions of the leaves left on the plants. New leaves gradually
appeared and on November 1 several of them bore the uredinial stage
of the rust. A large number of spores were liberated by these few
infections. Germination tests of the spores showed a large percent
germinating. A few days later spores were collected which had
germinated in situ on the rosettes and produced a small mat of my-
celium. Examination of the portions of leaves left on the rosettes
after cutting off the infected leaves showed that considerable mycelium
was present in the cells of the mesophyll just beneath the epidermis.
From November 15 to 22 the rosettes developed a few leaves from two
to two and one half inches from the ground, indicating a departure
from the strictly rosette habit due to the temperature of the green-
house. The lowermost ones developed uredinia in abundance. The
uppermost leaves as yet showing no infection were sprinkled with
urediniospores taken from the pustules beneath on November 18,
1916. On November 27 uredinia developed on the leaves thus inocu-
lated. Two control plants remained normal. The preceding data
indicate the presence of a biological species of P. pustulatum occurring
on E. adenocaulon and overwintering by means of mycelium and

334 JAMES R. WEIR AND ERNEST E. HUBERT

uredinia upon the rosettes which continue living until spring. The
rust is carried over principally by means of urediniospores which
reinfect the leaves of the rosette and continue throughout the year
infecting the leaves of the flowering stalk during the spring and
summer. The fact that this form of the rust on E. adenocaulon
produces no telia is evidence of its continuation in the uredinial stage
and also explains the absence of a corresponding aecial stage upon
Abies.

P. pustulatum occurring upon F. angustifolium produces telia
which are capable of infecting species of Abies. This plant develops
from perennial horizontal root stalks. Rosettes which overwinter
are not produced and no evidence has been found to indicate any
stage of the rust overwintering on the living plant.

Studies have also been made upon Coleosporium solidaginis (Schw.)
Thum. occurring upon species of Aster and Solidago.

On October 18 four pots containing rosettes of Aster spp., 2 of
Aster conspicuus Lindl. and 2 of A. laevis-gayert Grey, infected with
the uredinial stage of C. solidaginis, were placed in the greenhouse at
Missoula, Mont. All of the leaves of the rosettes were removed and
the chances for infection depended entirely upon such few uredinio-
spores as had become transferred from the infected leaves. The rust
had been mature for some time previous to placing in the greenhouse.
On October 28, uredinia appeared on such leaves or portions of leaves
as were then present. From this date on other leaves as they appeared
became infected and developed scattered groups of uredinia. Four
control plants remained normal. Collections of this stage of the rust
upon species of Aster and Solidago have been made during the months
of the year when only the rosettes of the plants were in evidence.
Most of these collections were made in late winter or in early spring
before the snow had left the ground. Mains” in his article on the
overwintering of Coleosporium solidaginis produces very good evidence
of the overwintering habit of this rust on rosettes of Solidago sp. The
collections in Idaho and Montana of infected rosettes of Solidago
missouriensis Nutt. and S. canadensis L. during the months of March
and April before the peridia of the aecial stage on Pinus contorta have
appeared confirms the conclusions of Mains as to the wintering habit
of this fungus.

2 Mains, E. B. The Wintering of Coleosporium solidaginis. Phytopathology
62271... TOL:

OBSERVATIONS ON FOREST TREE RUSTS 335

The overwintering of a fungus such as C. solidaginis on Aster and
Solidago spp. when developing in regions so far removed from the
alternate host Pinus as to be too remote for infection by spores carried
by the wind is a question which has remained unanswered for some
time. Clinton™ refers to this problem in 1907 and comes to the con-
clusion that the rust winters over in the rosettes principally by means
of the urediniospores. A more recent article by Ludwig" gives some
very substantial evidence leading to his belief that the uredinial stage
of C. solidaginis on Aster, Solidago, and other hosts propagates itself
through the winter upon the rosettes principally by means of uredinio-
spores. He concludes that the evidence is in favor of the rusts being
able to maintain a high degree of vigor for a long period without
sexual reproduction.

OFFICE OF INVESTIGATIONS IN FOREST PATHOLOGY,

BUREAU OF PLANT INDUSTRY,
MissouLa, Mont. *

43Clinton, G. P. Heteroecious Rusts of Connecticut Having a Peridermium.
for Their Aecial Stage. Report of the Station Botanist 1907: 369.

4 Ludwig, C. A. Continuous Rust Propagation without Sexual Reproduction..
Proc. Indiana Acad. Sci. 1914: 219.

ENDOUTHIA] PIGMENTS:

Lon A. HAWKINS AND NEIL E. STEVENS

The genus Endothia is characterized by a yellow or orange stroma
and all known species produce a yellow or buff color in the mycelium
and upper layers of the substratum when grown on starchy culture
media. In connection with cultural studies of this genus Shear and
Stevens? first called attention to the fact that certain species when
grown on cornmeal or other starchy media produced a bright color,
“perilla purple,’’ while the others produce no such color.

Continued study enabled them to divide the genus on the basis of
this color production. ‘This division does not, however, coincide with
the classification based on morphology. On the basis of spore form
the genus is arranged as follows?

Section 1.—Ascospores short—cylindrical to allantoid, continuous or
pseudo-septate.
E. gyrosa (Schw.) Fr.
FE. sinsularis (syd. Syd.) Sco:
Section 2.—Ascospores oblong—fusiform to oblong-ellipsoid, uniseptate
when mature.
E. fluens (Sow.) S. & S.
E. fluens mississippiensis S. & S..
E. longirostris Earle.
E. tropicalis S. & S.
E. parasitica (Murr.) And. & And.

Of these species the first three uniformly produce perilla purple on
such media as cornmeal, oatmeal or rice flasks while the others have
consistently failed to produce this color. It is noteworthy that
E. fluens is included in the group which produces the purple color
while E. parasitica is not. These two species are so similar morpho-
logically that at one time leading mycologists considered them iden-

1 Published by permission of the Secretary of Agriculture.

2 Shear, C. L., and Stevens, Neil E. Cultural characters of the chestnut-blight
fungus and its near relatives. U.S. Dept. Agr. Bur. Pl. Ind. Circ. 131: 3-18, 1913.

3 Shear, C. L., Stevens, Neil E., and Tiller, Ruby J. LEndothia parasitica and
related species. U.S. Dept Agr. Bull. 380. 1917.

336

ENDOTHIA PIGMENTS 337

tical. They are found within the same areas, the United States and
Japan, and on the same hosts, Castanea sp. Yet £. fluens is a sapro-
phyte while E. parasitica is one of the most uniformly destructive
fungous parasites known. This is perhaps the only case yet recorded
of two closely related fungi, growing on the same host, one of which
is a virulent parasite and the other a saprophyte. The two species
grow readily and can be easily distinguished on artificial culture
media. Itis obvious that cultural or physiological differences between
E. fluens and E. parasitica are of great interest.

It was to study the production of the various colors in species of
the genus that the work described in the present paper was taken up.
Some attention has been paid to the coloring matter produced by
E. parasitica. Pantanelli*t considers the pigment to be a lipochrome
but records no experimental work in proof of this statement. Ander-
son’ disagrees with Pantanelli on this point. He considers the pigment
to be an aurine and quotes unpublished work by Mr. C. T. Thomas to
substantiate his view. It was hoped in the present investigation to
obtain more evidence on this disputed point.

In taking up the study of the pigments various solvents were tried
to see which was most favorable for the extraction of the pigment
from the mycelium and the mass of rice upon which the fungi were
grown. It was found that the coloring matter of all the species was
soluble in ethyl alcohol, and a considerable portion of it readily
soluble in ether. Accordingly extracts were made of the culture
media and mycelium, with alcohol, at room temperature. The
alcohol was evaporated and the residue extracted with ether. The
ether extract was then filtered, the ether distilled and the pigments
taken up in alcohol again. All tests were made in alcoholic solution
unless otherwise noted. The coloring matter was found to be yellow
when acidified with either hydrochloric, sulphuric, nitric, phosphoric,
or acetic acids. When the acid solution was treated with dilute
alkali, sodium, potassium or ammonium hydroxides, or sodium or
potassium carbonates, it became a deep red. Apparently all the fungi
elaborated pigments which were bright yellow when acidified and red
when made alkaline. While the alcoholic extracts from all the fungi

* Pantanelli, E. Sul parassitismo di Diaporthe parasitica Murr. per il Castagno.

Rendiconti della R. Accademia dei Lincei, Classe di Scienze, Fisiche, Matematiche
e Naturali. V. 20: 366-372, IgII.

® Anderson, P.J. The morphology and life history of the chestnut-blight fungus.
Bull. Penn. Chestnut Tree Blight Comm. 7: 1-43, 1913.

338 LON A. HAWKINS AND NEIL E. STEVENS

reacted in the same general way, there were various nuances of red in
the extracts from the different fungi.

A study of the alcoholic extracts from pure cultures on rice® of
E. parasitica, E. fluens, E. fluens mississippiensis, E. tropicalis, E.
gyrosa, E. singularis, and E. longirostris was made with a spectro-

ae Cc Te al elle ala

Sp) LL eee

80

SUBBEBREGSr sauceecos
SRR AER eee) SEER RSe
EERE CES MRP 2egeae eee
ARR Ree Ree eScsS

RRB) Ronse
‘oo

TACO LAIN —PER CLNT
iS)
iS eee
ee ea) ae
i ne | ee Se SS a
oe eee one
ee Ee ee eS

es poet seas aa ae |S
PEI SS es a es ee]

a
Soe aa a
ia a

BEB AERE OGL

690 660 6&0 700 720730

BI

ie

Pai
Ve aS ed
Ns
Let
BeBe. S|

el
Se

Go 420 440 460 4§G0O 500 520

Fic. 1. Curves of percentage of spectral transmission of acidified alcoholic
extracts from pure cultures on rice of E. parasitica (P), E. longirostris (L), and
E. fluens mississtppiensis (M). The curves were plotted with the percentage of
light transmitted as ordinates and wave-lengths of light in uu as abscissae.

photometer. With this apparatus measurements were made from
which the percentage of light of various wave-lengths transmitted by
the solution was calculated. These data were used in plotting the
curves of spectral transmission.’ The alcoholic extracts of stromata of
E. singulans from chaparral oak, EF. gyrosa, from beech, and E£.
parasitica from chestnut were prepared by separating the stromata

6 Throughout this study the fungi were grown on rice flasks prepared according
to the method published by Shear and Stevens. Loc. cit., p. 13.

7 This part of the work was made possible through the kindness of Mr. C. G.
Peters, of the Bureau of Standards, who made the measurements and calculations.

ENDOTHIA PIGMENTS 339

from the bark and extracting the mass with alcohol. The curves of
spectral transmission of these extracts are included in figures 3 and 5.

From the curves shown in figures I to 3 it is noticeable that the
transmission spectra of the acidified alcoholic extracts of pure cultures
of the species of Endothia studied group themselves into three classes.

pSECTRAL 7RANSM/SSION

1S Sl ee ee
30
8&0
K 70
; Ff fb
& 60 ‘
a’, a
a oer ae
N) / |
ae a ee eS Pa ee ea
x C2 ae Sea eae eee
S20 ACES seas eee
f AChR sank ana
_ 32 Aaa ae Se aes
Pe ae ee CFE iS occ eee

2 SeRae esas ese

Q
HOAZ0 440 460 480 400 520 $40 560 580 600 620 640 660 680 700 720730

Fic. 2. Curves of percentage of spectral transmission of acidified alccholic
extracts from pure cultures on rice of E. fluens (F), E. gyrosa (G), and E. singularis
(S). The curves were plotted with the percentage ot light transmitted as ordinates
and the wave-lengths of light in py as abscissae.

The first of these includes E. parasitica, E. longirostris, and E. fluens
mississippiensis. The curves for these three fungi agree rather closely
in most cases, the region of greatest variation being in the shorter
wave lengths transmitted. The distinctive feature of these curves is
that they indicate that practically all the violet rays are absorbed.
Only a small portion of the blue is transmitted while most of the
yellow, orange and red rays pass through.

The curves of spectral-transmission for EZ. fluens, E. singularis, and
E. gyrosa make up the second group and are shown in figure 2. An
inspection shows that the greatest variation in these curves is again in

340 LON A. HAWKINS AND NEIL E. STEVENS

the shorter wave-lengths. There is some transmission of the violet
rays, more of the blue, and a gradual increase in the percentage trans-
mitted through the green and yellow to the orange. From this region
through the red the percentage of transmission is practically the same
for all wave lengths. The curves of this type are different from those
of group I in that more of the violet and blue are transmitted and
somewhat less of the yellow.

eens CTRAL 77 Gite ee

a

Fz as Ps eS es es a Oe

jap
ay,
4

300 $20 $40 560 $80 600 620 640 660 680 JOO 720730

TRANS/TISS/ON — PEP? CENT
ABER aS Sea ees

BEES SSR eases

(ERE NES. OSES See ee

5)

(7

Fic. 3. Curves of percentage of spectral transmission of acidified alcoholic
extracts of E. tropicalis (T) from pure culture on rice, E. parasitica (P) from stromata
from chestnut, E. gyrosa (G), from stromata from beech, and FE. singularis (S) from
stromata from chaparral oak. The curves were plotted with the percentage of trans~
mission of light as ordinates and the wave-lengths of light in py as abscissae.

Figure 3 shows the curves derived from the percentage of light of
the different wave-lengths transmitted by the alcholic extract of
E. tropicalis in pure culture and also the curves for the alcoholic extracts
of the stromata of E. singularis grown on oak, the.stromata of E.
gyrosa grown on beech and the stromata of E. parasitica grown on
chestnut. The curve obtained for the extract from £. tropicalis is
typical of the group. This fungus was grown on artificial culture

ENDOTHIA PIGMENTS 341

media and thus is the only one in this group directly comparable with
those of the other two groups.

The curves of this group are different from those of type I in that
more of the violet and blue are transmitted. They are separated
from those of group II by the fact that their minimum percentage of
transmission is in the blue. A higher percentage of wave-lengths
shorter than 460 wy and a slightly higher percentage of the yellow is
transmitted than in any of the others. The chief difference in these
three groups is in the degree of absorption of the shorter wave-lengths
of light by the solutions.

(See iu GEO Ges id }
b |E a

Eva
Sener BemEeo4
Peewee: memiarera
ESRB eRessceher Ay
_ Eee a ae
ace Sees eees aepeeee
Pamiei@e so Peer ert yr

TRANSYUSSION ~PER CLINT

ee ie es
- See Sees

Ee Pe a
ae att Eni

$0420 440 60 480 500 $20 $40 $60 580 me 620 640 660 680 700 720730

Fic. 4. Curves of percentage of spectral transmission of the alkaline alcoholic
extracts of pure cultures on rice of E. parasitica (P), E. fluens, (F), E. fluens mississip-
pisensis (M) and FE. singularis (S)._ The curves were plotted with the percentage of
light transmitted as ordinates and wave-lengths in uu as abscissae.

The curves derived from the pércentages of spectral transmission
of the alcoholic extracts were considerably different when these solu-
tions were made alkaline. They group themselves into two classes
and the curves within each class are more widely divergent than in
the acidified solutions. As in the acidified extracts one striking dif-

342 LON A. HAWKINS AND NEIL E. STEVENS

ference between these two groups is in the transmission of the shorter
wave-lengths. In figure 4, which shows the curves of spectral trans-
mission of E. parasitica, E. fluens, E. fluens mississippiensis, and FE.
singularis grown in pure culture, it is apparent that very little of the
light of a wave-length below 480 yy is transmitted. From this point
the percentages of transmission increase with increase in wave-length.

SPECTRAL TRANS/SUSSION
IF b lE D Ic

/00 ie | IB la
Sich SRR esaae RRR RAB BE
gol ESET ee oe SS
a Bee EBRERERRReee |
ee es ea
Pee Pe Pa)
PR RERRRRERe Ree E Ses |
PsP TADS TSP bed eT Te et
See ee Pee EREre aes. Ll
PS AE Te SO Sa ea
Dace Se See eae ann Dee Senn
Sle lated LAC | ee
SANT SE See 7 a2 ee
ist | ATS wo fo no
Case AeR REE Er eee Ree |
EEC NEEE See ee ee
GREER RRR ee ae Eee
ET AAD STE mace TTS STESTS caees TL
PL STs TTP eed ZED STES a] Aa PESTS SIE Tes a
RRERR REE eeA |

46 420 440 460 F§O 500 $20 540 560 580 600 620 640 660 6680 700. 720230

Fic. 5. Curves of percentage of spectral transmission of alkaline alcoholic
extracts of E. tropicalis (T) from pure cultures on rice, EF. parasitica (P), stromata
from chestnut, E. gyrosa (G), stromata from beech, and E. singularis (S) stromata
from chaparral oak. The curves were plotted with the percentage of light trans-
mitted as ordinates and wave-lengths of light in wu as abscissae.

The curve of the percentages of transmission of the alkaline extract
from the mycelium of EF. tropicalis in pure culture shown in figure 5
is an example of the other type of curve. With this solution about
52 percent of the light of a wave-length of 420 uu passed through and
60 percent of the wave length of 460 wu. This is the maximum trans-
mission of this solution.

The other three curves shown in figure 5 were derived from the
percentages of light transmitted by the alkaline alcoholic extracts from

ENDOTHIA PIGMENTS 343

E. gyrosa, E. singularis, and E. parasitica, stromata on beech, chaparral
oak, and chestnut respectively. These curves are the same type as
that obtained for E. tropicalis, but in the case of the extracts from
E. gyrosa and E. singularts considerably more light was transmitted
in the longer wave-lengths.

The chief difference between these two groups, as shown in figures
4 and 5 is in the percentage of violet and blue light transmitted by the
extracts.

In the foregoing pages it has been shown that there were three
types of curves of the percentage of spectral transmission in the acidi-
fied alcoholic extracts of these seven species of fungi. This fact
seemed to indicate the presence of several different pigments in the
cultures. The differences in spectral transmission might of course
_be due either to the presence of different pigments, or to the presence
of the same pigments in different proportions. Investigations were
accordinglv carried out to see if there was any common pigment and
to determine if possible the presence of other pigments. As it was
hardly possible to investigate carefully the pigments formed by all
the species, forms typical of the three groups were chosen for study.
Those selected were EF. parasitica as a type of group I, £. fluens from
group II and E. tropicalis from group III.

The fungi were grown on rice in Erlenmeyer flasks until the medium
was well covered with mycelium and showed a considerable amount of
color. This usually required from three to six weeks. The culture
medium and mycelium were then removed from the flasks and dried
at a temperature of about 60° C. The dried mass was ground, placed
in a percolating funnel and extracted with cold neutral alcohol. The
extract thus obtained was concentrated under reduced pressure until
nearly all the alcohol was driven off. The residue was washed into
a precipitation jar with water and from 10 to I5 volumes of water
added. A reddish yellow precipitate resulted. This precipitate was
collected on a filter and extracted with ether until the solvent came
through nearly colorless. It was noticeable that a residue always
remained on the filter and was especially large in the preparation
from FE. fluens. The ether extract was concentrated and four volumes
of petroleum ether added to it. A yellow amorphous precipitate of
pigment settled out. This pigment was separated by filtration, dried,
redissolved in ether and precipitated out with petroleum ether.
Pigments were obtained from all three fungi which, judged by their

344 LON A. HAWKINS AND NEIL E. STEVENS

appearance, their solubility in ether and alcohol and the fact that they
were practically insoluble in water, were very similar. No crystalline
compound was obtained by this method.

An acetyl derivative was prepared from this yellow pigment by
dissolving a quantity in acetic anhydride to which some anhydrous
sodium acetate had been added and boiling under a reflux condenser
for two and one half hours. It was allowed to cool and was poured
into a beaker of cold water and allowed to stand with frequent stirrings
for ten or twelve hours. The precipitate which had formed was
separated from the solution, dissolved in absolute alcohol and cleared
with animal charcoal. It crystallized from the alcoholic solution in
yellow needles. After recrystallizing three times the melting point
was determined and found to be between 240° and 243° C. uncorrected.
The acetyl derivatives were prepared from the yellow pigment from
all three fungi and were apparently identical. They had the same
appearance, solubility, and melting point. There was, however, a
considerable difference in the yield of acetyl derivative from the dif-
ferent pigments. The largest yield in proportion to the quantity of
pigment used was obtained from the pigment from E. tropicalis; the
smallest was from the yellow precipitate from E. parasitica.

The acetyl derivative was broken down and the original pigment
recovered by dissolving the crystals in concentrated acetic acid, with
heat, and then adding a drop or so of concentrated sulphuric acid and
warming again. Several volumes of water were then added and the
pigment was precipitated out. The mixture was filtered and then
washed with water and dissolved in alcohol. The acetyl derivatives
of this yellow pigment from all three species of Endothia were broken
down in the same way. The three alcoholic solutions were treated
with a number of common reagents and reacted in exactly the same
way in all cases. It is evident that the three species of Endothia
produce the same pigment when grown on rice flasks and, as these
three species are typical of the three groups mentioned earlier in this
paper, it is highly probable that all the species of Endothia studied
produce this pigment. This pigment will be designated pigment A
throughout the rest of this paper. It is soluble in acetic acid, ether,
carbon-tetrachloride, and a number of other organic solvents. It is
slightly soluble in petroleum ether. It dissolves with a green color in
sulphuric acid and in concentrated nitric acid. When acidified it is
yellow and is insoluble in water. When made alkaline it has a color

ENDOTHIA PIGMENTS 345

approaching crimson magenta. The dry pigment is soluble in dilute
aqueous solutions of sodium or potassium or ammonium hydroxides,
or sodium or potassium carbonates.

Fic. 6. Photomicrograph of crystals of pigment B on the mycelium of £.
fluens grown on rice. (The writers are indebted to Dr. Erwin F. Smith for this
photograph.)

As was mentioned earlier, not all pigments extracted from the
ground-up rice and mycelium with alcohol were readily soluble in
ether. The residue remaining after the ether extraction in the case
of the pigment mixture from E. fluens was considerable. The cultures
of this fungus on rice, as has been pointed out, show a brilliant purple
color in the medium while the mycelium on top is yellow. In the
present study a red pigment frequently crystallized out on the mycelium
of E. fluens, fig. 6. These crystals were not observed in cultures of
E. parasitica or E. tropicalis. It was noticed also that when the
concentrated alcoholic extract was treated with water the water was
a brilliant red. 7

These facts seemed to indicate that part of the pigment A was
alkaline and soluble in water and slightly soluble in ether, or that
some other pigment was formed by this species of Endothia along with
pigment A. The red watery solution obtained from the first precipi-
tation of the yellow pigment was evaporated to dryness and taken up
in hot dilute alcohol. The dark red solution thus obtained was con-

346 LON A. HAWKINS AND NEIL E. STEVENS

centrated in a desiccator and a red crystalline precipitate was formed.
The residue from the first precipitation of the pigment of E. fluens
after the extraction with ether was treated with alcohol. The greater
part of this residue dissolved. When this solution cooled a red crystal-
line precipitate was formed similar to that obtained from the watery
extract mentioned above. The crystals were red glistening plates and
were optically active. After two recrystallizations an acetyl derivative
was prepared according to the method already described. The com-
pound crystallized out of absolute alcohol in colorless needles. It was
recrystallized twice and dried. It melted at 196° to 197° C. uncor-
rected. A portion of the acetyl derivative was broken down with
sulphuric acid after the method followed with the acetyl derivative
of pigment A, and a pigment was obtained which had much the same
appearance as the original red pigment, which will be designated as
pigment B in this paper.

The properties separating pigment B from pigment A are: It
forms a different acetyl derivative, is only slightly soluble in ether,
insoluble in toluene, carbon-tetrachloride, petroleum ether, and con-
centrated nitric acid. It is soluble in water and may be crystallized
from a water or dilute alcohol solution. When a dilute solution is
treated with a drop of ferric chloride it becomes darker, assuming a
greenish raw-sienna color. The alcoholic solution when made slightly
alkaline closely approaches orange vermillion in color. Crystals of
the red pigment found on the mycelium of E. fluens were removed and
tested with various reagents. The reactions were apparently the
same as those just described.

Pigment B has not been discovered in cultures of either E. parasitica
or E. tropicalis grown on starchy culture media. It may be elaborated
by these two fungi, but if so it occurs in such small amounts as to
render detection exceedingly difficult.

It was mentioned in the description of the work with pigment A
that the yield of acetyl derivative obtained from the yellow precipitate
from the extract from E. parasitica was very small as compared to
the yield from a similar amount of the yellow pigment from F. tropicalis
or EF. fluens. It was also shown that the alcoholic solution of the
pigment from E. parasitica grown on rice has a considerably different
spectral transmission than that from FE. tropicalis. These consider-
ations made it seem quite possible that another pigment might be
present in the solution in addition to pigment A.

ENDOTHIA PIGMENTS 347

The alcoholic solution remaining after the acetyl derivative of the
pigment from E. parasitica had crystallized out and had been separated
was treated with about four volumes of water. A flocculent yellow
precipitate was formed. This was filtered off and washed with water.
The dry pigment was amorphous and of a bright yellow color. No
acetyl derivative was formed even on long boiling with acetic anhydride
and sodium acetate. When the pigment was dissolved in alcohol
and treated with dilute alkali, the color closely approached rose doré.
It is evidently another pigment and the properties which separate
it from pigment A are as follows: It does not form an acetyl derivative;
it has an entirely different color when treated with alkali; it is insoluble
in cold petroleum ether and dissolves in cold concentrated nitric acid
with an orange red color and is red when dissolved in cold sulphuric
acid. It is readily distinguished from pigment B by its solubility in
ether and nitric acid and insolubility in water and very dilute alcohol,
also by its appearance when dry and when in acid or alkaline solution.
This pigment, which will be referred to in this paper as pigment C is
also found in extracts from E. fluens. Its presence has not yet been
demonstrated in the extract from E. tropicalis.

From the experimental work just described it is evident that there
are at least three different pigments formed by species of this genus,
pigment A, apparently common to all species, pigment B found in
E. fluens and probably also in the other species having a similar
spectral transmission of the acid alcoholic solutions and pigment C
which is present in the two groups typified by EF. fluens and E. para-
sitica.

It is of course quite possible that these three pigments are closely
related chemically and may be derivatives of the same substance.
They are similar in many particulars. All three are composed of
carbon, hydrogen and oxygen. ‘That is, on incineration they leave no
ash and tests for nitrogen,’ phosphorus, sulphur and the halogens
showed that none of these elements were present.

The comparative solubilities of these three pigments are shown in
Table I. These tests were all at room temperature and indicate
whether or not the pigment is appreciably soluble in the reagent used.

From the table it is evident that these pigments, especially pig-
ments A and C, are readily soluble in a considerable number of organic

§ Fresenius, C. R. Quantitative Chemical Analysis. Cohn’s translation of
the sixth German edition, 2: 4-7. I9QII.

348

solvents.

LON A. HAWKINS AND NEIL E. STEVENS

They are not in all cases soluble in the same reagent,

however, and these points of difference are of use in distinguishing and

separating the pigments.
TABLE [|

Solubility of Pigments A, B and C, in Various Reagents at Room Temperature

Pigment
Reagent

| A | B c
ACELICLACIC 55.5305 cet gia eee Soluble | Soluble Soluble
Aoeticetheres\. ih =e rn | - if
ANGCCtONE ix, oes nee he eo | ; ms
Bthylialcohol ws fsencace le a | 4 At
Methyl alcohol....... ees nS | a o
Amyl-alcohol (normal)....... i a 2
|B ave) eerree enon see Ltiee | x Slightly soluble =
POlUeiG ncaa dees cer Gane ence | 4 Insoluble ii
BenzZ0le ea see a ee | Q Soluble i
Carbon-tetrachloride 2... | iN Insoluble ms
CHiGroLoninic ike sss hae oe | sn oe a
Carbon-bisulphid’:...2 o...| ‘ i Ff
Petroleum ether .4 o. 3.40. 04%.. | Slightly soluble : Insoluble
Sulpluntc acid ao 8 tae csi, | Soluble Soluble Soluble
Nitricsacid"(COnC))... ae | ui Insoluble s
Walco co ee i anette | Insoluble Soluble Insoluble

The statement by Pantanelli that the coloring matter of EL. para-
sitica is a lipochrome seems hardly to be corroborated by the evidence
brought out in these experiments. Lipochromes according to Zopf®
and Samuely!® break down readily when exposed to light and air.
They are soluble with a green color in concentrated sulphuric and nitric
acids. When saponified by boiling with dilute sodium hydroxide
they are insoluble in alcohol. According to Zopf they are soluble in
petroleum ether and insoluble in water.

The pigments obtained in this study did not break down when
exposed in solution to light and air. Solutions were kept in the
laboratory for more than a year without apparent deterioration.
Pigment A was the only pigment which gave a green color when dis-
solved in concentrated nitric or sulphuric acids. Pigment B was
insoluble in this last mentioned reagent. Boiling in a dilute solution
of sodium hydroxide did not, apparently, affect the solubility of the
pigments in alcohol. Pigment A was the only pigment soluble in cold

°Zopf, W. Die Pilze. Page 144. Breslau, 1890.

10Samuely, F. Abderhalden’s Handbuch der Biochem. Arbeitsmethoden II:
758, I910.

ENDOTHIA PIGMENTS 349

petroleum ether. Pigment B was soluble in water. It is obvious
then that these pigments are lacking in many of the properties of
lipochrome and there is little reason at present for assuming that they
belong in this rather indefinite group. |

There is some evidence to support the conclusion of Anderson that
the pigment in E. parasitica is aurine. According to Rota’s"' system
for the classification of coloring matters these two pigments might
be classed as aurin.

Aurin is the trade name applied to a red pigment obtained by
heating phenol and oxalic acid with sulphuric acid. According to
Dale and Schorlemmer,” it was applied to this preparation as pre-
pared by Kolbe and Schmitt.% Dale and Schorlemmer found this:
preparation to be a mixture of compounds and succeeded in separating:
out what they considered pure aurin. The name aurin is retained.
by Schultz and Julius“ and by Allen for para-rosolic acid which’
according to Allen has a formula CyHuO3. This dye, however, is
insoluble in benzol and carbon-bisulphide and by boiling with sodium
hydroxide and zinc dust it is decorolorized. This is not true of the
pigment from E. parasitica. Moreover, a comparison of pigments
A and C in both alkaline and acid solutions with solutions of com-
mercial aurin shows that they are not the same color.

Other points of difference might be mentioned. It is, however,
apparent that while aurin and the pigments from FE. parasitica have
some properties in common the conclusion that they are the same is
unwarranted. Whether any of the three pigments considered in
this paper are similar in structure to aurin is a problem which needs
further investigation. It lies within the scope of this paper to take
up the chemical and physical properties of these pigments only in so
far as is necessary for separating and distinguishing them from each
other. Their chemistry is under investigation and will be considered
in a later paper.

11 Wiley, H. W., et al. Official Methods of Analysis of the Association of
Official Agricultural Chemists. U.S. Dept. Agr. Bur. Chem. Bull. 107. rg10.

22 Dale, R. S., und Schorlemmer, C. Ueber das Aurin. Ber. Deutsch. Chem.
Ges. 4: 574-576. 1871.

18 Kolbe, H., u. Schmitt, C. Rother Farbstoff aus dem Kreosot. Annalen der
Chemie und Pharmacie. 119: 169-172. 1861.

14 Schultz, G., u. Julius, P. Tabellarische Ubersicht der Kiinstlichen Organ-

ischen Farbstoffe. Dritte Auflage. 124, 1897.
© Allen, Alfred, H. Commerical Organic Analyses. ed. 3. 3:310-311. 1902.

350 LON A. HAWKINS AND NEIL E. STEVENS

Early in this work it became apparent that all the members of the
genus studied elaborated pigments which were bright yellow when
acidified and red when alkaline. This suggested the possibility that
if EL. parasitica could be grown on a sufficiently alkaline medium it
might produce the purple color considered characteristic of E. fluens.
The writers were able to suppress the production of the purple color
in cultures of E. fluens by the addition of 10 cc. n/10 sulphuric acid to
each 100 cc. culture flask. Cultures of E. parasitica were made to
produce a wine color in the culture media by the addition of 2 grams of
calcium carbonate to each 100 cc. flask before sterilizing.

While not particularly significant, these tests furnish another
example of the necessity of carefully standardizing culture media
used in critical comparative study of fungi. This is especially true
since no character is more commonly used to distinguish species of
fungi in pure culture than the production or nonproduction of color
changes in the mycelium or culture media. Under carefully controlled
cultural conditions the ability to produce color on certain media may
be a distinguishing character of great value, as in the work of Appel
and Wollenweber!, Grossenbacher and Duggar,!” Thom,!8 and others.

In studying the growth of species of Endothia on liquid media it
was noticed that F. parasitica produced a red coloration in old cultures
when the medium contained peptone. This was found to be due to
the ammonia liberated by the growth of the fungus acting on the
yellow pigment. Anderson (loc. cit., p. 14) mentions the fact that
old cultures of E. parasitica often become purple or wine colored and
attributes this change to the fact that the fungus in its growth on the
agar gradually causes it to become alkaline, thus changing the pigment
from yellow to purple. The writer’s investigations indicate that these
conclusions were probably correct and that the change to an alkaline
reaction may have been due to the formation of ammonia in the
cultures,

In the experimental work described it has been shown that three

16 Appel, O., and Wollenweber, H. W. Grundlagen einer Monographie der
Gattung Fusarium (Links). Arbeiten aus der Kaiserlichen Biologischen Anstalt
fiir Land- und Forstwirtschaft. 8: 1-207. I910. :

17 Grossenbacher, J. G., and Duggar, B. M., A contribution to the Life-history,
Parasitism, and Biology of Botryosphaerta ribis. N. Y. Geneva Agr. Exp. Sta.

Tech. Bull. 18. 1911.
18 Thom, Charles. The Pencillium luteum-purpurogenum group. Mycologia 7:

134-142. I915.

*“ENDOTHIA PIGMENTS Sot

different pigments are elaborated by the fungi in this genus. It has
been shown further that the curves of the percentage of spectral
transmission of the acidified alcoholic extracts of the pigments from
the seven different fungi group themselves into three distinct classes.
An investigation of the pigments produced by typical fungi from each
of these three classes, 72. e., E. parasitica, E.. fluens, and E. tropicalis,
show that there is one pigment common to all three groups but that
each of the three fungi is characterized by some one of the three pig-
ments.

E. tropicalis apparently elaborates only pigment A in quantity,
although a very little of pigment C may be present. FE. parasitica,
the type of the group which contains also LE. longirostris and E. fluens
mississippiensis, secretes pigment A in small amounts, but pigment
B was not found at all. Pigment C is characteristic of this group.
The group containing FE. fluens, E. gyrosa, and F£. singularis, of which
E. fluens is considered typical, is apparently the only one of the three
which secretes all the pigments. Pigment B is found only in this
group, and is thus characteristic of the group. This pigment is soluble
in water and is the cause of the “‘perilla purple’’ color in cultures of
this fungus. It frequently forms crystals on the mycelium (fig. 6).

It is evident from this work that the curves of percentage of
spectral transmission shown in figures I to 5 are in most cases curves
of mixtures of these pigments, and that the difference in the curves of
spectral transmission for the three groups is due to the fact that differ-
ent pigments predominate in the alcoholic extracts from the fungi of
these three groups. It is probable that further investigation with
quantitative methods would show that the variation in the curves for
different members of the same group was due to the presence in varying
proportions of the pigments characteristic of that group.

It is of interest to note that the grouping of the species based on
the spectral transmission of the acidified alcoholic extracts of the
mycelium shows no apparent agreement with the division based on
morphology, host, or geographical distribution. £. tropicalis, which
apparently produces only pigment A, differs to be sure from all the
other fungi examined in host and geographical distribution, being
known only from Ceylon on Elaeocarpus. It is, however, rather
closely related morphologically to E. parasitica and other members of
this group.

The group characterized by a curve of spectral transmission in-

352 LON A. HAWKINS AND NEIL E. STEVENS

dicating the presence of pigments A and C contains only forms having
oblong fusiform to oblong ellipsoid ascospores, and with somewhat
similar stromatic characters. The members differ widely, however, in
host and geographic relations, E. longirostris being a tropical form
known at present only from Porto Rico and French Guiana. E.
parasitica and FE. fluens misissippiensis occur on the same hosts,
Castanea sp. and Quercus sp., but E. parasitica is the destructive
chestnut blight organism known already from China, Japan, and the
United States, while EL. fluens mississippiensis is a weak saprophyte
which has been found in only four localities in the United States.

The group characterized by a spectral transmission curve indicating
the presence of all three pigments contains species widely different, in
morphology and distribution. E. gyrosa and FE. singularis have
cylindrical ascospores and are found only in the United States, E. sin-
gularis only on the chaparral forming species of Quercus in Colorado
and New Mexico. FE. gyrosa is found on species of Quercus, Fagus,
Castanea, and Liquidambar, and is widely distributed in this country,
though abundant only in the southeastern portion. F. fluens, on the
other hand, is a cosmopolitan species found in the United States on
Castanea and on Quercus in Europe and Asia on a variety of hosts.
In their stromatic characters also these species are widely different.
The stromata of E. singularis are large and irregular, being 3-5 mm.
wide by 2-4 mm. high, and disintegrate into a powdery mass when the
wall is ruptured. The stromata of E. fluens are much smaller, being
only .75-3 mm. in diameter by .5 to 2.5 mm. high and very compact.

On the other hand, the two most closely related fungi of the genus,
E. fluens and E. fluens mississippiensis, fall in different color groups.
In fact, it may well be that the production of pigment B by E. fluens
is the chief character which distinguishes it from its variety. The
varietal name was proposed by Shear and Stevens to designate a form
which they were unable to separate from E. fluens on morphological
grounds, but which showed constant differences on culture media.

The fact that the red pigment in F. fluens is not found in £. para-
sitica grown on the same media and under the same conditions is of
especial interest since the two species are so much alike morphologically
and grow on the same host, yet differ so widely in their relation to their
hosts, £. parasitica being, as has already been pointed out, the uni-
formly destructive chestnut blight parasite, while E. fluens is a harm-
less saprophyte.

ENDOTHIA PIGMENTS 353

The fact deserves more than passing notice also that what is
apparently the same pigment is produced by all the known members
of a genus, which, although small, includes species from four conti-
nents, from both temperate and tropical climates and occurring on
rather unrelated hosts.

BUREAU OF PLANT INDUSTRY,
WASHINGTON, D. C.

OBSERVATIONS ON AN ACHLYA LACKING SEXUAL
REPRODUCTION

Wn. H. WESTON

The following paper embodies the results of the writer’s investi-
gation of a Saprolegniaceous fungus which shows the zoospore pro-
duction characteristic of the genus Achlya, but which lacks sexual
reproduction. The resistant function exercised in most species of
Achlya by the sexually produced spores is apparently assumed, in
the form to be described, by large, heavy-walled spores of non-sexual
origin, and sharply defined morphological characteristics. Structures
of a similar nature have been recorded in other Saprolegniaceae; but
this form is unique in that its reproduction is limited to the zoosporan-
gia and to these structures which are of a particularly distinct type.
In contradistinction to the zoospores, these bodies are resistant to
unfavorable conditions. It seems advisable, therefore, to avoid the
implications of the names ‘‘conidia,’’ ‘‘chlamydospores,”’ “‘gemmae,’’
‘““Sporangienanlage,”’ “‘resting sporangia,’ or “resting spores,” gen-
erally used for these structures; and to employ simply the term
“resistant spores.”’

The writer realizes only too well the incompleteness of this study;
and regrets that the untimely destruction of stock cultures has pre-
vented the further physiological and cytological investigations which
had been planned. These results are presented for publication in the
hope that they may prove of interest to those working in the same
field.

99 66

MATERIAL AND METHODS

The fungus appeared on dead flies dropped into a culture containing
sediment and algae from a stone watering trough at Waverly, Mass.
For over two years gross cultures of the fungus were successfully
maintained in battery jars cooled in running water, and covered with
glass to exclude the dust.

Pure cultures were obtained by the following methods.

1. A young sporangium was washed repeatedly in sterile water,
and allowed to discharge zoospores in a drop of sterile water on a

354

AN ACHLYA LACKING SEXUAL REPRODUCTION 355

slide. This drop was then added to 2 or 3 cc. of sterile water in an
atomizer, and the resulting mixture sprayed on the surface of 2 percent
beef-extract agar in petri dishes. By examination with a binocular
microscope, the positions of single isolated zoospores were noted and
marked. After two or three days of growth in a cool place, bits of the
uncontaminated peripheral portions of the mycelia arising from these
zoospores were transferred on agar chips to fresh nutrient substratum.

2. A large number of vigorous resistant spores were washed re-
peatedly in sterile water and plated out in a separation culture series _
of four or five plates of 2 percent beef extract agar. The spores were
sufficiently resistant to remain uninjured by their exposure to the hot
agar, and promptly germinated. In the last two or three plates some
of the mycelia arising by germination of the isolated resistant spores
were found to be uncontaminated; and from them transfers were
made to fresh nutrient material.

When uncontaminated mycelia of the fungus had been obtained
by the above methods, stock cultures were maintained on firm corn-
meal mush in 500 cc. flasks. For the investigation of the fungus,
mycelia were grown in petri dishes in nutrient solutions of beef-extract,
and of pea, corn, bean, and other vegetable decoctions, and transferred
to sterile water or to various solutions in petri dishes or hanging drop
cultures for further development according to the methods already
worked out by Klebs (7), Kauffman (6), Obel (13), Pieters (14), and
others.

DEVELOPMEN?| OF THE: FUNGUS

The normal life cycle of the fungus comprises the establishing of a
branched mycelium which gives rise to large numbers of zoosporangia
of the Achlya type. At first these zoosporangia are produced ex-
clusively; but gradually they are superseded by abundant resistant
spores which continue to be formed until the mycelium is exhausted.
This regular cycle was observed for a space of two years in the original
gross culture in which the fungus appeared; and since the conditions
there probably closely approximated those of its natural habitat, we
may infer that it would follow the same cycle in nature.

In battery-jar cultures like those mentioned above, species of
several Saprolegniaceous genera have been found by the writer to
maintain their normal cycles of sexual as well as non-sexual repro-
duction for long periods of time. This Achlya, however, under these
conditions continued for two years to follow its cycle of unbroken

356 WM. H. WESTON

non-sexual reproduction. Under pure culture conditions, also, my-
celia, grown on sterilized flies, or transferred from various nutrient
solutions to water, developed Achlya zoosporangia followed only by
resistant spores. |

The reactions of the fungus when subjected to culture conditions
designed to induce sexual reproduction will be discussed later, the
phases of its normal development now being considered in detail.

Zoosporangia.—The vegetative mycelium of the fungus, whether
in gross or pure culture, consists of non-septate, branching hyphae
that show no apparent difference from those of other species of Achlya.
In gross cultures, as has been stated above, the first reproductive
structures to which the mycelium normally gives rise are zoosporangia.
In pure cultures, well nourished mycelia can be induced to form zoo-
sporangia in abundance by a distinct and rapid decrease of the food
supply according to the methods of Klebs (7), Kauffman (6), Horn
(4), and Pieters (15).

The process of sporangium development and spore formation in
this Achlya is quite normal, agreeing in its external features with the
description of Ward (17) for Achlya polyandra de B., and Humphrey
(5) for Achlya Americana Humph. No detailed description need
therefore, be given here.

The fully developed sporangia are cylindrical (Fig. 12) to fusiform
(Fig. 1) in shape, and vary greatly in size. They are formed in basi-
petal series, or renewed by side branching (Fig. 12), but never by
growth of the sporangiophore through the empty sporangium.

Escape of the sporangiospores from the sporangium is through a
terminal papilla of dehiscence; and in connection with the mouth of
the sporangium is formed the sphere of the encysted spores (Fig. 1)
which is characteristic of the genus Achlya. From each encysted
spore (Fig. 2) thus situated there may emerge under the proper con-
ditions a zoospore of the laterally biciliate Achlya type. The en-
cysted spores average 10.5 uw in diameter while the zoospores measure
about 12.5 uw by Ou.

In shape these zoospores are ovoid with a flattened side bearing a
longitudinal groove or sinus from which arise the two cilia (Fig. 3).
It is unfortunate that investigators have been content to describe and
figure this type of zoospore in the Saprolegniaceae as ‘‘bean”’ or “ kid-
ney’’ shaped. A careful examination of the living spores under a Zeiss J
water-immersion lens readily proves these terms to be inadequately

AN ACHLYA LACKING SEXUAL REPRODUCTION B57

descriptive or even misleading. Edson (2) for his most interesting
new genus Rheosporangium describes a similar type of spore (p. 285)
as ‘‘plano-convex or slightly concavo-convex, with a central vacuole,
and on the flattened side a sinus from the bottom of which the two
cilia of unequal length arise.’”’ The description and the accompanying
figures agree in general with the writer’s observation of this type of
spore in Achlya, Thraustotheca, Dictyuchus, and other genera of the
Saprolegniaceae.

Upon emerging from the cysts the zoospores swim actively about
for’as long as thirty minutes if in pure water, and then round off,
lose their cilia, and encyst (Fig. 4). These encysted zoospores ger-
minate either directly or after a period of rest according as available
nutriment is present or absent. Germination takes place by the out-
growth of a tube (Fig. 5) which shows scanty development, or which
ultimately forms an extensive mycelium, the amount of growth being
proportionate to the amount of available nutriment.

Resistant Spores.—In the details of its life history as considered
thus far, the fungus is a typical Achlya. The formation of a hollow
sphere by the sporangiospores at liberation, the character of the
zoospores which emerge, and the method of sporangium renewal are
distinguishing characteristics of the genus. In its further develop-
ment, however, the fungus differs from any Achlya hitherto recorded
in that zoosporangium formation is normally followed by the produc-
tion not of sexual organs, but of heavy-walled resistant spores of a very
distinct type. Both in gross and in pure culture, formation of these
resistant spores follows the production of zoosporangia with remarkable
regularity. Resistant spores are produced exclusively, however, under
the conditions found by Klebs (7) to inhibit the formation of zoo-
sporangia, although favoring the development of ‘‘gemmae.”’ Par-
ticularly is this true of I percent agar solution. i

Although this Achlya is truly an aquatic fungus, these resistant
spores are produced in culture not only in liquids but also aerially on
solid media. In flask cultures on fairly dry cornmeal they are formed
in especial abundance. When such a culture is inoculated with a
bit of mycelium, a vigorous growth ensues, covering the surface of the
medium from which innumerable stout hyphae with dense whitish
content push up into the air. On examining such a culture with
a binocular microscope after three or four weeks, large numbers of
resistant spores can be seen especially around the edge of the culture.

358 WM. H. WESTON

Most of these spores are formed in chains, and in some cases the
terminal spores abjuncted by the pressure of the spores below are
tipped over sideways, although still remaining attached. That these
resistant spores should be borne aerially is of interest because of the
close analogy presented to the aerially borne conidia of certain of the
- Peronosporaceae. It does not seem justifiable, however, to regard
this phenomenon as of any further significance than an instance of
convergenec, of parallel development, in the two families.

The process of resistant spore formation takes place as follows:
The tip of a hypha becomes densely filled with coarsely granular pro-
toplasm carried up by the streaming of the currents in the peripheral
protoplasm. Gradually more and more material is accumulated in the
tip forming a dense mass which slowly extends downwards to the base
of the spore initial. Meanwhile the tip of the hypha swells to the
_ spherical or oval shape of the mature resistant spore; and finally the
spore initial is separated from the hypha by a wall. Around the inner
surface of the terminal cell thus cut off, a wall of varying thickness is
laid down by the dense, coarsely granular protoplasm which occupies a
peripheral position around a large vacuole (Fig. 9).

After the terminal resistant spore has been formed, as just de-
scribed, other resistant spores are generally formed in basipetal suc- -
cession under the first as in figures 6, 9, and 10, giving a torulose
series closely resembling the catenate oogonia of Saprolegnia torulosa
de B. In gross cultures, the resistant spores are generally formed in
this fashion, giving a very striking and characteristic appearance to
the plant. After a time the fragile outer wall of the resistant spore is
ruptured (Fig. 9), and the spore is set free in the water to be washed
about and then to slowly settle to the bottom. Further development
varies with the conditions of the environment, and will be considered
later. |

The resistant spores are generally formed in a terminal series as
above described, but they may be intercalary (Figs. 7 and 8) or may
form branching systems of various types. Inshape the resistant spores
are spherical, cylindric, oval, or less often pyriform, or clavate. Under
conditions closely approximating those of nature, the resistant spores
are spherical (Fig. 10) or oval; and on fairly dry cornmeal cultures,
they are the same. On vegetative mycelia transferred to sterile
water and to solutions of various sorts, the resistant spores which are
formed are at times cylindrical or club-shaped (Figs. 1 and 12). In

AN ACHLYA LACKING SEXUAL REPRODUCTION 359

solutions of harmful concentration or at a temperature approaching
the maximum, they tend to be irregular in shape (Fig. 8). In size
the spores average about 110 w in diameter, the size being apparently
correlated with the vigor of the mycelium. The walls of the resistant
spores formed aerially on rather dry cornmeal, or in water cultures
kept cold, are thick (Fig. 1c); but in spores grown in various solutions
at laboratory temperature, the thickness is often less (Fig. 12). By
chloroiodide of zinc the walls are colored not blue, but a muddy brick-
red, showing that their composition is not of pure cellulose, and re-
calling the reaction of the oogonial walls of other Saprolegniaceous
species. |

Germination of the resistant spores varies with the environmental
conditions to which these bodies are subjected. In water in the
absence of food material, a tube is sent out which, after greater or less
growth, gives rise to a perfectly normal Achlya sporangium (Fig. 12).
In nutrient solutions, however, or on nutrient agar, germination is in-
variably siphonoblastic, the tube or tubes of germination rapidly
giving rise to an extensive mycelium (Fig. 11). At laboratory tem- -
perature the resistant spores germinate in from twenty-four to forty-
eight hours in the presence of nutriment, and somewhat more slowly
in pure water. Germination need not be preceded by a period of rest;
since no particular difference in behavior was observed between spores
just formed and those two to four weeks old. The spores are very
resistant to cold, those from one culture surviving an exposure to
outdoor conditions during two winter months in which the water was
repeatedly frozen and thawed. The writer has made no extensive
investigation of the degree of resistance to extremes of temperature
or to desiccation shown by resistant spores. There has been demon-
strated, however, a degree of resistance that may perhaps be regarded
as an indication that these spores play for the fungus the resistant
role usually assumed in others of the family by spores of sexual origin.

The formation of zoosporangia and resistant spores as described
above completes the reproduction of the fungus. Under natural
conditions of growth, the formation of sexual organs was never ob-
served to take place throughout the two years in which the fungus was
under investigation. Many attempts to induce the formation of
oogonia and antheridia were made without success. Hoping to stim-
ulate sexual reproduction vigorous mycelia were subjected to the
influence of the solutions by means of which Klebs (7), Kauffman (6),

360 WM. H. WESTON

Obel (13), and others, had successfully induced the formation of
sexual organs in other members of the family. Various concentra-
tions of haemoglobin, of leucin, of potassium, sodium, and calcium
phosphates, and of other substances were tested; but in no case was
the production of sexual organs achieved. The effect of various
temperatures and of changes in temperature was also tried; and the
rapid or gradual drying of cultures was attempted. Moreover
starving was resorted to, the fungus being grown on synthetic media
containing a minimum of nutriment or on such natural substrata as
pomace flies that had been thoroughly leached out and then dried.
Dwarfed plants were produced, but no oogonia or antheridia developed.
In all the cases above, a more or less copious formation of resistant
spores resulted; but no production of sexual organs occurred.

DISCUSSION

The formation of bodies of non-sexual origin resembling to some
extent the resistant spores of the foregoing description has been re-
corded in nearly every genus of the Saprolegniaceae. In most of the
cases on record these structures appear to be transient resistant stages
which do not arise under conditions of the environment favorable to
zoospore formation, but are induced by extremes of temperature, by
foul water, or by other unfavorable conditions. As Klebs (7) has
shown, the production of ‘“‘gemmae’”’ (as he termed these structures)
takes place when the fungus mycelium is subjected to conditions of
environment permitting of growth and yet prohibiting zoospore
formation. The conditions to which the Saprolegniaceae were sub-
jected during investigation, before methods of pure culture were
introduced, quite generally resulted in the formation of numerous
structures of this sort. No reference to the occasional descriptions of
these bodies need be made here, since the literature has been carefully
covered in the monographs of Fischer (3) and von Minden (12). It
may not be amiss, however, to recall one interesting instance discussed
in an early paper by Walz (16). In Saprolegnia dioica Prings. he
described a type of reproduction which at that time was unknown for
the Saprolegniaceae, namely, the formation of the thick-walled bodies
produced in basipetal succession at the ends of the hyphae. The
resemblance between these structures (cf. Walz, Fig. 20) and the re-
sistant spores of the Achlya described above (Figs. 6 and I0) is re-
markably close. It would be of interest to know whether these bodies

AN ACHLYA LACKING SEXUAL REPRODUCTION 361

were formed in addition to the sex organs or took their place. Walz,
however, does not mention the zoosporangia or the sex organs; but
from his naming the form so definitely, we must infer that he observed
the sex organs of the species.

In quite a different category from the occasionally induced struc-
tures just considered, must be placed the resistant bodies which are
regularly produced in the life cycle of certain Saprolegniaceae forming
sex organs only rarely, or even lacking them entirely.

In the genus Saprolegnia there have been reported a number of
forms of this kind. Lindstedt (10) described a species of Saprolegnia
which produced zoosporangia and zoospores typical of the genus,
followed by spherical, pyriform, or more irregular bodies capable of
germinating at once by characteristic Saprolegnia zoospores, or of
remaining temporarily inactive and later germinating by a tube or by
zoospores on the renewal of favorable conditions. Maurizio (11)
also recorded three undetermined species of Saprolegnia which, after
the formation of the characteristic sporangia, produced not sex organs
but the irregular ‘‘Sporangienanlagen”’ which he considered significant
from the phylogenetic point of view. In all probability conditions of
culture are responsible for the behavior of the fungi in these instances;
since the method employed by both Lindstedt and Maurizio are shown
by the exact physiological investigations of Klebs (7) to favor ‘“‘gem-
mae’’ formation, and to hinder the development of sex organs. In
like manner the failure of Lechmere’s (8) more recently described
species of Saprolegnia to form oogonia under continued cultivation
may be ascribed to his culture methods. Although Lechmere used
pure cultures, he employed as a nutrient substratum egg albumin, a
substance readily broken down into injurious compounds by the
activity of the fungus. Indeed he even mentions (p. 168) changing
the water of his cultures every day to keep them fresh. Had his cul-
ture conditions been more favorable, it is reasonable to expect in the
light of Klebs’s investigations that the sporangia would not have
presented such confusing abnormalities of development; and the
torulose ‘““gemmae’’ would readily have given place to the normal
oogonia of the species (Saprolegnia torulosa) which Lechmere later
(9) determined this fungus to be. Recently, however, Pieters (15)
has subjected certain species of Saprolegnia to exact physiological
investigation with interesting results. One of these forms, Saprolegnia
Kauffmaniana, characteristically produced numerous sporangia and

362 WM. H. WESTON

“gemmae’’; but formed sexual organs rarely under normal culture
conditions, and in haemoglobin solution only at the specific concen-
tration of 0.025 percent. Another form, Saprolegnia monoica var.
vexans, after reproducing only non-sexually during sixteen months
of cultivation on various substrata which had induced sex organs in
other forms, finally produced an abundance of oogonia and antheridia
under the influence of a combination of M/200 levulose and M/200
leucin. Moreover, in addition to these two species, Pieters mentioned
a Saprolegnia (no. 66) forming numerous single spherical ‘ gemmae,”’
and yet, during eighteen months of cultivation, absolutely failing to
respond by the formation of sex organs to the usually successful
methods of culture.

In the genus Achlya, also, there have been reported somewhat
similar forms which show a lack of sex organs coupled with the for-
mation of resistant structures. Under the name of Achlya oidifera
Horn (4) described a form which under long continued observation in
pure culture, and with the most exact methods of cultivation, showed
consistent non-sexual reproduction. None of the conditions by which
Horn induced the formation of great numbers of oogonia in Achlya
polyandra de B. were successful in the case of Achlya oidiufera. In-
stead, there developed oidia-like hyphal segments which formed
zoospores at once, or rested even as long as a month and were still
capable of further growth. In his study of Achlya oidufera, Horn
observed oogonia only once within an ant egg on which he had grown
the fungus; and although no antheridia were found on account of the
‘advanced stage of -development, the pits on the oogonium wall, and
the eccentric zoospores, suggested Achlya polyandra de B. or some other
member of the prolifera group.

More recently Coker (1) made a careful! study in pure culture of
Achlya paradoxa, a species frequently collected by him in North
Carolina, and unique in the aberrant behavior of its zoospores. The
fungus when isolated from single zoospores produced zoosporangia
and ‘‘chlamydospores’”’ in abundance. Although the form was for a
long time maintained in pure culture under a variety of conditions,
sexual organs proved very rare. They were observed but a few times,
and their formation could not be induced by a great variety of sub-
strata including nutrient solutions of various kinds containing organic
or inorganic salts.

The peculiar Achlya described by the writer in the earlier part of

AN ACHLYA LACKING SEXUAL REPRODUCTION 363

this paper differs from A. oidtifera and A. paradoxa in the distinct
morphological character of its resistant spores and in the absolute lack
so far of any sexual reproduction. The destruction of stock cultures
has prevented testing the effect of mixtures of leucin and levulose in
the proportions found successful by Pieters (15) in inducing oogonium
formation in Saprolegnia. It is possible that this Achlya also would
yield to this combination or to some other more suited to its own
physiologic idiosyncrasies. In view of such a possibility it does not
seem advisable to assign a specific rank and name to the Achlya de-
scribed above. Both physiologically and morphologically this Achlya
differs from the other members of the genus; but further investigation
might result in inducing the formation of sex organs, and might prove
the fungus to be a variety of some already established species.

The significance of these Saprolegniaceous forms which partially
or completely lack sexual reproduction is not at once apparent. The
few explanatory theories advanced by early investigators were based
on the study of material under unfavorable conditions preceding
methods of pure culture; and hence may be disregarded. Even the
more modern investigators of such forms offer but little explanation
of their significance. Maurizio’s theory that asexual resistant spores
(Sporangienanlage) are reminiscent of primitive, non-specialized
structures from which both sporangia and oogonia have been evolved
has been largely nullified by Klebs’ exact researches on ‘‘gemma”’
formation. On the other hand, in Klebs’ interpretation of ‘'gemmae”’
as “Hemmungsbildungen”’ induced by conditions unfavorable to
other stages of development, we have an explanation undoubtedly
correct in most cases, yet hardly applicable to such distinct and con-
sistently formed structures as the resistant spores of the Achlya that
is the subject of this paper.

An extremely significant suggestion, however, has been made by
Pieters in a recent paper. Pieters’ (15, p. 483) suggestion that ‘‘the
production of sexual organs may depend on some special combination
of conditions, differing, doubtless, for each form’ is important in
emphasizing the physiological difference existing among the species
and varieties of the Saprolegniaceae. The members of the family,
probably because they are coenocytic in structure, and are completely
submerged in the culture media, are very sensitive to the nature of
their environment. Morphologic studies reveal no reason why under
the same conditions one form should produce oogonia and antheridia;

364 WM. H. WESTON

while another form seemingly identical should just as persistently
fail to produce these organs. The physiologic investigations begun
by Klebs, and carried on by a number of others, have gradually re-
sulted in a far more enlightening conception of the family. Because
of these investigations, we are, in the opinion of the writer, justified
in regarding the Saprolegniaceae as a series of forms ranging from
those which normally produce sexual organs, through forms which
produce sexual organs only under unusual conditions, to forms which
have entirely lost the power of sexual reproduction.

On the basis of this ability to form sexual organs the Saprolegni-
aceae may, for convenience, be divided roughly into three groups.
First there are those forms which are strongly sexual producing an
abundance of oogonia and antheridia even under adverse circum-
stances. Here may be grouped Saprolegnia monoica de B., S. diclina
de B., Achlya polyandra de B., A. prolifera de B., and others. Second,
there may be grouped together those forms which possess the power of
sexual reproduction, and perhaps even show a sexual stage as a normal
phase of their life cycle under unknown conditions of growth in
nature, but which under investigation remain imperfect; because the
exact conditions favoring sexual reproduction are not supplied.
Here belong Achlya oidufera Horn, A. paradoxa Coker, Saprolegnia
Kauffmaniana Pieters, and others. Finally, there are certain forms
which have lost the power to produce any sexual organs whatever.
The Achlya that is the subject of this paper, the Saprolegnia no. 66
of Pieters, as well as other forms not as yet recognized as imperfect,
should be placed in this category.

The statement that these fungi are absolutely incapable of sexual
reproduction cannot be made logically, of course, until they have been
grown under every possible condition. Even if these forms should on
later investigation be found to possess latent powers of sexual repro-
duction, it seems probable that the category of non-sexual Sapro-
legniaceous forms will still persist. It has long been recognized that
the male reproductive organs or antheridia are invariably developed
in some Saprolegniaceous species, occasionally in others, and never
in some. In the genus Saprolegnia particularly, one can trace all
gradations from forms normally developing antheridia in abundance,
through forms in which antheridia occur rarely but may be induced,
to truly parthenogenetic forms. In the opinion of the writer, a quite
similar condition also obtains with regard to the oogonia of the Sapro-
legniaceae.

AN ACHLYA LACKING SEXUAL REPRODUCTION 365

The lack of sexual reproduction, of a “‘perfect stage,” is not un-
common in the fungi in general. The resemblance of the condition
existing in the Achlya which is the subject of this paper to that of
certain Fungi Imperfecti is, of course, obvious. Moreover, the genus
Apodachlya in the closely related Leptomitaceae and the genus Blas-
tocladia in the Blastocladiaceae furnish an interesting comparison to
the Achlya described above. It is worthy of note that a loss of sexual
reproduction is, at least in the forms just mentioned, concomitant
with the production of more or less resistant nonsexual spores. This
is particularly noticeable in Apodachlya and Blastocladia; since all
the known species of these two genera (save perhaps the doubtful
A podachlya completa of Humphrey) alike exhibit a consistent pro-
duction of zoospores and resistant spores with an attendant lack of
sexual organs. If the dangerous luxury of theorizing about the evo-
lutionary origin of such structures were permissible, they might be
regarded as adaptations evolved to withstand unfavorable conditions
in forms that were gradually losing the sexually produced spores that
generally serve this purpose. Disregarding the purely theoretical
origin of this condition in the Achlya under discussion, however, the
actual condition itself remains; and in whatever way this Achlya may
be interpreted it at least presents an interesting case of a very distinct,
and well defined, but non-sexual representative of a Saprolegniaceous
genus usually strongly sexual.

SUMMARY

1. In the method of zoospore production and liberation, and in
the character and behavior of the zoospores themselves, the fungus
that is the subject of this paper distinctly belongs to the genus Achlya.

2. In contradistinction to most species of the genus, this fungus,
as far as observed, entirely lacks sexual reproduction, nor does it
produce oogonia and antheridia under the methods of culture usually
successful in inducing these organs.

3. The fungus is distinguished by the consistent production, under
widely varying conditions, of resistant spores of non-sexual origin
and distinct morphological characteristics. These spores differ from
the ““gemmae’”’ described for other species in their regular occurrence
and clearly defined structure. ;

4. The writer regards this fungus as an Achlya that has lost its
sexual reproduction—the resistant function usually assumed by the

366 WM. H. WESTON

spores of sexual origin being in this case taken over by the non-sexual
resistant spores.

16.

17.

FEDERAL HORTICULTURAL BOARD,
U. S. DEp’T oF AGRICULTURE

LITERATURE CITED

. Coker, W. C. 1914. Two New Species of Water Molds. Mycologia 6: 285-

20%.. .3%PIs:

. Edson, H. A. tI915. Rheosporangium aphanidermatum, a New Genus and

Species of Fungus Parasitic on Sugar Beets and Radishes. Journ. Agr.
Research 4: 279-291. 5 Pls.

. Fischer, A. 1892. Saprolegniineae, 7 Rabenhorst’s Kryptogamenflora,

Thiel 1, Abt. 4, 310-383. 12 Figs.

. Horn, L. 1904. Experimentelle Entwicklungsanderungen bei Achlya polyandra

de Bary. Ann. Mycol. 2: 207-241. 21 Figs.

. Humphrey, J. E. 1893. The Saprolegniaceae of the United States. Trans.

Amer. Phil. Soc. N. Ser. 17: 1-148. 7 Pls.

. Kauffman, C. H. 1908. A Contribution to the Physiology of the Saproleg-

legniaceae, with Special Reference to Variations in the Sex Organs. Annals of
Botany 22: 361-389. 1 PI.

. Klebs,G. 1899. Zur Physiologie der Fortpflanzung einiger Pilze. II. Sapro-

legnia mixta. Jahrb. Wiss. Bot. 33: 513-593. 2 Figs.

. Lechmere, A. E. tg10. An Investigation of a Species of Saprolegnia. New

Phytol. 9: 305-319. 2 Pls.

. Lechmere, A. E. tg11. Further Investigations of Methods of Reproduction

in the Saprolegniaceae. New Phytol. 10: 167-203. 6 Figs.

. Lindstedt, K. 1872. Synopsis der Saprolegniaceen und Beobachtungen ueber

einiger Arten. R. Friedlander u. Sohn, Berlin. 1-69. 4 Taf.

. Maurizio, A. 1896. Die Sporangiumanlage der Gattung Saprolegnia. Jahrb.

Wiss. Bot. 29: 75-131. 1 Taf.

. Minden, M. von. i912. Saprolegniineae, in Kryptogamenflora der Mark

Brandenburg 5, 209-608. Text Figs.

. Obel, P. 1910. Researches on the Conditions of the Forming of Oogonia in

Achlya. Ann. Mycol. 8: 421-443. 4 Text Figs.

. Pieters, A.J. 1915. The Relation between Vegetative Vigor and Reproduction

in some Saprolegniaceae. Amer. Journ. Bot. 2: 528-576. 2 Text Figs.

. Pieters, A. J. 1915. New Species of Achlya and of Saprolegnia. Bot. Gaz.

60: 483-490. I PI.

Walz, J. 1870. Beitrage zur Kenntniss der Saprolegnieen. Bot. Zeit. 28:
537-940, 553-5511 Lal.

Ward, H. M. 1883. Observations on the Saprolegnieae. Quart. Journ. Mic.
Sci. f., Ser..23%'272-201. TP.

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE XVIII.

WESTON : OBSERVATIONS ON ACHYLA.

AN ACHLYA LACKING SEXUAL REPRODUCTION B67,

EXPLANATION OF PLATE XVIII

The figures were drawn from living material at the level of the stage, with the
aid of an Abbé camera lucida. The approximate magnification of the combination
of lenses used is given in each case, but applies to the original figures, which have
been reduced to about % their diameter in reproduction.

Fig. 1. Emptied sporangium with hollow sphere of escaped sporangiospores
adherent to the tip. Beneath the sporangium a resistant spore has already formed.
X §50.

Fic. 2. Group of 5 escaped sporangiospores. From one now empty, a secon-
dary zoospore has already escaped; another shows a papilla of dehiscence forming.
X 1,400.

Fic. 3. Laterally biciliate zoospore showing its peculiar shape. The groove
from which the cilia arise is above. XX I,400.

Fic. 4. Encysted zoospore after coming to rest. XX I,400.

Fic. 5. Germination of such an encysted zoospore. X 1,400.

Fic. 6. Portion of mature mycelium bearing several resistant spores. XX 150.
Fic. 7. Intercalary resistant spore. X 220.

Fic. 8. Two intercalary resistant spores of somewhat irregular shape. X 220.

Fic. 9. Abjunction of a terminal resistant spore by the rupture of its delicate
outer wall. X 350.

Fic. 10. Formation of three resistant spores in basipetal series. XX 350.

Fic. 11. Germination in nutrient agar of a resistant spore by means of a vigorous
germ tube. X 350.

Fic. 12. Germination in pure water of a resistant spore forming a typical
Achlya sporangium. Some of the adherent spores have been washed away showing
their arrangement in a hollow sphere. X 350.

THE RUSTS OCCURRING ON THE GENUS FRITILLARIA
CHARLES C., REES

The Fritillaria rusts of Europe and America furnish an interesting
problem for taxonomic study. In the absence of inflorescence the
resemblance of the genera Fritillaria and Lilium, upon which occurs
Uromyces Lalu (Link) Fckl., as interpreted by European authors, is
striking. This host resemblance, coupled with the fact that the
rusts infecting them present a somewhat similar gross appearance,
has resulted in considerable confusion in the naming of many col-
lections.

Evidence derived from a study of types and other material of the
Opsis-Uromyces species on these hosts in the Arthur herbarium has
led to the conclusion that the species on Fritillaria in Europe is mor-
phologically quite separate and distinct from the rust with the same
life-cycle on Lilium in that region and from the American form on the
same host as well. The writer is therefore inclined to recognize as a
valid species, Uromyces Fritillariae (Schlechtd.) Thiim., to which all
of the European specimens on Fritillaria, seen by the writer, should
be referred. All of the Uromyces collections on Lilium in Europe
belong, according to the writer’s interpretation, to the species formerly
known as Uromyces Lalit (Link) Fckl. This species was first described
by Strauss! under the name Uredo aecidiiformis on Lilium candidum.
Although he used the generic term, Uredo, in designating the form, itis .
evident from the text that telia are described. For these reasons the
species should take the name Uromyces aeciditformis (Strauss). Re-
gardless of the fact that the type of Uromyces Miurae? Sydow has
not been examined, it is the writer’s opinion that the American rust
on Fritillaria is identical with it. Such disposition of a rust new to
this country is not unusual when it is remembered that certain species
of rust and other fungi have for a long time been known to be
indigenous to Japan and eastern Asia as well as to the western and
particularly to the northwestern coast of North America.

1 Ann. Wett. Ges. 2:94. I811.
*Ann, Myc. 11: 94. 1913.
368

THE RUSTS ON FRITILLARIA 369

The morphological differences of these three species are very
clearly defined, and the species are easily separated, as shown in the
analytical key which follows. Because of its close host relationship,
Uromyces Holwayi Lagerh.’ is included in the key. A description of
this species, however, is omitted because of the fact that its life-
cycle, which includes uredinia, precludes any possibility of its being
confused with the species under discussion. No species of Uromyces,
other than this one, has yet been reported on Lilium in North America.

KEY TO SPECIES ON LILIUM AND FRITILLARIA
Telia exclusively considered

Teliospore-wall moderately thick (2-3 »), chestnut-brown.
Teliospores globoid (exclusive of apiculus), 23-27 by

SATIS SIL ical ed a ae CO OP ee ne eS ee 1. Uromyces aeciditformis.

iteliosporesvellipsoid,. 18-25 by 29=39.@. <....-\. 220s. 2. Uromyces Holwayt.
Teliospore-wall moderately thin (1.5-2 4), cinnamon-

brown.
Teliospores narrowly ellipsoid, 14-23 by 24-35 pu, 1n-

conspicuously yerrucose in lines.............-..4. 3. Uromyces Mturae.
Teliospores ellipsoid to broadly so, 23-31 by 31-42 yp,

BeMe relive TI ROSE! sis faeces Saws pea dae ade ald 4. Uromyces Fritillariae.

1. Uromyces aecidiiformis (Strauss) comb. nov.

Uredo aeciduiformis Strauss, Ann. Wett. Ges. 2:94. I8IT.
Caeoma Lalit Link in Willd. Sp. Pl. 67: 8. 1825.

Caeoma aecidiiformis Schlecht. Linnaea I: 239. 1826.
Aecidium Meleagris Duby, Bot. Gall. 2: 904. 1830.

Erysibe variolosa Wallr. Fl. Crypt. Germ. 195. 1833.

Erysibe rostellata var. Lila Wallr. Fl. Crypt. Germ. 210. 1833.
Uromyces Liliacearum Unger, Einfl. Bodens 216. 1836.
Aecidium Liliacearum Unger, Einfl. Bodens 220. 1836.

Uredo Lalu Rab. Deutschl. Krypt. Fl. 1:12. 1844.

Uromyces Rabenhorstu Kunze, Rab. Fungi Eur. 1693a. 1873.
Uromyces Lilia Kunze, Rab. Fungi Eur. 1693). 1873.

O. Pycnia amphigenous, rather numerous, among the aecia, small,
0.I-0.2 mm. across, punctate, conspicuous, subepidermal, honey-
yellow, becoming darker later, chestnut-brown-colored ring encircling
base, globoid in cross section, 160-190 yp high by 160-190 win diameter.

I. Aecia amphigenous, for the most part hypophyllous, also

3.N. Amer. Fl. 7: 242. 1907. (As Nigredo Lilit (G. W. Clinton) Arth.)

370 CHARLES C. REES

petiolicolous and caulicolous, gregarious in round or elongated groups
of various sizes, up to 10 mm. in length, cupulate, low, 0.3-1 mm. in
diameter; peridium white, turning to yellowish-brown, opening by a
central pore, enlarging later, the margin erect or incurved slightly,
erose; peridial cells oblong in cross section, 26-29 by 32-39 uw, abutted,
the outer wall 7-13 uw thick, striate, the inner wall 9-12 wu in thickness,
very finely and almost inconspicuously verrucose; aeciospores angu-
larly globoid, 19-23 by 19-27 yu; wall pale yellow, 1.5—2.5 uw in thick-
ness, finely and closely verrucose.

III. Telia amphigenous, numerous, scattered, elongated, 0.2—1.3
mm. in length, tardily naked, finally dehiscent by longitudinal slits
in the epidermis, pulverulent, chestnut-brown, ruptured epidermis
conspicuous; teliospores broadly ellipsoid to globoid (exclusive of
apiculus), 23-27 by 31-35 m (including apiculus); wall chestnut-
brown, about 3 uw thick, a low hyaline apiculus at the apex, moder-
ately rugose with longitudinal parallel ridges, sometimes appearing
almost smooth when wet; pedicel very fragile, short, hyaline.

On LiviacEaE: Lilium bulbiferum L., L. candidum L., L. carni-
olicum Bernh. and L. croceum Chaix. Throughout central Europe.

TYPE LOCALITY: Europe, on Lilium candidum.

ExsiccatTi: Thiim. Myc. Univ. 2ogz; Thiim. Fungi Austr 646;
Rab. Fungi Eur. 1693; Sydow, Ured. 1504; Sydow, Myc. March.
3010; Kunze, Fungi Sel. 35.

Even in those collections showing teliospores only, Uromyces
aecidtiformis is readily distinguished from Uromyces Holwayt, since
the teliospores of the former are more nearly globoid and have slightly
thicker walls than the latter (Fig. 1). However, the presence of

Fic. 1. Teliospores of Uromyces aecidiiformis showing optical sections and
surface view. X 625.

aeciospores, which are quite different in the two species, enables one to
separate them readily.

THE RUSTS ON FRITILLARIA 371

3. Uromyces Miurae Sydow, Ann. Myc. 11: 94. 1913.

O and i. Pycnia and aecia, unknown. |

III. Telia amphigenous, petiolicolous, numerous, occasionally
crowded in groups of two or three sori, round or broadly ellipsoid,
0.2-0.7 mm. across, tardily naked, finally dehiscent by longitudinal
rents in the epidermis, pulvinate, becoming pulverulent, cinnamon-
brown, ruptured epidermis conspicuous; teliospores rather narrowly
and irregularly ellipsoid to terete, 14-23 by 24-35 mu, rounded or taper-
ing at apex, usually tapering at base; wall golden to cinnamon-brown,
of uniform thickness (1.5-2 wu), a low (1.5-3 »), hyaline apiculus: at
apex, moderately and very inconspicuously verrucose, markings
arranged to form longitudinal striations, appearing almost smooth
when wet; pedicel very short, fragile, a sac-like swelling at point of
attachment with spore.

On LiLtacEAkE: Fritillaria Kamtschatcensis Ker. Alaska, British
Columbia, Japan. Fritillaria lanceolata Pursh. Washington. 3

TyPE LocaLity: Mt. Shirouma, prov. Shinano, Japan on’ Fritil-
laria Kamtschatcensis. :

NY ‘aq \
i AY! Y f
SHAN AY AM
n\ oe ® % . A |
aa! LAN BS 8d
NAGS, A \ q '
WANES q A
TS | vi
\ A A A! Gh
i AVY Oe AS hi ff
\) WY URD AK } |]
\ ANY BA
A YOA Wh d7g
\ NY WW a7 A
\ LR | y
V €

Fic. 2. Teliospores of Uromyces infrequens showing optical sections and surface
view. X 625.

~
v Ko

DISTRIBUTION: South-central and east-central Washington north-
westward through Vancouver and Queen Charlotte islands to south-
eastern Alaska; also in central Japan.

Exsiccatl: Ellis & Ev. N. Am. Fungi 1663.

It is impossible of course with only the telial stage present to assign
this species with certainty to any genus in the classification based on
the length of life-cycle, proposed by Arthur.* However, in spite of
the lack of positive evidence which would indicate its proper taxonomic

4Eine auf die Struktur und Entwicklungsgeschichte begriindete Klassifikation.
Result. Sci. Congr. Bot. Vienne 331-348. 1906.

272 CHARLES C, REES

location, the writer is inclined to consider it an Opsis-form similar in
life history to Uromyces aeciduformis and Uromyces Fritillariae. Suc-
cessful cultures or additional material bearing other spore stages will
be necessary before the full life history is understood.

Morphologically this form is notably different.from any species
yet reported on any Liliaceous host either in North America or Europe
and the addition of the North American material extends consider-
ably the distribution of this distinctive species. The teliospores are
more narrowly ellipsoid, have considerably thinner walls and are
verrucose in longitudinal striations (Fig. 2). The teliospores of
the other species discussed in this paper are distinctly rugose.

4. Uromyces Fritillariae (Schlechtd.) Thiim.; Voss, Oesterr. Bot.
Zeits. 26: 207. 1876

Caeoma Fritillariae Schlecht. Linnaea 1: 240. 1826.

O. Pycnia amphigenous, rather numerous, scattered among the
aecia, small, punctiform, conspicuous, subepidermal, honey-yellow
becoming dark chestnut-brown, flattened globoid in cross section,
80-95 uw in width by 60-80 yp high; ostiolar filaments free.

I. Aecia amphigenous, caulicolous, petiolicolous, crowded in linear
groups, cupulate, low, 0.3-I mm. in diameter; peridium at first white,
becoming yellowish-brown later, opening by a central pore after a_
longitudinal splitting of the epidermis has taken place, the margin

\_ |

Fic. 3. Teliospores of Uromyces Fritillariae showing optical sections and surface
view. X 625.

incurved, erose; peridial cells oblong in cross section, 19-26 by 29-32 py,
abutted, the outer wall 9-13 w thick, striate, the inner wall 9-12 u
thick, very finely verrucose; aeciospores angularly globoid to ellipsoid,
16-22 by 21-28 yw; wall yellow, 2-3 uw in thickness, very finely, closely
and almost inconspicuously verrucose.

THE RUSTS ON FRITILLARIA 3/3

III. Telia amphigenous, caulicolous, petiolicolous, numerous,
scattered, elliptical, rather small, o.1-0.8 mm. in length, tardily
naked, dehiscent by longitudinal slits in the epidermis, somewhat
pulverulent, chestnut-brown, ruptured epidermis conspicuous; telio-
spores broadly ellipsoid to obovoid, 23-31 by 31-42 yu; wall golden- to
cinnamon-brown, I.5-2 p thick, low (1.5-3 mu) hyaline apiculus at apex,
rather delicately rugose in longitudinal striations; pedicel hyaline,
short, fragile.

On LILIACEAE: Fritillaria Meleagris L. in Europe.

TYPE LOCALITY: Southern Europe, on Fritillaria Meleagris.

EXsICccATI: Sydow Ured. 107; Thiim. Myc. Univ. 553, 728; Roum.
Fungi Gall. 2922.

This species differs in many respects from Uromyces aeciduformts
with which it has been included by European authors. The telio-
spores (Fig. 3) are larger, thinner walled and less prominently rugose
than those on Lilium.

Grateful acknowledgment is due Dr. J. C. Arthur for the unre-
stricted use of his herbarium material upon which this study is based;
to Professor H. S. Jackson, for his many helpful suggestions, the
writer is also deeply indebted.

PURDUE UNIVERSITY,

AGRICULTURAL EXPERIMENT STATION,
WEST LAFAYETTE, INDIANA

tility Pin pcipharium: ies ‘dag occurrence of self-fertile sine yer
ones booed progeny of, es stent plage nigee ihe s hye ay athe: i. TOUR BYR |

is oe Bara E. WHITE. 396

h

: influence ‘of light. ae chlorophyll Aceiatic on. the minimum toxic”
Oncentration of magnesium nitrate for the squash. -
| RB. HARVEY ‘dud R. H, ‘TRUE 407

é use ‘of the vibration’ galvanometer with: a 60-cycle alternating current in’
‘the measurement of the conductivity. of electrolytes. NEwToN B. GREEN. be
am inochemical studies of the plant Seat in. proteins of the wheat seed
oe wean and d other cereals. ‘Study PX. RaPs WODEHOUSE. 6) eel c.ccs «ve pes aT

ag alactoge and mannose che green anid and the antagonistic.
So sugars toward these. . Yo Says ae Dua Ghee Se rchluks KNUDSON 430

iN

“Tavis W. Bailey’ 3
«Bussey, Lastitution.

os ot BARTLETT, ORE

Pe ie of Michigan THON Sac! saa
“ . G Jounson, University of Wisconsin
ss eaare American Pegiepe teloaleal Soren,

‘The ‘Journal i is . pablished fnpukye except dhring Aa 1
, ear. Single copies 60: cents, plus po
Ie. charged . to-all” foreign ‘couhtries,, except Mexico, ,
Canal ee me of Panama Hawaii, ‘Ph i
‘ gee to’ Bel 20 « nts a volurr n
al Union, ” cents ay lume

Manassit offered for i pe ion sanwel

Paes

“Btusiaeons: a
“be reetiieted to approximately te
vay be ce bs cost rates =

eh Ger
e

_ Remittances should be made ae abl a © American. ournal ¢
; Age: City.

- Claims for missing H gumbers should be n
as te of faniling. The : ee ma supply mis

Nigel fares

e of change , a berg d¢

Noaeegn i

W609,

lyn Bot tanic Garden,

VOL. 4 PUN, LOW 7, No. 7

PemoncliyY IN ClCHORIUM INTYBUS: THE SPORADIC
OCCURRENCE OF SELF-FERTILE PLANTS AMONG
tHE PROGENY OF SELF-STERILE PLANTS

AB oTOur

The writer (1916) has already presented the evidence that the
. very prevalent self-sterility (and cross-sterility as well) in chicory can
be ascribed to a physiological incompatibility operating between sex
organs or sex cells that are fully formed, anatomically perfect, poten-
tially functional and of simultaneous development. It was noted that
this type of sexual sterility is sharply to be distinguished from sterility
due to anatomical incompatibility (more or less purely structural dif-
ferences and adaptations such as hercogamy), impotence (failure to
produce gametes) or embryo abortion (death of egg after fertilization
or death of young embryo). I also at that time discussed and sum-
marized the literature bearing on such phenomena.

From my studies made in 1912 and 1913, it appears (Stout 1916,
p. 365-366) that self-sterility is the ruleinchicory. Three plants (des-
ignated A, B, and C) of wild stock were found to be self-sterile, as
were 52 plants grown from the open fertilized seed of these plants,
and all plants tested of ten cultivated varieties were self-sterile. In
one variety (Barbe de Capucin), 29 plants of one planting and 5 of
another were tested, and of other varieties about five plants of each
were tested. The total of about 135 plants from these various sources
were self-sterile. ,

However, in the pedigreed cultures grown in 1913, a few plants
exhibiting varying degrees of self-fertility appeared quite sporadically
among the Fi progeny of various crosses between self-sterile plants.
Of the 75 plants derived by crossing plant A with plant E22 (of the
variety Barbe de Capucin), only eight were self-fertile. Of 21 plants,

[The Journal for June (4: 315-374) was issued June 29, 1917]
305

376 A. (Be SsrOule,

the offspring of A and £3, four were self-fertile. Seventeen plants,
the offspring of C and £3, were self-sterile, as were 30 plants from seed
of a cross’ between a white-flowered plant, (A XC) no. 1, of wild stock
and a plant of the variety ‘‘improved striped-leaf.’’ The number of
self-fertile plants, therefore, varied greatly in the different series, but
in no series was the proportion very large.

The self-fertile plants mentioned above appeared after only one
generation of ancestry known to be self-sterile. Furthermore, the
parents in each cross were not closely related and were somewhat dif-
ferent in vegetative habit and flower color. As previously recognized
(1916, p. 415), these results raised some question regarding the in-
fluence of wide-crossing as compared with that of inbreeding on the
development of self-compatibilities, especially as continued inbreeding
in the variety ‘‘red-leaved Treviso”’ had in two generations given only
one feebly self-fertile plant out of a total of 49 plants (complete data
given in 1916, Table 7).

In order to obtain further data on this question, it was planned to
continue inbreeding within this variety, increasing the number of
plants grown in I916, and at the same time to grow for comparison
an Fy, generation from crosses between plants of this variety and a
self-sterile plant of a wild stock. The present paper will deal especially
with the data obtained from these cultures.

DESCRIPTION OF THE CULTURES

The variety ‘‘red-leaved Treviso”’ is a cultivated salad chicory that
has been developed in continental Europe. As grown for commercial
seed production the variety is biennial, seed being sown one summer
for a crop that matures in the following summer. As grown in my
culture the plants are more nearly annual. Seed is sown in January
in flats, and the seedlings are potted and kept in continuous growth
in the greenhouse until spring, when they are planted in the field.
Under such treatment the plants, as a rule, reach full development in
the following August. The general habit of growth of the mature
plants is well shown in text-figure 1. The height has ranged from 44%
to 61% feet with the greater number of plants about 5 feet tall. The
plants are rather sparsely branched near the base but rather abundantly
branched above. In the early stages of growth the rosette leaves are
numerous, of large size, and erect. One of the marked characteristics
of the family I have grown is the development of a type of fasciation

FERTILITY IN CICHORIUM INTYBUS ae 377

Fic. 1. Typical plants of the variety red-leaved Treviso. The marker stands
by self-fertile plant (R. Ser. 10, No. 8); all other plants in the view were self-sterile.

From photograph taken in the afternoon when all flower heads are closed.

involving duplication and cohesion of the main axis. Two main stems
develop with a single root system. Occasionally these are separated

378 AB SOUR

from the root upward, but most usually the two are more or less fused
for a distance, the fusion finally becoming complete near the top of the
plant. None of the plants of this variety have shown any tendency
to live over winter. The maturity and death of the stems and branches
is accompanied by death of the roots. Several attempts to obtain
new plants from root cuttings taken at the time of the maturity of
plants have failed.

The wild white-flowered plant used in crosses with plants of the
red-leaved Treviso is perennial as are wild plants of chicory in general.
In the five years it has been under observation its mature height has
ranged from 214 to 2/4 feet. Its rosette leaves are few, much smaller
in size than those of the red-leaved Treviso, and are flat in habit of
growth. The branches are few and strongly horizontal, giving the
plant a sparsely branched and scraggly appearance.

The F, generation plants of the crosses between plants of the red-
leaved Treviso variety and the wild plant just mentioned were more
like the red-leaved Treviso in habit of growth. They were all blue-
flowered. Their height ranged from 4 to 6 feet, and they were abun-
dantly and profusely branched from the base. The degree of the
duplication of the main stem was much less than in the family of the
red-leaved Treviso. As shown in text figure 2, the plants of this hybrid
generation were large and well developed and of marked vegetative
vigor. They were far more robust and vigorous in growth than the
wild parent, and in respect to the degree of branching they were more
developed than plants of the red-leaved Treviso strain.

The sex vigor of these plants and of plants of the Treviso variety
in respect to production of flowers was commensurate with the vege-
tative vigor. From statistical data obtained in studies of flower
number, it was found that the total number of flower heads produced
by individual plants ranged roughly from 2,000 to 3,500 with the
average number of flowers per head at about 17. At the climax of
development as many as 100 to 150 flower heads opened in a single day.
These statements together with the descriptions given and the illus-
trations in the accompanying text-figures give some conception of the
full and complete sex vigor seen in the profuse production of flowers
that set seed when pollinated with pollen that was compatible.

It will also readily be observed that the inbreeding within the family
of red-leaved Treviso involved crosses between plants of close blood
relationship and of decided similarity, and that the fertilization in

FERTILITY IN CICHORIUM INTYBUS 379

Fic. 2. Typical plants of the F; generation of the crosses between plants of

the Treviso variety and the wild plant A.

these plants with pollen from the wild plant A constitutes a com-
paratively wide cross both in respect to blood relationship and to the
vegetative characteristics of the respective parents.

380 A. BB. STOUT

RESULTS OF THE SELF-POLLINATIONS OF THE I916 CROP OF
“IMPROVED RED-LEAVED TREVISO”’

A total of 103 plants were grown in this crop. All were descended
from two plants of the 1913 crop, and all but two had three generations
of parentage known to be self-sterile. The data for the self-pollina-
tions made on these plants are given in Table 1. Here the plants are
grouped, as they were grown, in series according to the immediate
parentage. The table gives the total number of heads upon which
controlled self-pollinations were made, the number of heads producing
no seed, the number of heads with seed, the number of seeds per head,
and the percentage of fertility. Frequently birds ate seeds, indicated
in the tables by “ B,’’ and thus interferred somewhat with the deter-
mination of the percentage of fertility.

The 10 sister plants of Ser. 7 were all self-sterile; of the 19 sister
plants of Ser. 8, one was feebly self-fertile; of the 25 plants of Ser. 9,
three were self-fertile; of Ser. 10, one of 10 plants was self-fertile; of
Ser. 11, five out of 19 were self-fertile; and one of the 18 plants of Ser. 12
was self-fertile. Of the total 101 plants descended from self-sterile
parentage, II were self-fertile in some degree. With the exception of
Ser. 7, one or more self-fertile plants appeared in each series. It may
be noted that three self-fertile plants, two of which were rather highly
self-fertile, appeared in Ser. 9, which was derived by crossing two
sister plants of the previous generation. This series was from a more
closely inbred parentage than were the other series.

The total number of flower heads pollinated in these series is 1205.
As a rule, not less than 10 heads were pollinated on a plant, and in
nearly all cases this total includes pollinations made on several dif-
ferent days. When it became evident that some plants were self-
fertile, special efforts were made to continue self-pollination on them
in order to secure an abundance of seed for future progeny. However,
_as shown in the table, the number of heads pollinated on plants com-
pletely self-sterile is also often high.

The degree of self-fertility, judged by the percentage of flowers
setting seed, varied considerably. Most of the self-fertile plants were
feebly self-fertile, producing as a rule only a few seed per head in
only a few of the heads manipulated. Others, as no. 34 of Ser. 9 and
no. 8 of Ser. 10, set seed in every head pollinated, and in numerous
heads the numbers were nearly equal or even equal to all that were

FERTILITY IN CICHORIUM INTYBUS

TABLET

381

Record in 1916 for Self-pollinations of the Cultures of “‘ Improved Red-leaved Treviso”

OWN HNBRW O

Record for Heads Pollinated

ome)

oe) Ho) (eee) fo) (2) 12)

oye) (OC e) (e) 1o) (eo) (oye ae exes AS) 12) TSAO) Me)

Ny
ONO OO

(ole) e) e} fe) femley (e)ie)

Seeds per Head. Remarks

Parentage, Series 4, no. 21. A self-

fertile plant

Parentage, Ser. 1, no. 8 XSer. 5, no. 5.

|Parentage, Ser. 4, no. I2 XSer. 5, no. 8.|

Winer Xe)

155,050; O15 Fon 9s 0% 95-9).0) OOO,
10 en Oe a ae a eens ed ie peed

Parentage, Ser. 5, no. I XSer. 5, no. 6.|

|
|

Fertility
(%)

9.06

0.50

382

Series
No.
“é

Series

vie eae
eee

mBWN

AB Su OUre

TABLE I—Continued

Record for Heads Pollinated

oe ee}

Total |

| oO
| Seed

With
N

OS 10 0,079

NO
foOoOOOCOCO 8°

oe)

CODODORPOWONODOMNONODOOSD

OFOEORORS

Seeds per Head. Remarks

ag baling Oy 2a ee aA

3s 6, 7) 9; 9, IO, IO, 10, Io, 10, Il, 12, I2,
12, 13; 13, 14,54) 14, 14) 14 Sano

Parentage, Ser. 4, no. 12 XSer. 5, no. 9

4,6; 8,88; 9;-0,;9,-10, 10,11, Tl wulat2,
E2012, 42. 13 STA tad SO neler,

Parentage, Ser. 5, no. I XSer. 1, no. 8

1,5
I, I, 3, 4,5

Parentage, Ser. 5, no. 9 XSer. 4, no. 18

Fertility
(%)

O71

0.63

0.03

0.03

0.04

FERTILITY IN CICHORIUM INTYBUS 383

TABLE I—Continued

Record for Heads Pollinated |
5 Fertility
Total With With (%)
No. No It Seeds per Head. Remarks
Heads | Seed Seed
NIGT O..1 9 9 O
Per es 10 10 O
eee) 13 13 O
en, © a2 12h Veo
tO ero 10 O
ee 13 Opts «4B eBs.3 By 3 By 3c By 5c By 778) 10;
10, II, 18, 22 O73"
re ee 2 10 10 fe)
ake 12 1 e)
LEO 10 10 O
He II II fe)
ee LO 10 10 O
ean Ee) 13 13 fe)
20 5 5 e)

possible. It is not to be considered that the degree of fertility is
absolutely determined, and especially in those cases when birds (Ser.
I2, no. I1) ate all or a part of the seed produced in certain heads.
The detailed data, however, make it quite clear that various degrees of
self-compatibility may exist. The evidence in this particular is quite
identical with that already reported in 1916. In Ser. II a compara-
tively large proportion of plants, 5 out of 19, were self-sterile, but the
fertility was low in each case.

The two plants of Ser. 6 were derived from self-fertilized seed of a
plant that was feebly self-fertile and which was the only self-fertile
plant that appeared in my crops of this variety previous to 1916. The
two plants were self-sterile.

SELF-COMPATIBILITIES AND INCOMPATIBILITIES AMONG PLANTS OF
THE I916 CROP OF F, GENERATION OBTAINED BY CROSSING
MePEANT (A) OF WILD STOCK WITH PUANTS OF THE
VARIETY “‘RED-LEAVED. TREVISO”’

The data for the self-pollinations of this generation are presented
in Table 2. The wild white-flowered plant A was the pollen parent
for Series 1-4 and the seed parent of Series 5. Five different plants of
the 1915 generation of ‘‘red-leaved Treviso’? were concerned in the
parentage, as indicated in the table. The uncertainties of securing
compatible cross-pollinations among self-sterile plants (Stout, 1916,

384

A. Bo STOUT

DABUES2.

Data for Self-pollinations of Fy generation Derived by Crossing a Wild White-
flowered Self-sterile Plant (A) With Self-sterile plants of the Variety ‘‘ Red-leaved

Treviso”’

Plant.

RA. Ser.

Oo CON DAuUHPW N HH

Data for the Heads Pollinated

Heads
With
No Seed

Heads
With
Seed

©516 © O'O7O © OC OO OF Oo

(SIN Ce) (ee) (2) 6) ) (eo) (oe) (e) lee) te) oe ioneyxe: O2O+O"S) ©

670-079

Seeds per Head. Remarks

Parentage, R. Ser. 1, no. 7XA

I, 1, 1,§

Parentage, R. Ser.I,no.2xA

Parentage, R. Ser. 1,no.6XA

vs als Sy lone 510
Parentage, R. Ser. 5, no. 16XA

Fertility
(%)

0.05

FERTILITY IN CICHORIUM INTYBUS 385

TABLE 2—Continued

Data for the Heads Pollinated
oa = Fertility
Plant Total | Heads | Heads (%)
No. With With Seeds per Head. Remarks
Heads |NoSeed! Seed |

No, .5.|-:252 12 fe)

ne Oe LS 15 fe)

Oe. LO 10 O

poe FO scl eha: 14 O

t= LO%! TO 10 oO

acldicils yn 13 O

ye 73 ae 76. 10 )

pies ec s E6) 10 fe)

ps 7 Aes Be 12 O

Beseae re ie II e)

ee LON Tho 10 O

PMIRG ECL Ly 5.3 Parentage, AXR. Ser. I, no. I

NOt -t.|> 10 10 fe)

ea ecco) Tal II O

Litton VEO IO fe)

peed 21 0 10 ce)

saa we estes 10) 10 re)

pe SOR TO 10 fe)

ee eel EO 10 re)

HeLEEOR ES 12 fe)

eee Oa te 12 fe)

/ 00>, 10 10 (e)

pee DE 10 10 O

ale ie) 12 fe)

pane be) 12 12 )

ORS Ge Me 3 12 oO

aera 13 ne fe)

ole 9 9 fe)

aes ii 12 12 0)

Sees fs: 8 8 0)

esto 12 12 O

Table 14) made it somewhat difficult to limit the parentage of the
various series to the same parents, which would, of course, be highly
desirable. Thus it happens that the immediate parents of these
series are not the same as those of the series reported in Table 1; the
plants involved are, however, closely related sister plants.

In making the cross-pollinations between the self-sterile parents
here involved, no attempts were made to emasculate or to depollinate
the seed parent. In brushing a flower head of a prospective seed
parent with a flower head from a plant selected for a pollen parent,
there was necessarily a full and rather thorough mixture of the two
lots of pollen with apparently an equal chance that both should be
distributed on stigmatic surfaces. A total of 54 plants were derived

386 A. B. STOUT

from four different seed parents of the red-leaved Treviso (Ser. I-4,
Table 2) and 19 were derived from the wild white-flowered plant as a
seed parent. All these plants were unmistakably hybrids. In no
case did a plant’s own pollen function in fertilization.

It may be noted here that East (1915) has made the suggestion
that the physiological conditions operating in self-incompatibility
involve a failure on the part of the plant’s own pollen to stimulate the
proper secretions in its pistil necessary for growth of the pollen tubes.
If this were the case, it would seem that self-sterility might be removed,
in part at least, by mixing pollen as I have done in the crosses referred
to above. Such, however, was not the result. It is possible that such
results might more readily be obtained in species in which the fertiliza-
tion processes are much less rapid than in chicory.

Of the 73 plants of this F; generation, only two plants were self-
fertile with percentages of 5 and 19. In only two cases were the
number of heads pollinated less than 10. The results are therefore
very decided. All of these plants were blue-flowered and were quite
similar in general vegetative habit and appearance. All flowered pro-
fusely throughout the season, and, as is the case with plants having
only this type of sterility (physiological incompatibility), all set abun-
dant seed in many heads open-pollinated.

Cross-INCOMPATIBILITIES AMONG THE PLANTS GROWN IN I9QI6

A brief summary may here be given regarding the results of cross-
pollinations made during 1916. Of the cultures of red-leaved Treviso
(R), 37 different plants were tested in a total of 34 different combina-
tions; of these 16 were sterile and 18 fertile in some degree. Among
the plants of the F, generation (RA), 24 combinations of cross-pol-
linations were made involving 37 different plants. Of these 9 were
sterile and 15 were fertile in some degree. As indicated by the figures,
the combinations among the R plants involved fewer plants propor-
tionally and more reciprocals than did those among the RA plants.
No particular emphasis can be given to the number here obtained in
its bearing on the influence of inbreeding or cross-breeding. The
data obtained from these plants selected at random, however, indicate
that cross-incompatibilities exist in marked degree. The results in
this respect are quite in agreement with those already reported (1916,
Tables 9-14), not only for the red-leaved Treviso but for other cultures
of chicory. 7

FERTILITY IN CICHORIUM INTYBUS 387

DISCUSSION AND CONCLUSION

The sporadic development of self-compatibility giving self-fertility
among the progeny of self-sterile lines of descent is in decided evidence
in the cultures reported above. No doubt if a larger number of the
‘“‘red-leaved Treviso’’ variety had been grown and tested, more than
one self-fertile plant would have been found previous to the crop of
1916. However, they were not found and the variety was kept in
pedigreed cultures by crossing self-sterile plants.

Self-compatibility is therefore a characteristic that was new in
expression, at least to the particular and immediate line of descent
involved. A total of 1o1 plants of the 1916 crop had three generations
of ancestry known to be self-sterile; of these 11 plants were self-fertile.

There is, therefore, much in the occurrence of these plants
that suggests discontinuous variation or mutation. However, the
fertilities of these self-fertile plants vary. They grade over to
complete self-sterility. The variation in the self-fertility of plants
grown from self-fertile parents (Stout, 1916, Table 6) is much more
continuous and is indicative that the irregular and somewhat dis-
continuous variation seen in the intensity of fertilities is only an
apparent one due to the few cases observed.

It is to be noted that there have been scarcely any attempts made
to study the progeny of self-sterile plants in species and varieties
known to be strongly self-sterile by continued inbreeding in pedigreed
lines of descent. Compton (1912, 1913) has reported that in Reseda
odorata “‘self-sterile plants when bred inter se throw self-sterile offspring
only,’ but he has not published data regarding the number of such
families, the number of plants, or the number of generations tested.
East (1915) has reported that the inter-specific hybrids between
Nicotiana forgetiana and N. alata grandiflora have been completely
self-sterile for four generations, and that a total of over 500 plants were
tested. Data on the behavior of the parent plants, or even of the two
parent species, were evidently not obtained. Correns (1912, 1913)
was especially interested in the study of cross-incompatibilities and
evidently tested the self-fertility of only 13 of the total of 60 sister
plants obtained by crossing two self-sterile plants of Cardamine
pratense. Of these, however, three plants appear to have been self-
fertile.

In view of the prevalence of self-incompatibilities in many plants
of economic importance, such as cabbage, rye, apple, plum, prune,

388 : A. B. STOUT

cherry, blueberry, etc., it is somewhat surprising that more searching
studies have not been made on the sporadic occurrence of self-fertile
plants. It is somewhat in doubt, therefore, whether there exists a
species, a variety, or even a strain of plants in which self-sterility due
to physiological incompatibility is absolute. However, such may exist
especially among certain hybrid strains as is suggested by East’s
data. Many further data are needed to allow of any adequate state-
ment of the various degrees and intensities of such self-sterility in
species or in different strains as a whole. The general evidence, how-
ever, suggests that in many such cases the sporadic occurrence of
self-fertile plants may be quite as it is in chicory.

The almost complete self-incompatibility of the F; progeny of the
crosses between plants of the “red-leaved Treviso’’ variety and the
wild white-flowered plant A is noticeable. The otcurrence of only
two self-fertile plants of feeble fertility out of 73 such plants emphasizes
the sporadic nature of the development of self-compatibility. This
may also be considered as evidence that wide crossing inside the
species does not especially favor the development of self-compati-
bility. In fact, a comparison of the behavior of these plants with
that of the 1916 crop of the inbred plants of the ‘“‘red-leaved Treviso”’
variety leads to the conclusion that inbreeding is more favorable to
the development of self-compatibility than is wide crossing. In East’s
results (1915) all plants tested, some 500 in number, of the Fy, Foe, Fs,
and Fy, generations were found to be self-sterile. As these were the
offspring of an interspecific cross, it does not seem that wide crossing
has here favored the development of self-compatibility. It should be
noted that East suggests that there may be some increase in the
development of cross-incompatibilities among the later generations,
which he considers may be due to an increased homozygosity, but
the evidence is not conclusive on this point.

I have not sufficient data to judge adequately of the frequency of
cross-incompatibilities among the various series and generations of
chicory grown. Cross-incompatibility has occurred in each generation
of the red-leaved Treviso variety (for data obtained in 1914 and 1915
see Stout, 1916, Table 14) as it has in all other families thus far tested
(Stout, 1916, Tables 9-13). Everywhere that I have tested for cross-
incompatibility in chicory it has been found to be very general and to
exist in various grades of intensity.

The numbers of self-fertile plants which appeared among the Fi

PERTILITY: IN -CICHORIUM INTYBUS 389.

generation of crosses between the wild plant A and plants of the cul-
tivated common chicory (£ Series) are somewhat higher than those
of the F, generation (RA plants) derived by crossing this same wild
plant with plants of the red-leaved Treviso here reported. The
strain (E) has not, however, been inbred in pedigreed cultures as has
the red-leaved Treviso strain, so there are less adequate data on the
comparative value of inbreeding and crossing with this variety.

The character of physiological self-compatibility giving fertility
appears in a very irregular and sporadic manner, and it exists in dif-
ferent degrees of intensity in different plants. It has appeared in
chicory in a family of the variety known as red-leaved Treviso after
three generations of self-sterile ancestry and no doubt would occur
with equal irregularity and intensity after many generations of such
ancestry. It seems very conclusive therefore that the causes of self-
incompatibilities are not to be ascribed to a similarity of nuclear
constitution involving definite hereditary units of germ plasm which
either directly determine incompatibilities (especially Correns’s view
of line-stuffs) or which indirectly determine them (East’s view).
Furthermore, the variability of the offspring grown from self-fertile
plants in chicory showsa very irregular inheritance of the characteristic
of self-compatibility and makes it quite clear that the expression of
self-compatibility is quite of the nature of a fluctuating variability,
and that self-compatibility and self-incompatibility, in chicory at
least, are not to be described in terms of dominant and recessive
characters which behave in any sort of Mendelian manner.

The evidence seems conclusive that the actual conditions giving
the various grades of self-compatibility, and of self-incompatibility
(undoubtedly there are various grades of incompatibility giving com-
plete sterility) as well, are decidedly individual. Various aspects of
this question in relation to conceptions of fertilization and to the
phenomena of serum incompatibilities have already been discussed
(Stout, 1916). It must be remembered that a plant whose two sets
of sex-organs are completely incompatible is itself derived from the
fusion of two cells that were compatible. The interactions between
pistil and pollen-tubes were compatible. The germ plasms of the
two sex cells were compatible in fusion, in the somatic life of the
diploid cell structure of the resulting individual, and in the more
intricate interactions involved in sporogenesis occurring in that indi-
vidual. Yet in cases of complete self-incompatibility none of the
pollen grains are functional on the pistil of the plant.

390 A.B STOUT

Such conditions emphasize the marked individuality of the develop-
ment of conditions giving incompatibility. The conditions are
fundamentally physiological and arise apparently in connection with
the differentiation of the two sets of so-called sex organs. Important
to an understanding of the facts of differentiation here involved are
the phenomena of cross-incompatibilities. Three sister sporophytes-
which are quite identical in all vegetative characters may possess sex
organs that are incompatible to the extent that complete self-sterility
is in evidence; no. I may be incompatible with the male sex organs
(microgametophytes and gametes) of no. 2, but compatible with those
of no. 3. This difference in relation is certainly indicative of dif-
ferences in the physiological qualities of the two lots of male gameto-
phytes. Conversely the microgametophytes and gametes produced
by a single sporophyte may act quite differently on the female sex
organs borne on two other sporophytes, being compatible in one case
and incompatible in the other. This indicates, likewise, a difference
in the condition of the two sets of female organs (including pistils).
Furthermore, the data as to the occurrence of cross-incompatibilities
in chicory even indicate that reciprocal crosses between two plants
may give quite the opposite results, showing that the relations of the
two sets of sex organs may not be interchangeable.

In such phenomena we may recognize a loss of sex-vigor which is
concerned with the function of gametophytes and gametes. The
decrease in fertility is entirely independent of a decrease in the pro-
duction of spores. Furthermore, there appears to be full and complete
development of the macrogametophyte and its egg; its development is
certainly not inhibited by the condition of the pistil in which it develops.
There is no evidence that the microgametophyte is not fully developed
with reference to its differentiation. Although often involving a
decreased vegetative growth of the pollen-tube, the inhibition appears
fundamentally to involve function.

The reactions involved in self- and cross-incompatibilities do seem
to involve, to some extent at least, as Jost (1907) and East (1915)
have especially emphasized, an interaction between the haploid
pollen-tube and the diploid tissues of the pistil. There may be some
question as to what extent these relations are involved.

Incompatibilities are evidently indicated not only by an inability
to produce embryos, but also sometimes by a feeble viability of those
that are produced. This death of embryos among seed produced by

FERTILITY IN. CICHORIUM INTYBUS 391

the self-pollination of different sister plants is quite as fluctuating in
degree as is the production of seed itself. In its effect it is often quite
like the conditions observed in the ‘zygotic sterility’? which Davis
(1915a, 19150, 1916) has observed in the Oenotheras, especially those
of hybrid origin. In chicory, however, the noticeable failure in seed
production suggests that much of the embryo abortion observed may
also involve a sort of sexual incompatibility. Embryo abortion, how-
ever, may be due purely to conditions of nutrition, especially in those
species which exhibit no physiological incompatibility.

The incompatibilities in chicory are obviously not purely a question
of haploid against diploid, but of a particular kind of haploid and
diploid relationship. In discussing these various points, the writer
(1916, p. 436-440) has pointed out that our knowledge of the physiology
of pollen-tubes is scarcely sufficient to decide whether the critical
point in the growth of the pollen-tube is determined by purely nu-
tritive reactions with the pistil as such or whether it is really deter-
mined by the diffusion of secretisns (hormones) from the macro-
gametophyte. The writer hopes to be able to state later somewhat
definitely from cytological investigation what the relative develop-
ments and nuclear phenomena in chicory are.

In discussing the various aspects of the relation of cell organization
to the development of compatibilities and incompatibilities, the writer
(1916, p. 416) has pointed out that the role of any particular combina-
tion of germ plasm elements, as far as can be judged by their expression
as characters in parents, in sister plants and in offspring, must be
quite secondary as far as incompatibilities are concerned to a more
general quality of the tissue and cell organization that develops in
connection with ontogenetic growth and development. ‘The con-
ceptions of Jost (1907), Morgan (1904, I910), and East (1915) are
fundamentally based on this same generalization as I there pointed out.

Much the same idea, if I understand their position aright, has since
been expressed by Goodspeed and Clausen in stating that such cases
of physiological incompatibility seem to involve ‘‘non-specific’’ dis-
turbances in the “reaction systems’’ (germ plasm) (1917a, p. 46).
These authors have embodied in the conception of “‘reaction systems”’
(1916, 1917a) a view which in some measure is a revolt against the
extreme formalism of the Mendelian factorial hypothesis, and in this
sense the conception is useful in the interpretation of the phenomena
of sterility especially of the type I have called impotence. In their

392 A Be SrOUr

application of this conception to the almost complete impotence of
the F, hybrids of Nicotiana Tabacum XN. sylvestris, they are dealing
with the well-known cases of degeneration so often observed during
sporogenesis in interspecific hybrids. They believe that the very few
perfect spores formed represent the Tabacum and sylvestris extremes of
a combination series. In other words, these few spores represent the
cases where the parental germ plasms segregated without mutual
influence. The greater number of recombinations, however, were
incompatible combinations of various elements derived from the two
germ plasms. ‘There are very few of the two original combinations
that survive reduction and sporogenesis. In somatogeneses the in-
compatibility is seen, they believe, in a complete dominance of the
Tabacum characters (1717@, 19170). Whether involving chemical or
mechanical reactions or involving differences in developmental ten-
dencies in the sense used by Tischler (1907), (Stout, 1916, p. 423-427)
such intra-cellular incompatibilities arise especially in the reorganiza-
tion of cells during or immediately following reduction as has long
been known.

In the case of physiological incompatibility, as in chicory, there
appears to be no impotence except of a purely accidental sort. Any
recombination system may survive, and in chicory sporogenesis in the
offspring of crosses between the red-leaved Treviso variety and a wild
white-flowered plant must, it would seem, give many new recombina-
tions. The range of these recombinations must be quite the same in
the various sister plants both of the F; generation hybrids and of the
various series of red-leaved Treviso. Yet for the self-sterile plants,
and these are here in greater number, all the pollen grains fail to func-
tion irrespective of the character of the particular germ plasm organ-
ization from which they came and of which they may be variously
composed. On the other hand in the self-fertile plants that are sister
plants of such self-sterile plants, germ cells of much the same hereditary
constitutions (as judged by the characters of the plants that bear them)
are compatible.

Furthermore, in the cases of self-fertility of any degree (or cross-
fertility as well), the evidence thus far obtained from hybrid genera-
tions does not indicate that the fertilizations involved selective or
preferential mating which favored fusion between particular recom-
binations of germ plasm with respect to hereditary characters.

The determination of whether physiological self- and cross-incom-

FERTILITY IN CICHORIUM INTYBUS 393

patibilities giving sterility involve similarity or dissimilarity of con-
stitutional organization is, of course, very fundamental to the under-
standing of the nature of fertilization. Although rather widely differ-
ing in particular applications, the conceptions advanced as to the
causes of physiological sexual self-incompatibility in such hermaphro-
dite plants as Eschscholtzia (Darwin, 1877), Cardamine (Correns,
1912, 1913), Reseda (Compton, 1912, 1913), Nicotiana (East, 1915),
and in such hermaphrodite animals as Ciona (Morgan, 1904, I9I0)
have in general agreed in considering that a similarity or lack of dif-
ferentiation is responsible for the sterility. The writer has already
(1916) discussed these conceptions and has presented for consideration
the view that the evidence is more readily to be interpreted on the
basis of the principle that in general a marked degree of similarity im
constitution is necessary for sexual fertility. In this relation it is to
be noted that inbreeding in the variety ‘‘red-leaved Treviso’’ has led
to asomewhat greater similarity in general characteristics than existed
in the original stock grown from commercial seed. In this sense the
continued inbreeding of sister plants has led to a greater homozygosity.
It is in the 1916 cultures of the offspring of inbred plants that self-
fertile plants appeared as noted above. As far as the results in chicory
extend, and it may be said that there are no more comprehensive data
to be had for any other species, the general results are not in disagree-
ment with the view expressed above.

The sporadic variability of the sex relations and their fluctuating
inheritance is very obvious in chicory. Self-fertile plants appear
irregularly among the offspring of wide crosses and among plants of
inbred strains which are prevailingly self-sterile. In both types of
offspring the number of self-fertile plants that appear varies con-
siderably. The manner of their appearance is not to be correlated
closely with similarities or dissimilarities as these are ordinarily
judged by the expression of characters. The condition of complete
functional sex vigor is in many hermaphrodites so complete that it
appears to be very definitely fixed in heredity. In chicory, however,
we see tnat highly individual and epigenetic developments may arise,
evidently in differentiation and in the transition to the gametophytic
stage, which lead to wide and sporadic variations in the functional sex
vigor.

The various phenomena of self- and cross-compatibility and incom- .
patibility raise many questions that are fundamental to an under-

394 ATE SLOUm

standing of morphogenetic differentiation involved in sexuality, but of
which we have at the present time only a superficial knowledge.

When does physiological incompatibility begin to develop? Is it
a steady and progressive development through the whole diploid
association of the two parental cell elements involved, or is it achieved
suddenly at some particular point in ontogeny? Also, when does the
sexual condition as distinct from the asexual condition actually arise?

Does incompatibility arise because of sex? Are the two the same?
It would seem most definitely that they are not and that incompati-
bilities are not merely due to sexuality. But even if independent,
where incompatibilities do arise, where, how, and to what extent are
they correlated with sex and is the development of the two ever
parallel? To what extent are the physiological interrelations of
sexuality and incompatibility dependent on such mechanical or
chemical interactions as are involved in reduction and sporogenesis?

Are the differences of intra-varietal physiological compatibility
and incompatibility (both self and cross) indicative of differences in
sexuality assuch? Are some of the organs of either sex (microgameto-
phytes and macrogametophytes with their respective gametes)
sometimes more sexual or of greater sex vigor than are others?

To what degree are the incompatibilities, and compatibilities as
well, determined by nutritive relations that are to be considered as
vegetative functions? Is sexuality in its origin and in its phenomena
of cell fusions, as some have held, to be considered in reality as a
phase of vegetative function? ‘To what extent are the sexual incom-
patibilities related to phenomena of serum incompatibilities and to
immunity and what are the fundamental reactions involved in the
development and operation of these?

These are among the fundamental questions that naturally arise
in connection with such sporadic behavior of functional sex vigor as is
seen in chicory in which self-fertile plants of varying degrees of fertility
arise among a progeny even after three generations of parentage
known to be self-sterile.

New YorkK BOTANICAL GARDEN

BIBLIOGRAPHY

Compton, R.H. 1912. Preliminary Note on the Inheritance of Sterility in Reseda
odorata. Proc: Cambridge Phil. Soc: 27 .Pt:i1-
—— 1913. Phenomenaand Problems of Self-sterility. New Phytologist 7: 197-206.

FERTILITY IN CICHORIUM INTYBUS 395

Correns, C. 1912. Selbststerilitat und Individualstoffe. Festsch. Med. Nat. Ges.
84. Versam. Deutsch. Naturf. Arzte.

—— 1913. Selbststerilitat und Individualstoffe. Biol. Centralbl. 33: 389-423.

Darwin, C. 1877. Cross and Self-fertilization in the Vegetable Kingdom. Edition

by D. Appleton Co., New York.

Davis, B. M. ig15a. A Test of a Pure Species in Oenothera. Proc. Amer. Phil.

Soc. 54: 226-245.

—— 19150. A Method of Obtaining Complete Germination of Seeds in Oenothera
and of Recording the Residue of Sterile Seed-like Structures. Proc. Nat.
Acad. Sci. 1: 360-363.

1916. Hybrids of Oenothera biennis and Oenothera Franciscana in the First and
Second Generations. Genetics I: 197-251.

East, E. M. 1915. The Phenomenon of Self-sterility. Amer. Nat. 49: 77-87.

Goodspeed, T. H., & Clausen, R.E. 1916. Hereditary Reaction-system Relations.

An Extension of Mendelian Concepts. Proc. Nat. Acad. Sci. 2: 240-244.
—— 1917a. Mendelian Factor Differences versus Reaction System Contrasts in
Heredity. Amer. Nat. 51: 31-46.

— 1917). The Nature of the F; Species Hybrids between Nicotiana sylvestris and
Varieties of Nicotiana Tabacum. Univ. Cal. Publ. 5: 301-346.

Jost, L. 1907. Ueber die Selbststerilitat einiger Bliiten. Bot. Zeit. 65: 77-117.

Morgan, T. H. 1904. Some Further Experiments on Self-fertilization in Czona.
Biol. Bull. 8: 313-330.

—— 1910. Cross and Self-fertilization in Ciona intestinalis. Arch. Entwickelungs-
mech. Organ. 307: 206-234.

Stout, A.B. 1916. Self-and Cross-pollinationsin Cichorium Intybus with Reference
to Sterility. Mem. N. Y. Bot. Gard. 6: 333-454.

Tischler, G. 1907. Weitere Untersuchungen iiber Sterilitatsursachen bei Bastard-
pflanzen. Ber. Deutsch. Bot. Ges. 25: 376-383.

°.

INHERITANCE OF ENDOSPERM COLOR IN MAIZE}!
: ORLAND E. WHITE

Few of the many characters in plants and animals studied by
geneticists during the last seventeen years are now to be regarded as
inherited in simple fashion. As more detailed and extensive studies
on the heredity of each type of character, once regarded as a simple
unit, are made, the more various the facts and the more complex the
interpretations have become. The present paper describes such an
increase in complexity of fact and interpretation in the heredity of
endosperm color in maize, which at the outset was regarded as a
single allelomorphic pair consisting of yellow and white, but which at
present involves possibly as many as four pairs of factors, one of
which brings about a dominance of white.

As studied and interpreted by Correns (3, 4) and Lock (10) and
others, yellow endosperm in maize is determined by the presence of a
factor for yellow, in the absence of which, the endosperm remains
white. Lock (10) found various degrees of yellow among the grains
classified as yellow, but all were easily separated from the white, so
that he regarded them as fluctuations of slight importance, and, in his
interpretation, did not distinguish between them. He notes, however,
that on the average, homozygote yellows are deeper colored than
heterozygotes. Lock studied very large numbers of Fe, F3, and F,
generation hybrid plants from crosses of yellow and white endosperm
varieties. Back-crosses of the yellow heterozygote with the recessive
white homozygote were also made in large numbers. The numerical
results are very slightly vitiated by the technique used, but the
numbers are so large as to make the small error from this source, in
this particular case, of comparatively slight importance. Lock’s
results are summarized in Table 1.

Further studies by East and Hayes (5, 7), Burtt-Davy (1) and
others on these endosperm colors brought to light a more complex
state of affairs, for they found two yellow endosperm ‘‘colors’’ in
maize, each behaving, when crossed with non-yellow (white) endo-

1 Brooklyn Botanic Garden Contributions No. 18.

396

INHERITANCE OF ENDOSPERM COLOR IN MAIZE 397

TABLE I
ie i | Ratios
Cross Yellow White Total (Ratios Act. Obt.' Picer
| xp.
Yellow heterozygote X white homo-
BNOue Wl) vce. ier ve cece ales .-| 1,963 | 1,982] 3,945 | 49.6 :50.4 | 50:50
Yellow heterozygote Xpure white
“VIF ISOS St erry Weems ee vacarane 26,792 | 26,751 | 53,543 | 50.03:49.97 | 50:50
White variety (homozygote) X yel-
lem: lreterozywote. oo... ae: 2,723 | 2,846] 5,569 | 48 :52 50:50 ©
IGNOU Aes NES A oe ese reg arya oe | 3478-1 31;579 63,057 | 40.6 750.4 || 50:50
F, white segregate xXpure white | |
ODL So AA A GR SE eae ea 59 all white ears | all
Yellow heterozygote (selfed)....... £6,592 [45,080 |:22,273 | 74.5 225.5 os

sperm varieties, as an independent allelomorphic pair. In the va-
rieties studied by East and Hayes, the two yellows were indistinguish-
able except that ears in which both were present in homozygous con-
dition were usually darker than ears homozygous for either one of the
yellows alone. Apparently either of the yellows, even in a homozygous
condition, could not be distinguished from the other, but, when both
were present in crosses with non-yellow races, the F, ratio approxi-
mated 15°Y :1 W. Designating the two factors as Y: and Yo, the
presence of either produced a similar effect. In some crosses, domi-
nance of yellow was complete and heterozygotes were indistinguishable,
while, in other crosses, as many as five shades of yellow were present
among the F, progeny, each shade signifying a difference in factorial
composition from darker to light in the following order: Y1:Y1Y2Yo,
YiYiy2ye or yiviYoY., VYiyiY oye, VYiyiY2Ye or VANoyay 9; Vuyiyeye Or
WiuyeYo. In Table 17, p. 56 (7) East and Hayes refer to an F,
population from white X yellow endosperm which gives only a 3Y:1W
ratio, but two shades of yellow are distinguishable—a dark and light
in the ratio of 1:2. This is interpreted as a I : 2 : 1 monohybrid
ratio in which the heterozygote is easily distinguished, owing to
imperfect dominance. The nature of the starch, whether soft (flour)
or horny (corneous), also causes a variation in the intensity of the
yellow endosperm color. From certain crosses, in a few cases, ap-
parently white segregate seeds, when planted, gave either pure yellow
or 2.Y :1 W progeny.
East and Hayes’s data are partially summarized in Table 2:

3908 ORLAND E. WHITE

TABLE 2

Segregation of Endosperm Color in Maize

SS

Cross Yellow | White | “Total-+Ratios Net OBR seme oe
| e

Exp.
F. generation of white Xyellow or
reciprocal or data on hybrid popu- |
lation:of same character... 4.52. 9,458 | ~-466'| 9,024 | 95.13.4290 93.8:6.2
(ES!)

F, generation of white Xlight yellow
or reciprocal or data on hybrid|
population of same character... ..| 6,792 | 2,428 | 9,220] 73.7:26.3 75:25

F; generation of white Xdark yellow, 5
or reciprocal or data on hybrid

population of same character....|2,376| 766 | 2, TAD Ih 7a eodas 75:25
F, generation of white Xyellow..... 609 dark yellow:1,143 light yellow: 589
white
PX PeCuCd einen hie cao ahaa 585.2 dark yellow:1,170.4 light yellow: 585.2
white

The varieties studied by Burtt-Davy (1) apparently consisted of
two distinct yellow types—a dark and a light, each of which with its
‘opposite’? represented an independent allelomorphic pair. The
darker of these yellows gives a I : 2:1 ratio, while “‘the other (the
paler) gives the ratio 9: 3:3: 1.’ The writer is at a loss to under-
stand the meaning of the above quoted statement, unless the two
yellows were crossed together and gave-in addition to three types of
yellow, whites in the ratio of 1W:15 Y. If each yellow is repre-
sented by an independent factor, crosses with white should give in
each case in F, only 1:2:1 or 3:1 ratios. This pale yellow has
sometimes been mistaken for a ‘dominant white.”” Burtt-Davy found
ten shades of yellow in the F3; seed generation from crossing yellows
with whites. Further, Burtt-Davy (p. 172, 173, 177, 188) refers to
a yellow endosperm color, which depends for its expression on the
presence of two factors—a color factor and a pigment factor. In
absence of either, the endosperm will be white, and by crossing two
white races, each carrying one of the factors, the resulting progeny
will all have yellow endosperm.

Emerson (8) obtained from F, populations of crosses of the orange
yellow Queen’s Golden with Black Mexican (white endosperm), two
yellow endosperm colors (a dark and a pale) in addition to the expected
orange and white endosperm segregates, and in F3, some of the pale
yellow segregates bred true, while others gave 3 pale yellow : 1 white.

INHERITANCE OF ENDOSPERM COLOR IN MAIZE 399

Collins (2) reports a déminant white (albinistic) mutation in en-
dosperm color. The mutation consisted of a single, wholly white
ear which appeared in a field of a carefully selected strain of dark
yellow dent with red cobs known as Gorham Yellow Dent.

The ‘‘albinistic’’ ear was fully matured, had white cobs, and the
seeds when closely examined had a faint trace of yellow at the base,
but would ordinarily be regarded as pure white. This white was
evidently ‘‘dominant’’ over the Gorham Dent yellow, since the original
mutant ear must have been in part, at least, cross-pollinated from the
surrounding plants—all of which were Gorham Dent. The immediate
descendants of this ear, either when selfed or cross-pollinated, con-
sisted of both yellows and whites, the former greatly predominating.
The yellows consisted of both dents and pale types, the former being
the more numerous. Seeds of the three color types were grown.
From 19 white seeds, 17 all white self-pollinated ears and two white
tinged yellow ears were obtained. From 16 light yellow seeds were
obtained 2 very light yellow ears, 11 ears with light yellow and white
seeds approximating in many cases a 3 Y :1 W ratio, I ear with dark
yellow and white seeds approximating a 3:1 ratio, and 2 ears un-
classified. Forty-nine dark yellow seeds produced 16 ears with yellow
of varying shades, 22 ears with both yellow and white grains in the
ratio of 3:1 (8,694 Y : 2,954 W), 4 ears with both yellow and white
' grains in the ratio of 15 Y:1W (2,548 Y : 177 W—theoretically
expected 2,555 Y : 170 W), and seven ears with all shades from dark
yellow to white. Crosses between a second generation progeny plant
from the albinistic ear (presumably heterozygous for yellow and white)
(no. 47) with two white seeded varieties of corn, in both cases gave
ears with both white and yellow grains—the ratios approximating
1:1. Selfed yellows (44 seeds) from various ears of the first cross
(no. 47 X White Dent) gave yellows and whites approximating a
ratio of 3 Y:1W (44 ears with 16,351 Y : 5,184 W, theoretically
expected 16,151 Y :5,384 W). Selfed yellow seeds from the second
cross (no. 47 X white Hopi) gave 2 ears with both yellow and white
seeds, the yellow predominating and one “pure white ear.’’ Seeds
of the selfed pure white ear gave 5 pure white ears and 4 faint yellowish
tinted white seeded ears. No data are given as to cob color in the
case of the second appearance of a white endosperm mutation, nor in
the case of the dozen or more pure white ears obtained in later gene-
rations. Since both the white cob and white endosperm appeared

400 ORLAND E. WHITE

together as a single mutation, it would be interesting to know whether
they were inherited together, and whether the white cob color was
also dominant over the red from which it sprung. Owing to ratio
discrepancies, and the occurrence of traces of yellow in descendants
of seeds classed as pure white, Collins regards the segregation of yellow
and white endosperm color as incomplete or imperfect. In other
words, factor contamination has occurred, though, in general, Men-
delian ratios were obtained. Interpreting his results as showing the
presence of at least two factors for yellow color and perhaps more,
Collins regards as both ‘violent’? and unwarranted. East (6, p.
404-405) however, in reviewing this paper, interprets Collin’s data as
demonstrating the presence of two factors for yellow endosperm color,
one of which is much less effective in producing the yellow color than the
other. East discredits the mutative reversal of dominance interpreta-
tion, suggesting the appearance of the original wholly white ear as due
to non-development of color brought about by abnormal environmental
causes, such as, perhaps, the “accidental presence’’ of some metallic
salt in the soil. This suppression of color development, East intimates,
is not extremely rare in experimental corn cultures.

NEw DaTA

The material consisted of an inbred strain of California Golden
Pop with yellow endosperm and a strain of white endosperm maize
obtained from Haage & Schmidt under the name of Zea Caragua.?
The latter bred true to a white endosperm intermediate between
flour and corneous in texture. A white endosperm variety of Hopi
maize, isolated from seed obtained from G. N. Collins, was also used.

In classifying the colors of the F, and F3 seeds, three methods were
used—(1) each seed as classified on the ear, was picked off and trans-
ferred to a black velvet background and contrasted either with the
parental varieties or with other pure white and yellow races; (2) most
of the seeds were classified independently by the writer and his assistant
Miss M. Mann; (3) both the writer and Miss Mann reclassified many
of the F, and F3 ears twice, first by re-examining both the yellow and
white groupings of each ear and second by mixing the two color
groups together again and reselecting. In each reclassification, the
ratios resulting from previous classifications were unknown. These
determinations were fairly accurate as demonstrated by the F» classi-

2 Described by Sturtevant (11). Apparently an old variety desseminated in
Europe by Vilmorin.

INHERITANCE OF ENDOSPERM COLOR IN MAIZE 401

fication and its F3; progeny. In some cases a photographic blue ray
screen was used.
F, GENERATION PROGENY

California Golden Pop pollinated by Zea Caragua gave uniformly
white endosperm grains with perhaps the very faintest suggestion of
yellow. This cross was repeated six times with similar results. The
reciprocal cross has not yet: been obtained. Crosses of Golden Pop
pollinated with white endosperm Hopi gave similar results.

F, GENERATION
Three independent classifications of the F. progeny from California
Golden Pop X Zea Caragua are given in Table 3. The first three
ears were classified with the aid of a photographic blue ray screen.

ABLE ?3
Endosperm Color of Fo Progeny of (Z 14XZ 21)

rst Classif. 2d Classif. 3d Classif. Total
Progeny No. Genae
White | Yellow] White | Yellow | White | Yellow

bo ALY SRO) at OO a aa ae 254 52) iieractically. 4) 5 wi, 336
a Bee) re Words, Orc eS 264 | 108 the same — | — | 372
Cais BP eet chiki isto vee oi 285 89 > us — | — | 374
Cae k= HORA RE ee 257 Wot 225. LOle |), 225 a0 Tit, 320
( i i Veer et tei har ws Sues 346 72> 200. fs1OO. 42607. 1522) AiO
( 2 ‘ Pe One ame ae: Bans oes 338 98 | 358 79 #|- 240 96 | 436
( ) a ahs eee eee age 346 77 a BTO. |) ehOO9-12 250. ke LOO | 425
es, 1 oO ARERR SOO SE a gE 248 7Out. 235 82 1274 | rod 1318
Ci ee ee en ee 250° | 146')|7 311= | 185/354 | 142 |, 496
Gt ee sO ia daate Pee cts S587 bei2o. |) 250" bals7 W220) Ve "hase 467
Gr Beh ee DOU, hi ocen ema taa A22, |. 126) AZO") #12814 410" W138) 548
(eon: Be OD ae rts as AQAT | LOOm 483: niet. | 472. | i132eiih G04
Ce BaD Bio hems vag as 336 Oise SUS TOO BES. | Tian
oe ae OU a ae Te ey at ATS LST, 3400.1 FAG) VeAgS Ale 177 4609
a Be GO hein een oe 300°). 143)| 406 | 127 | 388) 145. .| 533
Se OD wer Ue ara 373 Bae 303: | ELAM el Omen lr bacc4o7
ae MD ee 2a oe ele Vie irr ah B32F LShOAN 3325." OFM =43 0105. 486
peat We Jia el Tea thet Oe ae 427 Ta 37t L200" 396 || iii #1) 500
Ce De OC ge eal 221 52 | 209 64 | 211 62te|, 278
ey 1U7 Ve Sei 309 O31 205 We HO7s | eSOn leo.) -a02
cae oa OES eer 375 95 | 379 OL| 360% |" 1018) 470
( ak Wa ee eee 338 OF e326 e107. |e e624 b The Aes

3’The Fi; white endosperm of these seeds is indistinguishable from that of seeds
of several of the well known smooth white seeded varieties of pop-corn, when com-
pared with them. From data on non-guarded crosses referred to later, the recip-
rocal cross probably gives the same results, except the endosperm is opaque white
instead of translucent white.

4 Ears obtained from unbagged F; plants allowed to intercross with other Fjs of
similar pedigree.

5 One unclassified.

402 ORLAND E. WHITE

Owing to the increased experience, the third classification given in
Table 3 is probably the most accurate. Assuming that it is, from a
total of 9,663 progeny, 6,999 were classed as white and 2,664 as yellow.
On the assumption of a one-factor difference between the two maize
races with complete or practically complete dominance of the white
color, the theoretically expected numbers would be 7,248 W : 2,416 Y
(3:1). The deviation between the ratio actually obtained and
that theoretically expected is 249. The yellow segregates were far
from uniform in color, all shades from a dark yellow (not orange) to
a very light lemon yellow were present on the same ear. Further, in
some ears, the yellow was principally confined to the base of the grain,
nearest the point of attachment. A few dark yellow grains were
somewhat deeper colored than the yellow grand-parental California
Pop, but this may be due to segregation of various factors that effect
endosperm texture, as the dark yellow grains usually appear less
translucent than those of the grand-parental Pop variety.

BACK CROSSES OF F; WITH Z 21 (THE DOMINANT WHITE ENDOSPERM

PARENT)
(214 xX Z21)-20am- 72m All white
(LAL De 721) BAe Ziel. All white

These two ears, resulting from back crosses of the F, with the
dominant endosperm parent, came from two F, plants which also
produced two selfed ears (29) and 34a in Table 3). The back-crossed
ears were uniform in seed color, the white being more opaque than in
the endosperm of the F; grains. Nos. 29) and 34a gave typical Fe
ratios and the yellows were of several shades as in all the other ears
with Fe. seeds.

Fs; GENERATION

From self-pollinated ears Nos. 1, 2 and 3 (Table 3) of the Fe
generation, approximately 1,000 plants with F; endosperm seeds were
grown. Of these, 43 ears were self-pollinated, 27 of which came from
F, seeds classed as white, and 16 from F, seeds classed as yellow.
Nine of the white seeds gave all white F: progeny, while 19 gave both
white and yellow grains approximating the ratio of 3 W :1 Y (Table 4).
The 16 Fs» seeds classed as yellows gave 14 all yellow ears, and 3 ears
with both white and yellow seeds approximating the ratio of 3 W :1 Y.

INHERITANCE OF ENDOSPERM COLOR IN MAIZE 403

TABLE 4
F; Progeny of White-Seeded F2 Heterozygotes (Z 14 XZ 21)

Plant No. White Yellow Total No Grains

HM eR Wiese iE a eed) sn Cece oon saute setean ti et 269 55 324
1) Pte). ee ae Spee ee iit a eg Morn Aen RR UR er 294 98 392
1, SEROMA A PG al 119 38 L'57
PMN mera cae sada a aide <8 oh ae ae Ne 277 122 3909
ROE ke ong see RES ons Siar eM Macon s § 364° | 88 452
NSO Os caceetrcnusnec pat aie 476% felines eM a te 223, 74 307
Mae OOR 0 Accs e 2 Hed , toner yalend tap atals ties aise 298 108 406
HAN Geel OMe 4, ctr tay RM ere he oda Wit eee mec s 298 140 438
yen 0) Oa Oe Rng el et SR Se oe a 309 rh) 384
BN eA Utes, haath ahi, emtuian es giae tiene Ce es 113 | 48 161
Nig 94] eC 2 a ae ara Me a ig ee Sah Td aE 190 90 280
PMN eis ae cite ett ae Mae he where Meth 143 | 70 23
PRA) ail Newari ciecnuier es Dieccemreehses rs 119 39 - 158
ee en ies Tse: Ae chara? svn igiteligts 343 104 447
NN en oer ete ec ohne Wave pea a 3 216 56 272
2 MISO AR PL nn gan lane Se Pa 59 22 8I
a Ee Se PR eS SE eae 417 72 489
GB) NAVI ate ed Ue ee a ra aac ee a 252 78 290
MN ee MRM ec. Ret s alers ie this Parigalee cnalren ig the oP a 245 81 326
ee OO a PE ak SS Peo tiie FAR hae L352, | 47 222

iovalactually obtained’: cited 2c... « 4,703 | 1,505 6,208

Total theoretically expected......... A,050% fu) 552 6,208

Both yellow and white endosperm colors varied markedly in this
generation. In the case of white endosperm, the differences were
largely due to segregation of factors affecting the texture and degree
of translucency and opaqueness. Many ears had opaque caps, while
the remainder of the endosperm was corneous. In such cases, the
yellow was most apparent in the corneous region. ‘Translucent whites
such as one finds among popcorn varieties always appear slightly
yellow when contrasted with opaque whites such as are found among
the dent and wax varieties. No selfed ears were obtained of a deeper
yellow color than that in the California Pop ancestor. The all yellow
ears were of at least three distinguishable types: (1) a very light trans-
lucent lemon yellow, (2) a yellow as dark as the ancestral yellow and
(3) a yellow with opaque whitish caps.

UNBAGGED Ears OF Z14 AND Z 21

Unbagged ears on plants of Z 14 grown close to varieties with deep
yellow or orange endosperm color invariably have a large number of
dark yellow or orange grains, from which the dominance of these

°F, seeds probably wrongly classed as having yellow endosperm.

404. ORLAND E. WHITE

yellows over that of Z 14 (California Pop) is to be inferred, as bagged
ears are always of a uniform medium yellow.

Unbagged ears of Z 21 (Zea Caragua) on plants grown under con-
ditions similar to those mentioned for Z 14 have never been found in
my cultures with yellow grains. Further, cultures of Z21 grown
alongside of F; and F, generation hybrids (Z 14 X Z 21) have always
produced (in my experience) only white ears. The Z21 cultures
bloomed at about the same time as many of the hybrids, so that the
difference in flowering time would not account for the absence of
yellow grains.

INTERPRETATION

In the light of the preceding data, endosperm color differences
between Z 14 (California Pop) and Z 21 (Z. Caragua) may be regarded
as due to the presence and absence of a single factor A. The presence
of A prevents the development of the yellow color, when the factors
for yellow pigment are present, and gives no indication of its presence
in a variety from which these factors are absent. In the absence of A,
a given variety may be either yellow or white. In respect to this
factor A, then, and a single factor for yellow pigment, varieties of corn
may be of four kinds:

(1) AAYY (white endosperm)

(2) AAyy (white “abn )
(3) aaYY (yellow et )
(4) aayy (white ese)

Crossed with each other, these should give:

Cross Fi Fe
TWHOX SOW Ee ee white (ANYa) eee all white
TOW) XS GY oieewhites Aa OY) ae ee 2 Wil Y
LOW) XA Wiese white (Aa Viv Ga ie EEN CGN
2( WI >< OND Ae ee Winiter(Aa Vyas meta eae 13 Wie aXe
ZW) SCAND wine (Aaya)... ce eee all white
BNO DAW mene yellowa(aay yon cst Bus VeoT ANNs

So far as the data on Z14 X Z21 are concerned, California
Golden Pop would be represented on the above scheme as aaYY,
while the formula AAYY would be the only one applicable to Z 21
(Z. Caragua). All of the common white endosperm varieties of corn
which are wholly or partially recessive to yellow endosperm color

INHERITANCE OF ENDOSPERM COLOR IN MAIZE 405

have the formula aayy, neither the suppression nor the pigment factor
being present.

First generation (Fi) progeny from crosses between whites such
as I X 4, or whites and yellows such as 2 X 3, should give, when
backcrossed with white endosperm aayy (4) plants, white and yellow
endosperm seeds in the ratio of 3 W:1 Y. In other words, two whites
crossed together in F; give a certain proportion of yellows. The
obtaining of such results in partially worked out experiments on
heredity of endosperm color, in which it was taken for granted that
white endosperm color was always recessive, might be temporarily
interpreted as due to the presence of a color factor in a heterozygous
condition in one of the races experimented with. Perhaps Burtt-
Davy’s statements regarding the presence of a color factor for endo-
sperm (I, pp. 172, 173, 177) resulted from an experiment of this type.
I have not had access to papers with the data on which these state-
ments are based.

The preceding discussion assumes only one factor for yellow en-
dosperm pigment, whereas East and Hayes, Collins and Burtt-Davy
have each found at least two such factors. Further, Emerson and
East (9, p. 11) suggest that orange endosperm color, such as is char-
acteristic of Queen’s Golden Pop, Tom Thumb Pop, Yardstick and
some Chinese varieties (much intensified), is due possibly to the
presence of a color intensifying factor. The F: and F; data on crosses
of Z14 X Z21 show the presence of other color modifying factors,
especially one which dilutes ordinary yellow to a very pale lemon
color. Other investigators have also obtained this type.

SUMMARY

I. Crosses of a yellow endosperm variety of maize (California
Golden Pop) with a white endosperm variety (Z. Caragua) gave
uniformly white progeny in F; and a ratio approximating 3 W:1 Y
in Fs. The F, generation white grains, when planted, gave either all
white F; generation progeny or a mixture of white and yellow grains
approximating a ratio of 3 W:1Y. The Fy» yellow grains, except in
two cases, produced all yellow Fs self-pollinated ears. The yellow
grains in both the F2 and F3 generations varied considerably, and, in
F3, ears wholly of very light lemon yellow grains were obtained. Un-
protected ears of Z21 in close proximity to varieties and hybrids
having yellow endosperm always gave wholly white endosperm ears.

406 ORLAND E. WHITE

2. These results are interpreted as mainly due to the presence and
absence of an endosperm color suppression factor A. A factor Y for
yellow pigment is present in both races studied. Zea Caragua (Z 21)
is to be regarded as homozygous for both A and Y, while California
Golden Pop (Z 14) is homozygous for the presence of Y and the absence
of A.

3. The segregation of other endosperm factors, such as those for
flint and floury texture, opaque caps, etc., also modified the endosperm
color expressions.

4. Including the suppression factor A, at least three and possibly
five pairs of factors are primarily responsible for endosperm color in

maize.
LITERATURE CITED

1. Burtt-Davy, J. Maize, Its History, Cultivation, Handling and Uses. Long-
mans, Green & Co., London, pp. xl+831. Fig. 1-245. 1914.

2. Collins, G. N. Heredity of a Maize Variation. U.S. Dept. Agr. Bur. Pl. Ind.
Bull, 2722 1-22) 1912;

3. Correns, C. Untersuchungen iiber die Xenien bei Zea Mays. Ber. Deutsch.
Bot. Ges. 17: 410-417. 1899.

Bastarde zwischen Maisrassen mit besonderer Berticksichtigung der
Xenien. Bibliotheca Botanica 53: I-I6I. I9OI.

5. East, E. M. A Mendelian Interpretation of Variation that is Apparently Con-
tinuous. Amer. Nat. 44: 65-82. I9gI10.

Inheritance in Maize (Review). Bot. Gaz. 55: 404-405. 1913.

7. East, E. M. and H. K. Hayes. Inheritance in Maize. Conn. Agr. Exp. Sta.
Bull. 167: 1-141. Pl. 1-25. 1911. (Also Contrib. Lab. Genetics, Bussey
Inst., Harvard Univ. No. 9.)

8. Emerson, R. A. Latent Colors in Corn. Ann. Rep. Amer. Breeder’s Assoc. 6:
233-237. al Oils

9g. Emerson, R. A. and E..M. East. The Inheritance of Quantitative Characters in
Maize. Nebr. Agr. Exp. Sta. Research Bull. 2: 1-120. Fig. 1-21. Tables
I-39. I913.

10. Lock, R. H. Studies in Plant Breeding in the Tropics. III. Annals Roy.
Bot. Gardens, Peradeniya 3: 95-184. 1906.

11. Sturtevant, E. L. Varieties of Corn. U.S. Dept. Agr. Off. Exp. Sta. Bull.
5721-108. 21899,

THE INFLUENCE OF LIGHT AND CHLOROPHYLL FORMA-
TION ON THE MINIMUM TOXIC CONCENTRATION
OE MAGNESIUM... NITRATE FOR: THE SOQUASH*

R. B. HARVEY AND R. H. TRUE

In testing the absorption of magnesium nitrate by the squash
(Early Prolific Marrow), varying results in different series led the
authors to investigate the causes of these differences.

The results presented are taken from four experiments, the first
of which was run in full sunlight under a glass cover in the greenhouse

(ae as a ee ee Poe ee
RS ee ze
5 2 .a5hm SQUASH, MG NO) Lope IC St
LS] eS ee
5 RG a a a fn eee
Bs,
: 3.00000 oe ee =
pee Py | alae
Bae TT iocuee ele
nen ee | bre -|-
Dea ee Roe le | | a
a I “ 4 “a “a ‘ 2 3 ea I 6
Sa eee eee ie
aoe in Ge ee a
eae et aie
esleta caters = loot) s [27
2 See nee a
ea Ree fs Mea
ea aL ea
al Japs |
(Eine so Eee = BA
Leal leer | | zon Bal

Fic. 1. For explanation see text.

at a temperature varying between 17° and 30° C. This series (see
graph) showed a concentration of 2007 X 10-6 Mg(NOs)2 to be toxic
to the squash while 120” X 10-6 was not toxic. <A series run in the
dark at a temperature of 18° showed both the preceding concentra-
tions to be toxic.
* Published by permission of the Secretary of Agriculture.
407

408 R. B. HARVEY AND R. H. TRUE

SQUASH Mg(NO,),
(on ie

100. oe
NORMAL CONCENTRATION X 10

* Dark s:

i

Fic. 2. For explanation see text.

To eliminate the effect of temperature, the above series were
repeated, both at a constant temperature of 18° C. Concentrations
of exactly the same value were used in each. In the series exposed to

TOXIC CONCENTRATION OF MAGNESIUM NITRATE FOR THE SQUASH 409

the light, a diffused daylight intensified by a tungsten lamp was used.
Heat radiation from the lamp was minimized by a double glass window
in the constant temperature room. In every plant of this series there
was a deep green coloration equal to that of plants grown in daylight.
As is shown by the graph, a concentration of 150” X 10-6 Mg(NOs)o
is not toxic to the squash under the above conditions of illumination.
In the dark, however, this concentration is toxic. The toxicity here
is plainly shown both by the leach of electrolytes with an increase in
concentration as shown by the graph, and in the lack of root growth as
shown by the photograph. The growth of tops in the etiolated seed-
lings is of course greater than those exposed to light. The concentra-
tion 125” X 10-6 seems to be just at the border line of toxicity for
etiolated squashes. It, therefore, appears that such light exposure
and chlorophyll formation is accompanied by a rise in the minimal
toxic concentration of the solution.

Either of two conditions can produce this rise in the minimal toxic
concentration. Either the resistance of the protoplasm to the toxic
effect of magnesium may be increased, or the concentration of mag-
nesium within the cell may be reduced by light exposure. In regard
to increased resistance of the protoplasm no evidence is offered here.
It seems probable, however, that the decrease in concentration brought
about by light exposure may be sufficient to account for the change.
To become toxic the magnesium must reach a certain concentration
within the cell. From the etiolated series this minimum toxic con-
centration is seen to be a little below the equilibrium concentration
established within the cells in 120” &K 10-6 Mg(NQOs3)o. Since a con-
centration of 150” X 10-6 Mg(NQOs)e is not toxic to plants exposed
to light, the concentration of Mg(NQOs3)o actually in condition to
produce toxic effects within the cell is probably less than the minimum
toxic concentration found for etiolated seedlings. This decrease in
concentration may be brought about by the removal of magnesium to
form non-toxic compounds. One such group of compounds com-
paratively rich in magnesium whose formation in the squash depends
upon light is the leaf-green compounds found by Wiillstatter (1) to
consist of two parts, chlorophyll @ and chlorophyll 0.

From the work of Willstatter and others (2), it has been shown
that magnesium forms an important part of the chlorophyll molecule.
Mameli (3, 4), has shown that the presence of magnesium favors
chlorophyll formation.

ATO ee R. B. HARVEY AND R. H. TRUE

By means of the Grignard reagent Willstatter and Stoll were able
to introduce magnesium into the substance aetioporphyrin C3,H3.N4
to form aetiophyllin C3,H3sNaMg, one of the cleavage products of the
chlorophyll molecule. The results of experiments here presented seem
to indicate that the introduction of magnesium into the compounds of
the leaf takes place to a greater degree when there is sufficient illu-
mination to cause a green coloration, that is, the squash requires light
for the later steps of chlorophyll synthesis and these steps are asso-
ciated with the removal of magnesium from the field of toxic action.
No quantitative measure of the amount of chlorophyll compounds
present has been obtained on account of their instability and complex-
ity. However, calculations using the formulae found by Willstatter
indicate that the increased amount of magnesium used in the light is
well within the limits of the amount used for chlorophyll synthesis as
determined by Willstatter in nettle leaves.

In testing the toxicity of ferric chloride solutions under similar
conditions, no differences were observed between cultures grown in
the light and darkness.

SUMMARY

The minimal toxic concentration of magnesium nitrate for the
squash grown in water cultures was found to be 125” X 10-6 in the
dark and 200” X 10-6 in the light. The increase in the minimal toxic
concentration is probably correlated with the removal of magnesium
from toxic compounds to form chlorophyll.

BUREAU OF PLANT INDUSTRY,
U. S. DEPARTMENT OF AGRICULTURE.

LITERATURE CITED

. Willstatter, R. Chlorophyll. Jour. Amer. Chem. Soc. 37: 323. I915.
. Willstatter, R., & Stoll, A. Untersuchungen tiber chlorophyll. 1913.

* Mameli, E. ~Attt.soc, Ital. Procr. sci..5; 703. ole

. Mameli,E. Attr Ist. Bot: Univ. Pavia.(2)\15.) 1) (19t2.

BW NH &

THE USE OF THE VIBRATION GALVANOMETER WITH
A 60-CYCLE ALTERNATING CURRENT IN THE
MEASUREMENT OF THE CONDUCTIVITY
OF ELECTROLYTES

NEWTON B. GREEN

In 1898 Kohlrausch (1) published a description of his method for
determining the conductivity of electrolytes, and since that time much
has been done by various investigators to increase the accuracy of the
method. Notable among these are E. W. Washburn (2) and R. P.
Hibbard in collaboration with C. W. Chapman (3). Still more
recently has appeared an article by W. Taylor and S. F. Acree (5).
Among the sources of error, which they have removed, may be men-
tioned the following: an alternating current from an induction coil
which is neither strictly alternating nor of constant frequency; resis-
tance coils which are inaccurate because of capacity and inductance;
and lack of sensitivity in the telephone detector. At the present
writing a series of articles is appearing in the Journal of the American
Chemical Society by Dr. Washburn (4) which sums up the latest
researches on the subject. To these articles any investigators who
desire absolute accuracy of results are referred. The plant physi-
ologist is concerned more with precise comparative data than with
absolute physical accuracy, which must of necessity include experi-
ments extending over long periods of time and involving great elabora-
tion of method.

Dr. Washburn’s method overcomes the difficulties mentioned above
in the following manner: He uses for a source of current either the
Vreeland Oscillator, which gives a pure sine wave at a frequency of
one thousand cycles per second, or a constant-speed high-frequency
generator which delivers an alternating current at the same frequency.
Both of these pieces of apparatus and the principles involved are
described in Catalog No. 48 of the Leeds and Northrup Co. (6). For
resistance coils he uses the Curtis type, which have a minimum of
inductance and capacity, and for the detector he uses a telephone
receiver tuned to the frequency of the current. He also finds it

AII

412 NEWTON B. GREEN-

necessary in connection with the high frequency generator to maintain
the correct resonance in the bridge circuit by means of a double con-
denser. Consequently a complete outfit for making conductivity
measurements, using the generator and excluding bridge, conductivity
cell and resistance coils, which are of course necessary in any form of
apparatus, would cost over two hundred and fifty dollars. If the oscil-
lator is substituted as a source of current the price is increased to about
three hundred dollars. In either case a condenser must be used to
balance out the capacity in the conductivity cell.

Hibbard and Chapman have met the problems as to source of
current and detector in a different manner. They use a 60-cycle
rotary converter and an alternating current galvanometer of the
electro-dynamometer type. The latter costs about one hundred
dollars. In addition, a rheostat is necessary to regulate the primary
current. Consequently this apparatus exclusive of bridge, cell and
resistances must cost as much or more than that employed by Wash-
burn, without being applicable to as wide a variety of conditions.
The physical chemist will occasionally need a current of higher fre-
quency than 60 cycles, and such variation is impossible with the
ordinary rotary converter. Obviously the main objection to both of
these methods from the standpoint of the plant physiologist is the one
of expense involved. This is especially significant when one considers
that the apparatus employed serves one purpose only in the plant
physiology laboratory, namely, that of measuring changes in permea-
bility, or the electrolytic content of plant tissues and juices. Con-
sequently the writer considers it likely that an apparatus embodying
the latest methods of procedure, which fulfills all the requirements of
precision, will be welcomed by workers along this line.

Since the investigators, whose results are cited above, have con-
ducted the most exhaustive researches on the subject of conductivity
measurements, it is certainly desirable to follow any procedure which
they all recommend. We may assume then the necessity for a constant
temperature bath, in which to immerse the conductivity cell, and a
condenser to balance out the capacity in the cell. It is also certain
that the Curtis coils are the most reliable of all available resistances,
because they are so wound as to reduce inductance and capacity to a
minimum. By standardizing the apparatus to this extent we are
sure that the results obtained will have at least precise comparative
values. We now come to the question of the source and type of current

MEASUREMENT OF THE CONDUCTIVITY OF ELECTROLYTES 413

to be used. Washburn is of the opinion that to avoid undue polariza-
tion in the conductivity cell, a frequency of 1,000 cycles per second is
necessary. Moreover Taylor and Acree have shown that as the fre-
quency approaches infinity, variations in the resistance and capacity
of the cell approach zero. If we adopt this high frequency, the only
available type of detector is the telephone receiver. On the other
hand, Hibbard and Chapman, after exhaustive experimentation with
lower frequencies, assert that at 60 cycles per second, using a cell with
platinized electrodes, the amount of polarization is practically negligible
in all but a few exceptional types of solutions. If this is the case the
plant physiologist may feel secure in using this frequency, which has
several great advantages over higher frequencies as will now be
explained. An additional security rests in the fact that polarization,
if present, is easily detected in the ‘“‘creeping’’ of the balance point,
and can be immediately remedied by cutting down the amount of
current and the period of time in which the circuit is closed. The
main advantage in using a low frequency lies in the fact that another
detector than the telephone may be used. Such a detector is the
alternating current galvanometer, of which there are two general
types. The advantages of such a substitution are many and are fully
discussed by Hibbard and Chapman. All who have worked with the
telephone as detector will understand the difficulties attending the
constant strain of listening, and will appreciate the substitution of a
method which enables sight to take the place of hearing.

At this point the alterations in apparatus devised by the writer
may properly be considered. If a frequency of 60 cycles per second is
possible without a sacrifice in precision of comparative results, there
should be some source of current more available and entailing less
initial cost than the rotary converter. Such a source of current is
present in practically every laboratory, and needs only to be reduced
to the proper E.M.F. and potential. This is the ordinary 110-volt
alternating-current lighting circuit. As supplied to the laboratory it
is practically always a single-phase, 60-cycle system, having in most
cases a frequency variation of not more than one percent and a re-
markably pure wave form. ‘Taylor and Acree in their article have
inserted oscillograms of the Madison (Wis.) city current, which are
by no means exceptional, and can be duplicated elsewhere. For
example when this type of current is used to supply a bridge network
in which the bridge-wire has a resistance of 1.2 ohms, the current

AI4 NEWTON B. GREEN

can be sufficiently reduced by the insertion in series of two I6 c.p.
lamps. When the connections are made as in the diagram, the current
passing through the bridge wire, if R and R’ are open, is about .28
amperes at a potential of .336 volt.!. In practice the resistances
Rand R’ are in series with each other and the two connected in parallel
with the bridge-wire. Then when the bridge is balanced, the current

B’ BRIDGE WIRE. ia =

Fic. 1. For explanation see text.

divides itself between R-+ R’ and the bridge-wire, so that the two
divisions of the current are in inverse proportion to the resistances of
these parallel branches. Since the resistance of the bridge-wire is in
this case 1.2 ohms and the resistance of the unknown (R’) + its balance
resistance (2) varies between 200 ohms for concentrated solutions and
10° ohms for conductivity water, it is easily seen that the amount of
current passing through R’ will be exceedingly small in all cases,
running from .o016 ampere to 3.36 X 1077 amp. If however heating
occurs when a concentrated solution is being measured, it can easily
be obviated by the introduction of a rheostat in series with the bridge-
wire and lamps thus cutting the current down still further.

There now remains the consideration of the detector. For, fre-

1 The writer is indebted to Dr. Alan T. Waterman, of the Department of Physics
of the University of Cincinnati, for assistance in calculating the electrical data.

MEASUREMENT OF THE CONDUCTIVITY OF ELECTROLYTES AI5

quencies around 60 cycles the vibration galvanometer, as made by
Leeds and Northrup for seventy-five dollars, is the most sensitive.
It is easily tuned to the exact frequency of the current supply and once
tuned needs attention only on rare occasions. Its sensitivity is such
as to make profitable the use of the most accurate bridge with ex-
tensions on the bridge-wire. An added advantage is that the moving
coil returns quickly to its neutral position when the circuit is broken,
so that the band of light from the mirror follows closely in its width
the position of the slider on the wire. This enables a speedy deter-
mination of the balance point and cuts down the chance of polarization.

The total cost of such an apparatus using the best bridge, resis-
tances, conductivity cell and condensers will be much less than either
the Washburn or the Hibbard and Chapman outfits. This is made
possible by the substitution of the city current for an expensive piece
of apparatus, which is itself often a source of annoyance because of
noise. Moreover the vibration galvanometer is less expensive than
the electro-dynamometer type, and there is no sacrifice in precision.
The writer believes that these advantages will appear to be of distinct
importance to plant physiologists and to others interested in conduc-
tivity measurements.

SUMMARY

Since Messrs. Hibbard and Chapman have shown that polarization
is in nearly all cases a negligible factor using a current of 60-cycle
frequency, the ordinary single-phase, I10-volt, a.-c. lighting circuit
can be used as a source of current in making measurements of the con-
ductivity of electrolytes.

With such a frequency the most sensitive and convenient detector
is the vibration galvanometer.

The use of this method in preference to those previously known
enables the investigator, who desires only precise comparative results,
to make a considerable saving in first cost of apparatus without any
attendant sacrifice in accuracy.

BIBLIOGRAPHY
I. Kohlrausch, F. and Holborn, L. Das Leitvermégen der Elektrolyte. Leipzig,
1898.
2. Washburn, E. W. and Bell, J. E. An Improved Apparatus for Measuring the
Conductivity of Electrolytes. Journ. Amer. Chem. Soc. 35:174. 1913.
Hibbard, R. P. and Chapman, C. W. A Simplified Apparatus for Measuring
the Conductivity of Electrolytes. Mich. Agr. Coll. Techn. Bull. 23. 1913.

416 NEWTON B. GREEN

4. Washburn, E. W. The Measurement of Electrolytic Conductivity. Journ.
Amer. Chem. Soc. 38: 2431. I916.

5. Taylor, W. and Acree, S. F. Studies in the Measurement of the Electrical Con-
ductivity of Solutions at Different Frequencies. Jour. Amer. Chem. Soc. 38:
2396. I9QI6. |

6. Leeds and Northrup Co. Philadelphia, Pa. The Measurement of the Conductivity
of Electrolytes. Catalog no. 48. 1916,

UNIVERSITY OF CINCINNATI,
DEPARTMENT OF BOTANY.

(WNIT ONOCHEMICAL STUDIES OF THE: PLANT PROTEINS:
PROTEINS, OF TRE WHEAT SEED AND) OTHER
CERT Acss Ss LUDY mi?

R. P. WoDEHOUSE

That wheat foods are active in causing asthma has become an
established fact and it has been shown that watery extracts made from
practically any of the wheat foods except those cooked at very high
temperatures (Goodale ’16) will give positive skin reactions when
tested by means of the skin test (which is generally considered a test
for anaphylactic sensitization) to wheat asthmatics. This present
work was undertaken in order to find out which protein or proteins
of the wheat seed were responsible for the production of asthma by
isolating them individually in as pure a form as possible and testing
by means of the skin reaction.

The reserve proteins of wheat have been thoroughly studied by
Osborne and his co-workers (Osborne ’10, ’10A, ’07). By these in-
vestigators many different protein preparations were made from the
wheat seed and it was shown by very careful analyses of these that
there are five and probably only five distinct reserve proteins present
in the seed and these fall into the well-known protein classes as follows:

Albumin of wheat = leucosin

Globulin of wheat = wheat globulin

Prolamine of wheat = gliadin

Glutelin of wheat = glutenin

Proteose of wheat = cae Baar Sous
‘wheat artificial proteose

These proteins can be distinguished by their elementary composi-
tion and by their amino-acid content which these investigators have
worked out (Osborne ’10, Osborne and Clapp ’06) but they are most
readily distinguished by their solubility characteristics which are used
to place them in the groups of plant proteins to which they belong.
For convenience they are briefly here summarized as follows:

1 Made possible through a gift by Mr. Charles F. Choat, Jr., Boston, to the
Peter Bent Brigham Hospital for the study of bronchial asthma.

417

418 R. P. WODEHOUSE

Leucosin is soluble in pure water or in water faintly acid or alkaline
in reaction but is precipitated in an insoluble coagulum from a faintly
acid solution by heat.

Wheat Globulin. In neutral media this is soluble only in saline
solutions, the one generally used being Io percent NaCl.

Gliadin is soluble in 50-80 percent alcohol but insoluble in water
or absolute alcohol.

Glutenin is insoluble in water, alcohol or neutral salt solution but
readily soluble in weak alkali, 1/5 percent being generally used.

Proteose is soluble in water and not precipitated by heat.

All of the proteins soluble in water or salt solutions may be pre-
cipitated by saturating their solutions with ammonium sulphate.

Whether these proteins are chemical individuals or whether we
have to seek further separations is not definitely known. None has
been prepared, as yet, in a crystalline form. Though some vegetable
proteins have been crystallized, wheat globulin which, of the wheat
proteins, approaches nearest to this condition, occurs only in more or
less regular spheroids. :

There is pretty good evidence to show that all of these proteins,
including the “‘natural proteose,’’ exist preformed in the seed. Pro-
teoses are generally regarded, however, as being the first product of
hydrolysis of the higher proteins. Still Wells and Osborne (Wells
and Osborne, ’15) furnish good evidence that the “natural proteose”’
is not a product of hydrolysis but is a naturally existing protein of the
seed.

The artificial proteose was prepared by hydrolysis from glutenin (as
will be shown) and was used throughout the wheat experiments for
comparison with the natural proteins in general and more especially
to see if a proteose, known to be the product of hydrolysis, would give
biological reactions similar to those of natural proteose.

Besides these individual proteins there were used in this work two
“whole wheat’? protein preparations made, one from raw wheat and
the other from bread, by the author (Wodehouse, ’16). Of course
these two preparations do not, as their name might imply, contain
all five of the wheat proteins in anything like equal proportions. On
the contrary the solubilities, as indicated above, would prevent such
preparations from containing more than a trace of glutenin and
gliadin, and the bulk would be made up of the natural and decom-
position proteoses together with leucosin and globulin where they were
not precipitated by heat.

IMMUNOCHEMICAL STUDIES OF THE PLANT PROTEINS 4I9Q

METHOD OF PREPARATION
Proteins Soluble in Aqueous Salt Solution

Since these proteins occur mainly in the embryo of the wheat
grain and not very much in the endosperm (Osborne and Campbell,
00), ordinary white bread or pastry flour should not be used. The
commercial ‘“‘entire wheat”’ preparations give very good results except
that the yield is small.

Leucosin, being soluble in pure water, can be extracted from the
flour by simply soaking in cold water (with the addition of some pre-
servative as thymol or toluol to prevent bacterial decomposition) for
a few days, and decanting the supernatant fluid which then contains
this protein together with some globulin, dissolved by virtue of the
mineral salts contained in the grain, and proteose, together with
sugars, etc. This method of preparing leucosin was discarded early
in the work because it was found much easier to extract it together
with the globulin and separate them afterwards. So the entire wheat
flour was stirred in 10 percent salt solution (about 1,200 gm. to 3,500
cc.) and allowed to stand at room temperature for about three days
(no preservative is necessary). The flour settles to the bottom and
the supernatant solution (which is pinkish and syrupy) can be siphoned
off. It is desirable to allow the flour to separate completely from the
supernatant fluid so that further clearing will be unnecessary, for the
viscosity of this solution renders it difficult to filter. Less than one
half of the volume of the salt solution is recovered, the rest remaining
entangled in the flour, so it is profitable to make a second extraction
from the same flour by adding a volume of salt solution equal to that
removed by decantation. This second extraction is almost as rich
in protein as the first. Since the globulin is insoluble in water at
neutral reaction it may now be separated out by dialyzing the whole
solution in water until free from Cl (the other proteins may be separ-
ated from each other as will be subsequently shown), or all the proteins
may be salted out together by saturation of the extract with am-
monium sulphate, and separation effected by dialysis after again
being dissolved.

In the preparation of natural proteose it was found desirable to
follow the former method. When the salt extract is freed from NaCl
by dialysis all of the globulin and possibly parts of some of the other
proteins are thrown out of solution and can be removed by filtration.

420 Rei-P. WODEHBOUSE

If the filtrate be faintly acidulated and boiled the leucosin is coagulated
and forms a flocculent precipitate which can be removed by filtration.
The proteose, which now remains in solution, can best be obtained
by reducing the volume by boiling on the water bath and then dialyzing
against 95 percent alcohol when it appears in the form of a white
precipitate which can be washed in alcohol and ether and dried over
sulphuric acid giving a white powder. When prepared in this manner
it is perfectly soluble in water or 0.01 M KOH, giving a clear solution.

The globulin precipitated by dialysis from 10 percent NaCl solution
is largely insoluble when treated a second time with Io percent NaCl.
For this reason this method cannot be used advantageously for the
preparation of globulin. It was found best to follow the method of
first saturating the 10 percent NaCl extract with ammonium sulphate
thereby throwing out of solution all of the proteins together. This
precipitate can then be dissolved in 5 percent NaCl and the solution
dialyzed free from Cl and SO, when the globulin is thrown out of
solution in spheroids or imperfect crystals which can be separated
from the solution by centrifugalizing. Globulin thus prepared can
nearly all be redissolved in 5 percent NaCl and precipitated again
by dialysis. In making the globulin preparations used in this work
this was done twice in order to purify the preparations. It was then
washed in water, 95 percent alcohol, absolute alcohol and ether and
dried over sulphuric acid under diminished pressure. This preparation
is completely soluble in 10 percent NaCl or weak alkali. It appears in
more or less regular spheroids or imperfect crystals, so it can be con-
sidered to be reasonably pure.

Leucosin is very difficult to separate from proteose without coagu-
lation, therefore no attempt was made to prepare it entirely free from
proteose. The leucosin used in these experiments was prepared as
follows: The NaCl extract, after the salt and globulin had been re-
moved by dialysis was saturated with ammonium sulphate and the
proteins thereby precipitated were filtered out and pressed as dry as
possible between filter paper and redissolved in a small amount of
water, in which they proved to be almost completely soluble. This
solution was then dialyzed until free from the remaining ammonium
sulphate, or until it failed to give a precipitate with barium chloride.
This caused the production of a very small amount of an insoluble
protean which was filtered out. The solution was then dialyzed
against 95 percent alcohol until further reduced in volume. This

IMMUNOCHEMICAL STUDIES OF THE PLANT PROTEINS 421

caused the production of only a small precipitate, so the whole was
poured into three volumes of a mixture of acetone and ether (80-20)
and the precipitate so formed was centrifugalized out and washed in
alcohol and ether and dried over sulphuric acid under diminished
pressure. When desiccation was complete it formed a gray powder
which was more or less insoluble according to the time of exposure to
the alcohol baths.

PROTEINS INSOLUBLE IN WATER OR NEUTRAL AQUEOUS SALINE
SOLUTIONS

Gliadin and glutenin occur mostly in the endosperm of the seed
and are the proteins which make up gluten, the substance which gives
what flour its capacity for making dough. In order to obtain the
gluten the flour is mixed with water and kneaded into a stiff dough.
This is then wrapped in muslin and kneaded under water until the
starch is washed out. When it is mostly removed the dough may be
taken from the muslin and the kneading continued under water until
- no further starch can be removed. In preparing the proteins for this
work the gluten was next kneaded in several baths of 10 percent NaCl,
then chopped fine and allowed to remain in a large volume of salt
solution over night to complete the removal of the globulin, then
(still chopped fine) it was allowed to remain in running water for some
hours to remove the salt. |

In order to dissolve out the gliadin it was now subjected to ex-
traction with alcohol. This was done by boiling in 70 percent alcohol
on the hot water bath for about an hour, using a reflux condenser to
keep the alcohol from evaporating. It was then strained off through
muslin and the extraction of the undissolved gluten was continued
with a fresh bath of alcohol. In the meantime the first extract was
filtered and the alcohol distilled off by heating in a retort on a boiling
water bath. The alcohol recovered from this distillation was then
diluted with water to make again 70 percent by the hydrometer test,
and used for the next bath. This process was repeated five or six
times or until nearly all the alcohol soluble protein was removed from
the gluten. Care was taken during the evaporation of the gliadin
solution in the still not to let too much alcohol evaporate. No
attempts were made to see what percentage of alcohol remained or at
what temper: ture it was boiling. Distillation was discontinued,
however, while there was still enough alcohol left to lower the boiling

422 R. P. WODEHOUSE

point of the solution sufficiently to cause it to boil vigorously on the
hot water bath. In this way the risk of heating it to too high a
temperature was avoided.

When this gliadin solution was allowed to cool.a small part of the
protein settled out in a gluey mass at the bottom, and part assumed
the form of a fine suspension which would pass through filter paper
and could not be removed by centrifugalizing. So it was warmed up
enough to cause complete resolution and while still hot poured into
the dialyzers and dialyzed against tap water for three days, using thy-
mol as a preservative. At the end of this time the protein had all
settled at the bottom. The supernatant fluid was discarded, the
dialyzers torn open and the gliadin scraped off. At this stage the
protein was light gray in color and resembled malleable rubber in
consistency. It was thoroughly washed in distilled water, then cut
up into fine pieces and digested successively with acetone and ether,
absolute alcohol, ether, and dried over sulphuric acid under diminished
pressure. When desiccation was complete it was ground in a mortar
to a fine gray powder which could be used conveniently for making
the tests in these experiments.

The residue from the gluten, remaining after the extraction of the
gliadin, was now soaked in five or six volumes of 0.2 percent KOH to
dissolve out the glutenin. This solution formed a thick opaque white
fluid which could not be filtered; it was centrifugalized at high speed
for about one hour; this precipitated a considerable amount of insoluble
material. The supernatant fluid was poured off and very carefully
neutralized by adding 1 percent HCl. This caused the production of
a voluminous curdy. precipitate which reached a maximum at neu-
trality to litmus and would readily redissolve if made slightly more
acid. This precipitated glutenin was removed by centrifugalization
and the supernatant fluid, which was found to contain a large amount
of protein, was evaporated down to about one fifth of its original volume
on the water bath; the small amount of precipitate which had: formed
was discarded and the solution dialyzed free from KCl. Since no
further precipitate was formed the dialyzer was transferred to 95
percent alcohol which reduced the volume still further and caused the
appearance of a precipitate. Dialysis was continued in fresh baths
of alcohol and finally absolute alcohol and when thus dehydrated the
precipitate was removed, washed in absolute alcohol, ether and dried
over sulphuric acid under diminished pressure. This gave a fine white
powder which is called in these experiments “artificial proteose.”’

IMMUNOCHEMICAL STUDIES OF THE PLANT PROTEINS 423

The precipitated glutenin was now dissolved in 0.2 percent KOH
and centrifugalized to remove a small part that would not dissolve.
It was then precipitated again by neutralization with HCl, washed in
water, several baths of 70 percent alcohol to remove all traces of
gliadin, absolute alcohol, ether, and dried over sulphuric acid, giving
a fine white powder soluble in 0o.o1 M KOH.

The following table shows the reactions obtained from these
wheat protein preparations. The tests, the results of which are here
recorded, were all done under my observation by Dr. I. C. Walker
(nos. I-15 incl.), Dr. Turnbull (nos. 16-19 incl.), Dr. J. L. Goodale
(oss 20-22-incl,), Ors Fritz B:. Talbot (Mo. 23), and to them am I
indebted for the use of their results.

Briefly described, the skin test, by means of which these results
were obtained, is done by making small scarifications in the skin of
the inner side of the forearm and applying separately to these the dif-
ferent proteins which are then moistened with o.o1 M KOH. A
reaction is considered positive when an edematous swelling, which is
usually surrounded by a red areola, makes its appearance about the
scratch within a few minutes after the protein is applied. The in-
tensity of the reaction is gauged by comparison with a control scratch
upon which nothing but a drop of 0.01 14 KOH has been put. Fora
more complete description of the test the reader is referred to the
publications of the above mentioned investigators (Walker, this series
no. V’17, Turnbull ’16, Goodale ’16, Talbot 16).

In recording these results + is used to indicate a reaction scarcely
stronger than the control and should probably very often be regarded
as negative; + represents a quite definite reaction and the intensity
of the reaction is represented in an arbitrary fashion by the number
of plus signs, 4++ representing an edema about the size of a silver
dollar. The size of the reactions alone, however, is a very inadequate
comparison index of the anaphylactic activity of the proteins in
question. For this reason the proteins were dissolved in 0.01 M KOH
at a concentration of I percent and from this solution dilutions were
made, in the same medium, in the series I : 100, I : I,000, I : 10,000,
etc., and wherever possible the tests were repeated using the dilutions
instead of the dry proteins. Wherever this was done the lowest con-
centration to give a reaction is recorded in the table together with the
size of the reaction.

In all, about seventy patients were tested, but only those are

R. P. WODEHOUSE

424

~

‘JAI}ES9U 9q 0} PUNO}J a1aM Ady} UOYM JUIUI}ZeII} 107Je [IQUN Pd}S9} JON ¢

+ - - +. O O O = + one weiemetce N
++ ee tear ae per ean Ye
+++ fee Pte hee te lee | ee Jee foe sac nd

O + + + oO +4 ae toh, = see Noa 5
pope | eee jhe Paeede fe 8 S
OOI:I
O + + + O O O + Siler ’e7 te) e/a q
OO121= | 00171
Oo ++ ++ O O + Oo le ee J
+ O +4- O O O + GAP Onolg © YW
sree gs | hig (OR ee etre | here Bese: nea eee aes H
OOO‘I:I | OOOSI:1 | OOI:I |OOI:1}| OOOSI:I zs ooS: I
a + -+ -- + ae Saitireman cots ‘al
OOI:1 Se OOI:1 OOI:I
ae aE ar lip aRoE 4S a ae aF == i Bae gCT 5)
le OOI:I
O O O O a a + O =e = O = ae oq
OOOI:I | OOOI:I ooS: I
+++}/++]+ ) 0 + |++| + + | +] ++ | = 4h WH
oO 6) O + + oO O O oO Oo qi
O O O O O oO O + SE, ae O ry
O O + O O O O oti waleik sie: S'gq
0 ae ae al g e ¢ g O SAN
0 Bi le S g g g g Ie ale PAG!
ae 0 ae 0 g g g g Ale alg os IN
2 : ° 0 0 ais “lg te We la seca. = at AND
+ O O O O + Se ome Sie)
Oo 6) O O Oo as O ae O ae = Wwe
Oo= Oo= o=
O O O O O O + OOI:I} OOI:I OOI:1
miskat ote fetes ldecte aes ae ae gach S
3301 ula ure} 19101 19101 9s0a} I *JO01dg “WO1g
roe Peer Beye hoes: ae ean uioae Bete pacer ulsooney] | uluan[y ra ulyngoly ee eee jused

IMMUNOCHEMICAL STUDIES OF THE PLANT PROTEINS 425

included which gave a positive reaction with one or more of the wheat
proteins.

For the sake of comparison with the other cereals, corn, oat, rice,
barley and rye proteins were tested, using for this purpose the pre-
parations made by the author (Wodehouse, ’16).

The writer is well aware of the incompleteness of the records
shown in the following table but this is due to the fact that the in-
vestigation was carried out upon patients the number of which reacting
to wheat proteins was limited. Then with patients it is not possible
to repeat tests with the frequency and thoroughness that is possible .
when using animals for anaphylactic tests.

When this work was begun the writer was expectant of finding that
some one of the proteins of wheat was entirely responsible for its ana-
phylactogenic properties or else that they all behaved in the same
fashion and it was with the idea of isolating the ‘active principle”’
that the work was undertaken. ‘That no such simple state of affairs
exists can be seen from a glance at the table. On the contrary this
work shows that, though all of the proteins are capable of calling forth
anaphylactic symptoms, their method of action is so complex that with
our present state of knowledge it baffles explanation. Several in-
teresting obtentions to which attention is drawn in the following
paragraphs should, however, be noticed.

In many cases where the whole wheat preparation gives a doubtful
reaction or even in some where it gives a negative, some one or more
of the individual proteins give a quite definite reaction and always
one or more of the individual proteins are as active or, as 1s usually the
case, more so than the whole wheat preparations.

Of the total number of patients with which any part of wheat gavea
reaction :

Percent
Wi holeawheatprotein: 8.0). as ete cee hs reacted with 60
REAM OLCI | it ic cid. a corto nets whee sk . 45
CeO MUllutiemme cere: ay. . 3i2 atte, eee twas Paks ty 57
Sliver intavaanne ten tie oa 2) bl ace uct as a 31
SME CTT see Sed, “ir 6, sth ee ees Bains eS er 38
PE CHIC OSU eehe crane ig Oe a se cctsa pink ty 61
Natitralipnoctease:. 4.25... wuwne se als 2 vee ae ty 72
Antiicial proteose...% 5 ta sh wlbe «', see * 36

* These preparations were made by soaking the uncooked flour or meal of the
cereals in water until a solution rich in protein was obtained. From this the protein
material was precipitated by alcohol and the precipitate dried in alcohol and ether.

426 R. P. WODEHOUSE

In computing these figures doubtful reactions were always counted
as negative and this accounts for the small proportion (viz., 60 percent)
of reactions obtained with whole wheat. Since it is composed of
several different proteins this preparation would be expected to have
proportionately more chances of producing reactions. On the other
hand, however, being a mixture the active proteins would be diluted,
to a large extent, by the inactive proteins and with cases with which
the active proteins are few in number or weak in reaction this dilution
might be sufficient to obscure, their activity almost to, or even below,
the limit of sensibility of the skin test, thus accounting for the large
proportion of doubtful or negative reactions of the whole wheat while
with the same cases at the same time the individual proteins reacted
quite strongly.

In case No. 10 is seen a good example of this. Here ‘natural
proteose’’ is the only one active but in the whole wheat preparation
it does not call forth a response because its activity is obscured by the
other four. With the case of No. 13 “natural proteose’’ and gliadin
are the active parts but their activity is masked by the three other
inactive parts. Except to Nos. 1 and 4 this explanation can be applied
to all. However it is only tentative and a definite explanation must
await further investigation.

An even more interesting result to be observed here is that the
‘natural proteose’’ is the most ative, producing reactions with 72
percent of the cases while the ‘‘artificial proteose’’ only shows activity
towards 36 percent. ‘This shows that these two proteins are not im-
munologically alike and lends support to the contention of Wells and
Osborne (Wells and Osborne, ’15) that the ‘‘natural proteose’’ exists
preformed in the seed and is not formed, as is the ‘‘artificial proteose”’
by reagents used in extraction and purification.

It is also to be seen that heating to a cooking temperature does not
destroy the anaphylactogenic properties of wheat. However the
heating employed in the cooking of bread somewhat reduces its ac-
tivity in most cases. Nevertheless with some the reverse is true. In
order to test this further a concentrated watery extract of flour was
boiled for several hours. Another was heated in the autoclave at a
temperature of 114° C. and a pressure of 15 pounds per square inch
for one hour. When the coagula formed by the heat were filtered off
and skin tests performed with the filtrates it was found that neither
heating to a temperature of 114° C. nor prolonged boiling had reduced

IMMUNOCHEMICAL STUDIES OF THE PLANT PROTEINS 427

their activity in the slightest degree. Nevertheless when the whole
seeds of wheat and some of the other cereals were heated in a crucible
until they became a light tan color all through, they were found to
have completely lost their activity. This was also found to be true
of the prepared cereal foods, which are said to be cooked at tempera-
tures much exceeding any reached by an ordinary autoclave, such as
‘“‘Puffed Wheat,”’ ‘‘Puffed Rice,’’ Kellogg’s “‘Toasted Wheat Biscuit,”’
“Shredded Wheat,” etc. When concentrated aqueous extracts made
from these were tested’ by means of the skin test upon patients who
were strongly sensitive to the corresponding cereals in the raw form
no reaction whatsoever was obtained. In this connection it is inter-
esting to note that in the preparation of the most active of the in-
dividual wheat proteins (viz., ‘‘natural proteose’’) considerable
boiling is employed. From this it is seen that only very high temper-
atures tend to diminish the anaphylactogenic activity of wheat and
of some of the other cereals in relation to sensitization as revealed by
the skin test.

In cases sensitized to the pollens of the Gramineae it has been
pretty definitely shown that idiosyncrasy to pollen of one species of
grass is almost always accompanied by sensitization to the pollens of
all the grass family (Goodale, ’15). However, in cases allergic to the
seed proteins of the Gramineae we see that this is not generally so,
though sometimes it may be, especially with cases highly sensitized.
This is entirely in keeping with Nuttall’s findings in the immunological
relationships among the animal proteins. He says: ‘‘The more power-
ful the antiserum obtained the greater its sphere of action upon the
bloods of related species. For instance, a weak anti-human serum pro-
duced no reaction with the blood of the Hapalidae, whereas a powerful
serum did produce a reaction’’ (Nuttall, ’o1). This is confirmed by
Uhlenhuth (Uhlenhuth, ’o1) in experiments upon the relationships
between the ox, goat and sheep.

The question as to what extent subjects which are hyper-sensitive
to the seed proteins of the Gramineae respond to the pollens of this
family should be further investigated. It is interesting to note in
passing that wheat pollen was entirely negative with the one case
upon which it was tried although this case was extremely sensitive to
the proteins of the seed. The same preparation of wheat pollen, how-
ever, gave good reactions with some grass hay-fever cases.

4 These tests were made by Dr. J. L. Goodale with the materials prepared by the
author.

428 R. P. WODEHOUSE

SUMMARY

The five proteins globulin, gliadin, glutenin, leucosin and natural
proteose were prepared from wheat according to the method of T. B.
Osborne, and when they were compared in their anaphylactogenic
properties with each other, with an artificial proteose prepared by
hydrolysis from glutenin, with the whole wheat preparations and with
the proteins of other cereals, it was found that (1) all are anaphylacto-
genic, but no two are immunologically exactly alike, (2) the natural
proteose is the most active, (3) the natural proteose is different from
the artificial proteose, (4) in any given case where whole wheat gives a
reaction and in some where it does not some one or more of the in-
dividual proteins are sure to be found to be more active, (5) it does
not necessarily follow that because a case is allergic to wheat it will
be found to be also hypersensitive to the other cereals (though this is
sometimes the case especially if sensitization is of a high order), (6) it
probably does not follow that sensitization to the seed proteins of
cereals necessitates sensitization to the pollens of the same species,
though not enough experiments were done upon this to more than
suggest that it is a problem that ought to be further investigated.

It is also shown that heating, except to very high temperatures,
does not materially affect the anaphylactogenic properties of the wheat
proteins.

MEDICAL CLINIC OF THE

PETER BRIGHAM HOSPITAL,
Boston, Mass.

LITERATURE CITED

Goodale, J. L. Pollen Therapy in Hay-fever. Bost. Med. Surg. Journ. 175: 2.
1915.
Diagnosis and Management of Vasomotor Disturbances of the Upper Air
Passages. Bost. Med. Surg. Journ. 175: 6. 1916.
Nuttall, G. H. F. The New Biological Test for Blood in Relation to Zoological
Classification. Proc. Roy. Soc. Lond. 69: 150. 1901.
Osborne, T. B. Die Pflanzen Proteine. Ergebn. Physiol. 10:47. IgIo.
Darstellung der Proteine der Pflanzenwelt. Abderhalden, Handb. Biochem.
Arbeitsmeth. Berlin. I9gIo.
Proteins of the Wheat Kernel. Carnegie Inst. Publ. 84. 1907.
Osborne, T. B. and Clapp. Chemistry of the Protein Bodies of the Wheat Kernel III
Hydrolysis. Amer. Journ. Physiol. 17: 231. 1906.
Osborne, T. B. and Campbell. The Nuclei Acid of the Wheat Embryo and its
Proteid Compounds. Journ. Amer. Chem. Soc. 22: 379. 1900.

IMMUNOCHEMICAL STUDIES OF THE PLANT PROTEINS 429

Talbot, Fritz B. Asthma in Children. II. Its Relation to Anaphylaxis. Bost.
Med. Surg. Journ. 175: 191. I916.

Turnbull. Anaphylactic Action of Grains in the Respiratory Tract. Bost. Med.
Surg. Journ. 175: 266. 1916,

Uhlenhuth, P. Weitere Mittheilungen iiber die Praktische Anwendung meiner
forensichen Methode zum Nachweis von Menchen- und Thierblut. Deutch.
Med. Wochenschr. 27: 499. I9QOI.

Walker, I. C. Studies on the Sensitization of Patients with Bronchial Asthma to the
different Proteins in Wheat. Journ. Med. Res. 35: 3, 509. I917.

Wells, Gideon and Osborne, T. B. The Biological Reactions of the Vegetable
Proteins. VI. The Anaphylactic Reactions of the Proteoses of Various
Seeds. Journ. Inf. Dis. 17: 259. 1915. :

Wodehouse, R. P. Preparation of Vegetable Food Proteins for Anaphylactic
Tests. Bost. Med. Surg. Journ. 175:6. 195. 1916.

THE TOXICITY OF GALACTOSE AND-MANNOSESEOR
GREEN PLANTS AND THE ANTAGONISTIC AGiaon
OF OTHER SUGARS TOWARD THESE:

LEwIis KNUDSON.

In experiments concerned primarily with the utilization of certain
sugars by certain green plants (Knudson, 1915, 1916), the noteworthy
fact developed that, while other sugars may be of benefit, galactose is
toxic. The injurious effect was manifested in a killing of the root or a
retardation of root growth, depending upon the concentration of
galactose employed. It was observed, furthermore, that glucose
can antidote the toxicity of galactose, but this antagonism occurs only
when the glucose is present at a concentration equal to or greater than
that of the galactose. So far as the writer has been able to determine,
this is the only recorded case of a hexose sugar being injurious to
plants and of antagonism among the sugars.

In view of the fact that glucose exhibited such a marked antago-
nistic action toward galactose, it seemed advisable to extend further
the investigation to include various other sugars. A considerable
number of experiments have been made to this end and the paper
here presented records briefly the results obtained.

Methods.—For the experiments either Canada field pea (Pisum
arvense L.) or wheat (Triticum sativum L.) was used. ‘The plants were
grown in all cases under conditions insuring freedom from micro-
organisms. For this purpose the plants were grown in culture tubes
200 mm. X 20 mm. in size, on a nutrient agar medium. Pfeffer’s
nutrient solution, slightly modified, was made up as follows: Ca(NOs)s,
4 grams; KNO;, 1 gram; KeHPQO:, 1 gram; KCl, 0.5 gram; MgSO,,
0.5 gram; FeCl;, 20 milligrams; distilled water, 12 liters. The agar
used had been previously rinsed three times in distilled water and then
air-dried. One percent of agar was used. The medium is faintly
alkaline to methyl red. The different sugars were dissolved in this
medium and stock solutions were made up of double the concentration
of sugar used in the experiments. Dilution of the sugar was effected

1 Contribution from the Laboratory of Plant Physiology, Cornell University.
430

TOXICITY OF. GALACTOSE AND MANNOSE FOR GREEN PLANTS 43I

by addition to the nutrient solution or by addition of another sugar
solution. For example, to obtain a solution containing 0.25 mol.
galactose + 0.025 mol. saccharose, it was necessary to mix equal parts
of 0.5 mol. galactose and 0.5 saccharose. The volume of the medium
in each tube was 25 cc. Sterilization was effected by autoclaving at
fifteen pound§ pressure for fifteen minutes.

All the sugars used, with the exception of arabinose, were supplied
by Dr. C. S. Hudson, in charge of the carbohydrate laboratory,
U.S. Bureau of Chemistry, and are stated by him to be of very high
purity. The arabinose used was a Merck reagent.

Character of the Injury.—The injurious action of galactose is made
evident first in the roots. The primary root coming in contact with
the agar may first become brown and in a few days death results. In
other cases the tip of the root is killed and this stimulates the pro-
duction of a large number of lateral roots, the tips of which, on coming
in contact with the agar medium, are soon killed. A short primary
root with many laterals results, the appearance of which is somewhat
centipedal. Two plants in the same culture may, however, vary in
the manner of injury, and the presence of certain sugars may alter
the extent of the injury.

For the sake of clearness and definiteness, it seems desirable to
describe the injury by a numerical system as well as by root lengths.
Accordingly the following key is given: 0, no injury; I, primary root
tip killed, laterals not injured; 2, the primary root tip may be killed,
but the laterals may attain a length of a few centimeters and then
growth is stopped or the roots are killed; 3, the primary root may
penetrate the agar, but becomes brownish and five or six centimeters
long; 4, the primary root may attain a length of a few centimeters, but
becomes brown in color and the laterals do not grow beyond 0.5 cm.;
5, the primary root tip is killed and all laterals suffer likewise; 6, the
_ primary root is entirely killed.

Antagonistic Action.—In the following experiment the galactose
was supplied at a concentration of 0.025 mol. and the other sugars
were used at the same concentration. In order to demonstrate con-
clusively that the total concentration was not responsible for any |
toxicity, a few cultures were made with the nontoxic sugars supplied
at 0.05 mol. The experiment was begun on January 29, 1917, and
concluded on February 13, 1917. The cultures were placed:in the
greenhouse and grown in the light. All cultures were made in trip-

432 LEWIS KNUDSON

licate, but contamination or failure to germinate caused a loss of
some of the cultures. The seed were sterilized by immersion in a
solution of calcium hypochlorite (calcium oxychloride, Baker) ac-
cording to the method of Wilson (1915). The peas. were treated for
two hours and the wheat for five hours. The results are given in
Table 1.

TABLE I

Influence of Sugars on the Toxicity of Galactose

Length of Averare | 1B h of Cl
The Concentration of Each Sugar Equals 0.025 Mol. Primary Of gee | Tap (emn.) oF me

Root (Cm.) Root (Cm.)
Galactose (3. Cultures): .ance an hate meee I iy 6 5
Galactoses(@rcultures)r sfc teste een none O fe) 3 6
Glucoses(egcultures)e ne ee es 10 8 15 fe)
Levulose,@) cultumes 0) Waucce o saee cee: 8 8 14 re)
ArabinosesGr culture) se 525 ta). 2 ee a 10 7 15 O
Saccharoses(sicultunes) sal, v2.0 ae ee 9 8 14 fe)
Maltose® (a:cultures m4 2s.5-satae eee ae IO 8 13 e)
Rafiinose (2ecultunes) s0..0.0..6) nls ene ie ee 8 14 O
Pfeffers, no‘sugar (3cultures)tic.- Gea. 4 10 8 14 ce)
Galactose+glucose (3 cultures).......... hue eG 7 13 O
Galactose-+levulose (2 cultures) ......... 5% yy 10% 4
Galactose+levulose (1 plant)............ 7 re) baie ha 5
Galactose+saccharose (3 cultures)....... 2 7 13 I
Galactose-+saccharose (I culture)........ 10 9 14 fe)
Galactose-+lactose (3 cultures).......... I yy 9 5
Galactose-++maltose (2 cultures)......... | I yy 10 5
Galactosé--ratiinose ye Wee wtie Aa eae Ee, y% 6 | 5

From the table it will be noted that the toxicity of galactose is
prevented by glucose or saccharose, the former being slightly more
effective than the latter since the primary root is not killed in the
presence of glucose. None of the other sugars are effective in pre-
venting the injurious action of galactose, although in the presence of
levulose the primary root may continue its growth to a limited extent.
Representative cultures are shown in Fig. 1.

All of the preceding experiments except those with levulose were
repeated and similar results were obtained.

In some earlier experiments (Knudson, 1916) it was noted that
glucose does not antidote galactose if the concentration of the former
is less than that of the latter. It was thought that some relation might
be found between concentrations and antagonistic action. Accord-
ingly the galactose was supplied in each case at a concentration of
0.0125 mol. solution, and the other sugars used at double this concen-

TOXICITY OF GALACTOSE AND MANNOSE FOR GREEN PLANTS 433

Fic. 1. 1. Galactose .025 mol. + saccharose .025 mol.
o “+ maltose .025 “
A “es Talhinose .025)-
s Hy, lactose. 025 °

+ arabinose .025

+ levulose .025
. Pfeffer’s solution. No sugar.

TABLE 2
Influence of Sugars on the Toxicity of Galactose

Average
Length of Length of | Length
Culture Solution Primary | [Lateral | of Top Class
Root Root (Cm.) of Injury
(Cm.) | (Cm.) |
Galactose .0125 mol. (2 cultures).......@:.....| I 05.5 | 7 5
Galactose .0125 mol.+glucose .025 mol. (1 cul- |
POLTE YPN eS a 9 Se vib) srs O
Galactose .0125 mol.+lactose .025 mol. (1 cul-'
EILUES pay Ghee ce aR ln ae a ea Ih 1.5 12 2
Galactose .0125 mol.+levulose .025 mol. (1 cul-
(CURE) as Gy retain CA en CR 3 6 m3 aes
Galactose .0125 mol.-+arabinose .025 mol. (1 cul-
(LEO) oy adh Re ect Glo A ane | ee 5 I 13 4
Galactose .0125 mol.+saccharose .025 mol. (1 cul-
‘TUNE Bie oe Succ alia dl te tA eae ra 2 8 15 1
Galactose .0125 mol.+lactose .025 mol. (2 cul-
MEE) Me ett n oe pie Mine see, Oe comet Shen, le 6 0.5 IO 4
Galactose .0125 mol.+lactose .025 mol. (2 cul-
TENGE S))) oe eRe ph Re ction tent renee Re I I 12 5
Galactose .o12 mol.+maltose .025 mol. (2 cul-
(CLONES) ORR IR Red AaB sie Hicker A eae cor GRE Aa erty I 0.5 12 5
Galactose .0125 mol.+raffinose .025 mol. (2 cul-
(LEU Gf So) igen rah oP eae Recah ieene® brine: Aeeentt to RAE I 0.5 10 5

A34 LEWIS KNUDSON

tration, or 0.025 mol. It was noted that 0.0125 mol. galactose is as
toxic as 0.025 mol. galactose. The results are given in Table 2.
Since there was variation in some of the series, the results of in-
dividual cultures are recorded in these cases; for plants that were
alike, the averages are recorded.

In general the results are similar to the preceding experiment,
though levulose is somewhat more effective than with the higher con-
centrations of galactose. Cultures are represented in Fig. 2.

Fic. 2. 9. Galactose .0125

+ saccharose .025
age " + raffinose .025
-+ glucose .025
+ levulose .025
ni te + lactose .025
15. Lactose .05

16. Pfeffer’s solution. No sugar.

Toxicity of Mannose.—In the course of certain experiments on the
use of various sugars by vetch (Vicia villosa) grown in water cultures,
it was noted that mannose, supplied at a concentration of 0.025 mol.,
killed the tips of roots that came into contact with the solution. Ex-
periments were then made with pea and wheat to determine whether
the effect was consistent, and agar cultures were used as in the ex-

TOXICITY OF GALACTOSE AND MANNOSE FOR GREEN PLANTS 435

periments previously described with galactose. It was found that
mannose at a concentration of 0.025 mol. behaved very much as did
galactose, with the possible exception that the browning of the roots
was not so intense as with galactose. The same general effect,
however, was noted as is evident in Figs. 3 and 4.

sicinate

Ge cn eae eee me

Ree oncom
cient ee

y
i
;
/
:
is
'
:
it

. Galactose .025 mol,
. Mannose .025 ‘

PIG. 33. <i
2
3. Mannose .025 ‘‘ -+ galactose .025 mol.
4
5

66

. Mannose .025 + glucose .025 “
. Pfefter’ssolution. No sugar.

Antagonistic Action.—Since glucose and saccharose had been found
to antidote effectively galactose, these two sugars were tested with
respect to their preventing the toxicity of mannose. Test-tube cul-
tures were used as for the galactose experiments, the amount of the
medium being 25 cc. Both peas and wheat were used. The cultures
were permitted to grow for two weeks, and the data were then re-
corded. All cultures were in duplicate, and those for pea, with the
exception of the cultures containing saccharose, were repeated.

The cultures employed were as follows:

436 LEWIS KNUDSON

. Galactose .025 mol.
6é .025 6c

I
2
3. Mannose .025 mol.
4
5

FIG, 4.

+ mannose .025 mol.

be

KO25h 4
") .025

+ glucose .025 mol.
+ saccharose .025 mol.

bc

Galactose .025 mol.

Mannose .025 mol.

Mannose .025 mol. + galactose .025 mol.
Mannose .025 mol. + glucose .025 mol.
Mannose .025 mol. + saccharose .025 mol.
Glucose .025 mol.

Saccharose .025 mol.

. Pfeffer’s solution alone—no sugar.

Both Path wheat and with peas it was found that the toxicity of
mannose is antidoted by either glucose or saccharose. Mutual an-
tagonism was not found between galactose and mannose. ‘The plants
grown in the other solutions were in every way normal. Representa-
tive cultures are shown in Figs. 3 and 4.

Discussion.—The hexose sugars glucose, mannose, and galactose
are stereoisomers. All of them are used by various fungi, and
mannose is as readily fermented by yeasts as is glucose. Galactose,

SOM DAE No

»

TOXICITY OF GALACTOSE AND MANNOSE FOR GREEN PLANTS 437

however, is fermented with greater difficulty, and it is suggested
(Armstrong, 1912) that perhaps a different mechanism is involved in
its fermentation. Mannose has a common enolic form with glucose
and fructose, and any one of the three may be converted into any other
under the influence of alkalies. It is therefore all the more surprising
to find mannose behaving similarly to galactose and not like glucose.

In a previous paper it was suggested that the toxicity of galactose
might be due to its oxidation products. The first oxidation products
of glucose and galactose are gluconic and galactonic acids. Various
cultures were made with Canada field pea in which the effect of calcium
galactate and calcium gluconate was to be noted. The experiments
were made as were those previously described. In no case was any
injurious action of calcium galactate noted.

It is not yet possible to offer any explanation neevuning for the
toxicity of the two sugars. An explanation of antagonism is suggested
by the phenomenon commonly observed with fungi, namely, the
election of organic substances, whereby if two organic substances are
offered only one may be absorbed. Various cases of this nature have
been reported even for stereoisomeric compounds. According to this
view, in a mixture of glucose and galactose, the toxicity of the latter
would be prevented because of the absorption of glucose and the
nonabsorption of galactose, and a similar condition would hold for a
mixture of saccharose and galactose. The failure of the other sugars
to antagonize the toxicity of galactose would be due to their inability
to prevent the absorption of the galactose. Work is being continued
on this subject.

REFERENCES

Armstrong, E. Franklin. The Simple Carbohydrates and the Glucosides, pp. 72-76.
TOL,

Knudson, Lewis. Toxicity of Galactose for Certain of the Higher Plants. Annals
of the Missouri Bot. Garden 2: 659-666. I915.

Influence of Certain Carbohydrates on Green Plants. Cornell Univ. Agr.

Expt. Sta. Mem. 9: 1-75. 1916.

Wilson, J. K. Calcium Hypochlorite as a Seed Sterilizer. Amer. Journ. Bot. 2:
420-427. I9QI15.

a4 ee eet As i Pi
* By (KGS pray : 4 é
re DARE ROS (i ty y \
A ag Nal fe
fees i
er NE Reis York Wee:
j, ira re Vas Y l
+" i) } A
i y ’ r

tis)

“ocToRER, 1917

OFFICIAL PUBLICATION OF THE. ie ie
BOTANICAL SOCIETY OF AMERICA ee

} ey aes eh ob it / $ }
lode . bay a od ~ 4 : i \
7 { ’ / ia ices “ayy \
by a ware Re ‘ ote \t
s = £ e ‘ i ¢ i ic
{ nd) FAS ‘ ¥ j sae \
\ } , oe ie: ey re 3 t S ; A
* ; @
; ae t : Ti { )
: an A 4 as Hb! {
: , Ff Wh AX ST ey 1
ds yee Meet ae \
—— EEE
“ ‘ or A = ‘5 ae . Sarah (
ERS \ y
/ { % Q& \
\
, }
rea c } >. i {
s . y : - i
:
( Ck 4 = } iv
a }
{ ye
: y A ‘
ot bes ,
; \ \ :
hi aye iN , >
+ ek a - lars ay
bs x : : . i :
\
, whale Seah ch le 5 }

“Taxonomic characters of the genera Alternaria and Macrosporium. As OS area Niet alee
Se: . Be Pe RC a NE eB UNG a Dees ELLIOTT. 439 ee

0

Hi

ona uso

CONES ye Waa | PUBLISHED ELS TAM Tar ec ate Gt Ae

“IN COOPERATION 1 With THE BOTANICAL sociery OF AMERICA
ue a HR inc May Oa AS BY THE. ES Oe oe ek) yi oe Saas AN
oie ie BROOKLYN BOTANIC. GARDEN oe (eas

Pe Ar at Norra Queen Serer, Lancaster, Pa

i iy: ;
‘| Z Ae sXe 4 pia: Lint k

“Entered as second-class matter February 21 1974, at the post etied at Lancaster, Pennsylvania,

Z

5 : vance sore act of March 3» eh ; ae
} Xa i } 52
a ¢ a a } » ae
Nu pte A y ) j ie «
PA o P
U cue iY aga § j ) i
¥ x c o eet . i t
i ¥Y ¢ % id
Mg Me Bh i %
bis tI q " “i .
he Bey ye i i Val) Ss ‘ SOeor alh A

_ DevoTep. Koy Me ue 'B

EDITORIAL COMMITTEE

~ oe. Ge ‘Newcomer, Editor-i -in- ‘Chief,
Ga eee ehh me University of RSS

ee wer se 5G STUART Casi, ediouse Manger :
: aye a A or ih Brooklyn Botanic Garden ray oe
ee ens a, hy “Irvine W, BAILEY, le ! Se
Dt ae ees te s eon Bussey Lastitution Se SRS Retin ee
| | ae AL BARTLETT, Me GE eee moe ie einem obi
yee _ Citinereiy of Michigan eare Be eh 1 Bene Pale :
Pee ae eS ee : Yee con Jounson, Untiersity ie Haneen tac
Ee ay a | Repterenting American Ehren acetal Lee)

ne Kcieed to all ae countries, eee Mexico: Cube, Porto. as ile
“ Canal'Zone; Republic of Panama, Hawaii, Philippine Islands, Guam, “Samoan ‘ ay
| Islands,. and Shan ghai.’ Postage to Canada 20 cents a volume on. annual subscrip ie
Poe ‘tions ;. to all other countries } in, Lae Postal Union, ess cents a 2 volume on annual stb-_ ;
ny “scriptions. Pe BAe Ria on fd, pe Cpe a Be a a8 a
ut | The pages of ie J sara ei open to members. of the , Botanical ;
A ORE “America, ¢ or een ee peeves for mom DES eee :
for Bbcoen) wee ao : Se ue a
Separates. should bs ordered ethan a brook is niet | “Fifgre copies ee i
will be Nee free; cover and additional copies at cost, : ne
a Cie of mailing. The Sobtlahers will suppl missing numbers free e only
es “have been sed in | the mails. : Pane

AMERICAN
JOURNAL OF BOTANY

VoL. IV OCTOBER, IQI7 No. 8

TAXONOMIC CHARACTERS OF THE GENERA ALTERNARIA
AND MACROSPORIUM

Joun A, ELLIOTT

LY aINPRODUCTION

Alternaria and Macrosporium are among the most universally
distributed and most common forms of the Fungi Imperfecti, em-
bracing, according to the ‘‘Sylloge Fungorum”’ of Saccardo, 41 species
and varieties of Alternaria and 189 species and varieties of Macro-
sporium, these numbers including some synonyms but not the new
species which have been described since the publication of the last
volume of Saccardo’s work. Some species, such as Alternaria solant
(E. & M.) J. & G., A brassicae var. nigrescens Peglion, and Macro-
sportum sarcinaeforme Cav. are well known and destructive parasites,
but the great majority are sapraphytes or have been described from
non-important hosts. The ascigerous stages of a few species are
known, the connection in all such cases being with the genus Pleospora.

Even a casual survey of the literature dealing with the genera in
question would reveal the fact that the generic names, Alternaria and
Macrosporium, are in many cases used synonymously in dealing with
the best known of the parasitic species. This condition could be due
either to there being no basis for distinction between the two genera,
or to this basis being ill defined. The studies of the writer were
undertaken with the hope of adding to the knowledge of these two
genera. The work was necessarily limited and the result is in no way
of the nature of a monograph.

I HISTORICAL

The genus Alternaria was described and figured by Nees (15),
A. tenuis being the type and only species described. The description

[The Journal for July (4: 375-438) was issued July 14, 1917]
439

440 JOHN vA] ELLIOTT

is incomplete and in some particulars inaccurate, but it is definite and
complete enough to leave little doubt that what Nees described was
what is now generally recognized as Alternaria.

Fries (7) described the genus Macrosporium, differentiating it
from Cladosporium, Helminthosporium, and Sporodesmium. The
muriform spore, now given as one of the characters of the genus, is
not mentioned in the generic description, otherwise it fits the present
current conception fairly well.. Having dropped the genus Alter-
naria, Fries makes no mention of it in his description of Macrosporium.

Macrosporium and Alternaria are placed by reason of their muri-
form spores in the section Dictyosporae of the family Dimidiaceae of
the order Moniliales, the muriform spores separating them from the
genera Cladosporium and Helminthosporium, which in some species
are in many particulars similar. Among the Dictyosporae there is
little basis, as the genera are described, for separating Stemphylium,
Septosporium, or Mystrosporium from Macrosporium. ‘The separ-
ation of the genera Alternaria and Macrosporium rests solely on the
catenulation of spores in the former genus. The fact that many of
the species of Alternaria now recognized were first described as Macro-
sporiums indicates the uncertainty of this basis for generic distinction.
In the specific descriptions in both genera, while mycelium, conidio-
phores, and spores may all be taken into consideration, spore characters
are the most used basis for distinction.

The question of the validity of the separation of the two genera
arose over the study of their ascigerous connection with Pleospora
herbarum Tul. The Tulasne brothers (19) figure P. herbarum bearing
both Alternaria and sarcinaeform spores on the same hyphae. Gibelli
and Griffini (8), Mattirolo (13), Bauk (1), and Kohl (12), studying
P. herbarum in pure culture, concluded that it should be divided into
two varieties or species, one having Alternaria conidia and the other
having sarcinaeform conidia. Miyake (14), studying the life history
of Macrosporium parasiticum Thum., found no Alternaria stage in the
life cycle. Halsted (9), in studying the life history of Pleospora
tropaeolt Hal. in pure culture, found that the cycle included only
Pleospora and Alternaria stages.

As the ascigerous stage of most species of Alternaria and Macro-
sporium is unknown or non-existent, the basis for the distinction of
genera and species must rest, in general, on the conidia. Jones (10-11),
in studying Macrosporium solani E. & M. and M. fasciculatum C. & E.

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 44I

on artificial media, found that they formed chains of conidia and
placed them in the genus Alternaria. Buioletti (3) reported securing
pure cultures of Macrosporium sp. and Alternaria sp. from olives in
California. Others have reported isolating species of one or the other
genera from various sources and growing them in pure culture under
conditions where the morphology of the fungi ought to have been
pretty certainly ascertained, but by far the greatest amount of liter-
ature on the two genera deals with their pathogenic effects or with the
mere description of species.

Miyake (14), Prillieux and Delacroix, (17), and others (4) have
shown by careful experiments or have suggested that many of the
specific names are synonyms. Constantin (4) and Planchon (16)
have reported great variations in Alternaria due to growth on different
media. Planchon (16) expresses the opinion that Macrosporium is
merely Alternaria with dissociated conidia. Noextensive comparative
work, either between the two genera or between species of the two
genera, has yet been undertaken.

Mi. METHODS: IN GENERAL

To aid in the comparison of published descriptions, a tabulation
of the species and varieties of Alternaria and Macrosporium given in
the “‘Sylloge Fungorum”’ was made on the basis of spore length, the
species being arranged according to the maximum length given.
Specimens from all available exsiccati were studied and compared.
Cultures of the principal types found in the exsiccati were studied
under varied conditions in order to learn something of the constancy
of the characters which are made the basis of specific distinctions.
The original generic descriptions were studied in order to ascertain
the basis of generic distinction.

THE VALUE OF WRITTEN DESCRIPTIONS

An examination of the specific descriptions of the two genera in
question showed that spore measurements were most constantly used
in distinguishing between species; in many cases several species were
alike in every character given except that of size of spores. In order
to learn something of the variation in measurements which may be
due to the personal element and to the use of different micrometers
and microscopes in measuring, three slides were prepared and together
with cultures of Phoma destructiva Pat. were sent to a number of persons

442 JOHN A. ELLIOTT

actively engaged in descriptive mycological work. On one slide of
Pleurosigma angulatum Sm., a single frustule was enclosed in a circle
and indicated; on a second slide two spores of Alternaria fasciculata
were similarly indicated. A third slide of A. fasciculata contained
several hundred spores. Identical typewritten directions were sent
to each one who made the measurements, asking that no more care
be used than: would ordinarily be taken in measuring for the purpose
of describing a new species. A tabulation of results follows:

TABLE A.

ote ae Sines es One Spore of | One Spore of | Many Spores of Many Spores or

5 Value | P. angulatum Alternaria Alternaria Alternaria Phoma
te, 2:4 /|:285-0 X45:0:| 48° *X7.2 | 10:2-< 9:6 5 12-8076 12 2-5 1-2
QE. 24 1284.4 X46.8 | 49.2Xx6 L9.2X0:61|) 94=36. x o-12 2-4 X28
Bilas 1.8. 284.4 X45.9. | 50:60>¢7.2-| 19:8 X0-0 15 *LigAtp6— 13
As 3.16 | 282 XK46.5 | 48 X7.5 | 19.5.X9.5. |10.5-24.X7-5-13] “5-02 bee
Bie 3.4. 282) X40. Ao 7, 20 £10 17-34 X 7-10
6.. 3.2) |. 280° = X45) | 48. 17 205 <7 13-20. x9-11 AS X2—2355
yaaa? 4. 280% AA 14S G7 20° XO 20X9 5-6 X 2-2.5
shore 3:2 4 276° = XA5 48 X8 21° X77 12-35 X9-I2 5-9 X2-3
© cers] LO. Slee Se RS OC Goats He eaten L3—25.<7o1f 5. eee
BOG ai Set 275) XA Wear 3 eZ 20° X10 £2-0 3) 6-12 32-82-43
II er 270° X48"! | 50 X75 |) 2T 6154) S540 6 Ie 5-8 X35
Variation: 24: 15.05<0 |, GOxK2 . i. 8o<3 65=17 x 353 B42

Variation: 9. |)" “5. cla" 4 72x25 8.5 X30 43-41 X 33-23 | 60-44 X 66-60

The second column gives the value of the smallest division of each
eyepiece micrometer in micromillimeters. All the measurements
given in the table are in microns. The arrangement is according to
the maximum measurement given for Pleurosigma angulatum.

The third, fourth and fifth columns show the variations in measure-
ments when all, without any doubt, were measuring the same things.
The variation is least in the longest measurements, being a little over
5 percent in the greatest length given, and greatest in one of the two
shortest measurements, amounting to 30 percent in the width of the
spore indicated in the fifth column. That this variation is not due
to the eyepiece used is shown by the fact that observer No. 10 who
returned the lowest measurement for the long spore, column 4, also
returned the highest measurement for the width of the spore given in
column 5. The variations given for the Phoma spores are the greatest,
being 66 percent for the shortest measurement and 44 percent for the
longest measurement. In this case each mycologist made his own
microscopic preparation.

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 443

The most instructive results appear in the sixth column of the
table where the maximum and minimum lengths and breadths of the
spores are given under conditions such as would obtain in describing
a new species. Here there was a variation of over 41 percent of the
highest maximum measurement returned for the length of the spores.
The variation for the minimum lengths was greater. Assuming vari-
ation equal to that shown in column 6 of Table A, 34 species of Macro-
sporium and 5 species of Alternaria in the ‘Sylloge Fungorum”’ are
inseparable by measurements of both length and breadth.

Applying to all the species of Alternaria and Macrosporium in the
“Sylloge Fungorum”’ the variability shown in column 6 of Table A,
they can be combined into thirteen groups, taking into consideration
the measurements for both length and breadth of spores. In other
words, if other characters are disregarded, in so far as actual spore
measurements are dependable, there are only thirteen species of Ma-
crosporium and Alternaria adequately described in the “Sylloge
Fungorum.” ?

STUDY OF EXSICCATI

Following the study of descriptions in the ‘‘Sylloge Fungorum,”’
examination was made of the specimens of Macrosporium and Alter-
naria in the exsiccati immediately available. One hundred and thirty-
four specimens labeled as 85 species, were found. Of this number 17
were marked “sp. n.”’ 2. e., of or nearly of the value of type material.
Eighteen other species not marked ‘‘sp. n.”’ were found in the exsiccati
of the authors or one of the joint authors of the species. This gave a
total of 35 species, the material of which can be regarded as reasonably
authentic.

Mounts were made from each of the specimens, from which spore
and conidiophore measurements, and the character of each, were
recorded. The nature of the growth, whether apparently parasitic or
saprophytic, was also recorded and the descriptions so made were
compared. There was no doubt that in the collection, many speci-
mens morphologically indistinguishable appeared under different
names, and that in some instances the same name was given to speci-
mens which were in no way similar or which could readily be dis-
tinguished from each other.

In the following summary of the study of the exsiccati the names
are given as they appeared on the specimen, followed by the title of

A44 JOHN A. ELLIOTT

the collection and the specimen number in the collection. Two
asterisks (**) following the species name indicates that the specimen
was marked ‘“‘sp. n.’’; a single asterisk (*) indicates that the specimen
was found in the exsiccati of the author of the species.

Group I

The following species had only globular or packet-shaped spores
and were essentially alike:
Macrosporium sarcinaeforme* Cavara, Fungi Par., Bri. & Cav., 116.
M. cladosporioides Desm., Fungi Sel. Ex., Roum., 5596.
M. stilbosporoideum Bri. & Cav., N. Amer. Fungi, Ellis, 2080.

Group 2

A second group was made of those having globular or packet-shaped
spores like those of the first group, but having in addition some ovate
or pointed spores which might be due to variation in the shape of spore
or to a mixture of two forms:

M. parasiticum** Thum., Myc. Univ. Thum., 667; Fungi Par., Bri. &

Cay, 152.

M. consortiale** Thum:, Myc. Univ. Thum., 1373:
M. sarcinula Berk., Fungi Columb., 3032.
M. chartarum Pk., N. Amer. Fungi, Ellis, 648; Fungi Sel. Ex., Rowan

6560.

M. heteronemum (Desm.) Sacc. Fungi Sel. Ex., Roum., 6647, 6562, 6358.

The following were possibly the same as the above but they either
showed minor differences or else the material was not sufficient to
afford positive judgment:

M. chartarum Pk., Fungi Columb., 396.

M. zimmermaneu Thum., Fungi Sel. Ex., Roum., 396.

M. polytrichum Cke. & Rav., Fungi Par., Bri. & Cav., 191.
M. puccinioides E. & And., N. Amer. Fungi, Ellis, 2876.

Group 3

A third group was made of species having long, narrow, regular,
tapering spores with few longitudinal septa. All were apparently
parasitic:

M. euphorbiae** Bart., Fungi Columb., 2633.
M. carotae* E. &. E., N. Amer. Fungi, Ellis, 3289; Fungi Columb., 2632.

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 445

M. amaranthi Pk., Fungi Columb., 2631.

M. brassicae f. solani Faut. et Brun., Fungi Sel. Ex., Roum., 6559.

M. chetrantht Fr., Fungi Amer., Rav., 303.

M. solam E. & M., Fungi Columb., 891; 398; 3635; Fungi. Par., Bri. &
Cay., 191; Economic: Fungi, Sey: & Earle, 339; 340; -N. Amer:
Fungi, Ellis, 1265; 2485.

Aliernaria solant (E. & M.) J. & G.,.Econ. Fungi, Sey. & Earle, 521.

A. brassicae (Berk.) Sacc., Fungi Par., Bri. & Cav., 87.

Two forms similar to the above were:

M. caudatum* C. & E., Fungi Columb., 397. This was like the above
but uniformly shorter.

M. antennaeforme B. & C., Fungi. Columb., 2531. This species was in
general form similar to those in the above group but the spores
were uniformly more slender.

Group 4

Three other specimens were found with spores of the same form as
the above group but the spores were much larger. These were:
M. herculeum* E. & M., N. Amer. Fungi, Ellis, 1263.
M. commune* Rabh., Fungi Europ., Rabh., 1360.
MM. saponariae Pk., N. Y. Fungi, Shear,-397; Fungi Sel. Ex., Roum.,
3868.

Group 5

Another group was formed of species with spores similar to those
of Alternaria solani but generally wider and shorter and always more
markedly muriform. The species in this group are not morphologically
identical but merely similar.

. heteroschemon** Faut., Fungi Sel. Ex., Roum., 6942.

. cucumerinum™ E. &. E., N. Amer. Fungi, Ellis, 3396.

. neruu Cke., N. Amer. Fungi, Ellis, 964.

. sucaviae Trabut., Fungi Sel. Ex., Roum., 4098.

. convallariae Fr.; Fungi Sel. Ex., Roum., 1897.

. brassicae Berk., Fungi Sel. Ex.; Roum., 6442.

I. martindale: E. &. M., N. Amer. Fungi, Ellis, 1262; Fungi Europ.,
Rabh., 2282:

. cherrantht Fr., Fungi Sel. Ex., Roum., 7235.

Alternaria malvae* Roum., Fungi. Sel. Ex., Roum., 3393.

A. brassicae (Berk.) Sacc., Econ. Fungi, Sey. & Earle, 515.

SSSS5585

-

446 JOHN A. ELLIOTT

Group 6

In the following group the specimens were essentially all alike;
quite variable in color, shape, and size. The color varied from light
to dark olive both in spores and conidiophores. Almost all appeared
to be growing saprophytically.

M. hibiscanum** Thum., Myc. Uni., Thum., 979.

baptistiae** Thum., Myc. Uni., Thum., 1271; Fungi Sele
Roum., 4897.
. casstaecolim** Thum., Myc. Uni., Thum., 1270; Fungi Sele
Roum., 4795.
gossypinum** Thum., Myc. Uni., Thum., 1469; Fungi Sess.
Roum., 4808.

raveneliu** Thum., Myc. Uni., Thum., 2071; Fungi Sel. Ex., Roum.,
4680.

rubi** Ellis, N. Amer. Fungi, Ellis, 544.

truncatum** Laub. & Faut., Fungi, Gallici, Roum., 6752. *
inquinans* C. & E., N. Amer. Fungi, Ellis, 369.

ornatissimum™ E. & B., Fungi Columb., 1741.

porr* C. & E., Fungi Columb., 1279; N. Amer. Fungi, Ellis, 370.
caudatum* C. & E., Fungi, Amer., Rav., 607; Fungi Columb., 890;

Fungi Columb.,397; N. Amer. Funet, Ells, 816:

aridis* C. & E., N. Amer. Fungi, Ellis, 51.

canificans* Thum., inid, Myc. Uni. Thum., 2280; Fungi, Gally exe
Roum., 4794. |

leguminum* Cke., Fungi Amer., Rav., 300; Fungi Amer., Rav., 603.
maydis* C. & E., N. Amer. Fungi, Ellis, 3098; Rabh.-Winter, Fungi
Europ., 3592.

catalpae* E. & M., N. Amer. Fungi Ellis, 1264; Econ. Fungi, Sey. &
Earle, 144.

martindaler* E. & M., N. Amer. Fungi, Ellis, 1262.

tomato Cke.,* N. Amer. Fungi, Ellis, 2484; Fungi Amer., Rav., 603.

. tenuissimum Fr., Myc. Uni., Thum., 980.

. convallariae Fr., Myc. Uni., Thum., 1965.

. chartarum Pk., Fungi Columb., 396.

M. floridanum Cke., Fungi Amer., Rav., 299.

M. florigenum Ell. & Dear., N. Amer. Fungi, Ellis, 3097.

M. togenariae Thum., Fungi Columb., 1367.

M. clematis Pk., Fungi Columb., 1830.

M. bulbotrichum Cke., Fungi Amer., Rav., 604.

SSESS § SE SS SSESES § S 8 &

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 447

. erumpens Cke., Fungi, Amer., Rav., 605.

. graminum Cke., Fungi Amer., Rav., 606.

. caespitulus Cke., Fungi Amer., Rav., 906.

. chetrantht Cke., Fungi Sel. Ex., Roum., 4490.

. consortiale Thum., Fungi Sel. Ex., Roum., 4992.

. saponiariae Pk., Fungi Sel. Ex., Roum., 3868.

M. phomoides Thum., Fungi Sel. Ex., Roum., 6145.

M. caespitulorum Rabh., Fungi Sel. Ex., Roum., 7236.

M. fasciculatum C. & E., Fungi Sel. Ex., Roum., 1058; N. Amer., Fungi,
Ellis, 523: Funer,-Columb:, 300; Myc. Uni, Thum., 1870:

M. commune Rabh., Fungi Sel. Ex., Roum., 2068, 4239, 3288, 6443;
Fungi Amer., Rav., 304; Fungi Columb., 2330; N. Amer. Fungi,
Ellis, 418. |

Alternarta fasciculata (C. & E.) J. & G., Econ. Fungi, Sey. & Earle,
522; Fungi Columb., 1368.

Except for being lighter colored, M. peponicolum** Rabh., Fungi
Europ., Rabh., 1285, was like those in the above group. Several
others were in most particulars like the above but either differed in
some respects or else the material was too scanty for judgment. These
were:

M. abruptum* C. & E., N. Amer. Fungi, Ellis, 127; Fungi Amer., Rav.,
202.

M. phaseolt** Faut., Fungi Sel. Ex., Roum., 6247.

M. cercosporoides* C. &. E., Fungi Columb., 1740.

M. valertanellae* Roum., Fungi Sel. Ex., Roum., 3690.

M. elegantissimum* Rabh., Fungi Europ., 2883; Fungi Sel. Ex., Roum.,
2067.

M. concinum B. & Br., Fungi Sel. Ex., Roum., 6443.

M. commune Rabh., Fungi. Sel. Ex., Roum., 4240.

M. puccinioides E. & And., Fungi Columb., 1172.

SSSS88

Group 7

Two quite similar species which differed from any other specimens
were:
M. junci** Lamb. & Faut., Fungi Sel. Ex., Roum., 6444.
M. brassicae Berk., Fungi Sel. Ex., Roum., 2363; N. Amer. Fungi, Ellis,
2483.
They were like Alternaria brassicae var. microspora, which the
latter undoubtedly was.

448 JOHN A. ELLIOTT

One species, A. cucurbitae** Let. & Roum., Fungi Sel. Ex., Roum.,
3694, did not afford enough material for judgment.

The above study of exsiccati and descriptions brings not only
species into question but genera as well, since in all but the first and
second of the above groups both Alternaria and Macrosporium are
included in groups as morphologically similar.

IV. (EXPERIMENTAL RESULAS

Cultures of Alternaria and Macrosporium and material upon which
either was growing were secured from many sources.! Of eighty cul-
tures thus obtained all but two produced chains of spores regularly
on artificial media and accordingly belonged in the genus Alternaria.
All of these had clavate, elongate or ovate, more or less pointed spores.
The two which did not ordinarily produce chains of spores had glob-
ular or sarcinaeform conidia. One of these very rarely produced
chains of two spores, in which cases the bottom spore was pointed.
Eleven of the cultures were selected as representative of all the forms
present and as most suitable for extensive study. These eleven cul-
tures also represented all of the morphological forms found in exsiccati.
They were: Alternaria solani (E. & M.) J. & G., isolated from blighted
potato leaves (Solanum tuberosum L.); A. solani isolated from Datura
leaf spot (Datura stramonium L.); A. brassicae var. nigrescens Peglion,
isolated from blighted cantaloupe leaves (Cucumis melo L.); A. bras-

EXPLANATION OF GRAPHS 1 TO 9

The measurements of spores are indicated in microns by the base line, each
space representing one uw. The frequency is indicated on the perpendicular lines,
each space representing one spore. Measurements are at intervals of 2.4 uw except
in the case of the narrower spores where the width is taken at intervals of 1.2 u.

The following letters are used to indicate the host and media: A, natural host;
B, bean agar, 30°; C, bean agar, 10°; D, bean agar, + 20 Fuller’s scale; E, bean agar,
— 20 Fuller’s scale; F, synthetic agar; G, synthetic agar minus glucose; H, synthetic
agar with double amount of glucose; J, synthetic agar with double amount of as-
paragin; J, synthetic agar without asparagin; K, leached agar.

The most striking facts to be observed are: general reduction in size of spores on
the synthetic agars (F to J), over that on beanagars (Bto E). Relative constancy
of species given in graphs 2, 4, 5, and 6 over those given in graphs 3, 7, 8, and 9.
Increase in size of spores at lower temperature (Graphs 7 and 9; C). Extreme range
of variation in size of larger spores (graphs I, 2, 3, 8 and 9).

1 Cultures or material was received from J. J. Davis, L. R. Jones, B. F. Lutman,
W.. G. Sackett, S. M, Tracy, H. 1: Gissow, ©. ©: Stakman;-G, P. Chnten sir
Cook, G. F. Atkinson, B. D. Halsted, I. M. Lewis, C. W. Edgerton, C. R. Orton,
B. B. Higgins, J. W. Eastham, Miss Jean MacInnes, G. L. Peltier, F. C. Stewarts

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 449

3 r*) 79 () y U. OF 1.5.5. FORM ®

450 JOHN A. ELLIOTT

sicae var. microspora (Berk.), Sacc., isolated from cabbage leaf spot
(Brassica oleracea L.); A. fasciculata (C. & E.) J. & G., isolated from
potato leaf spot; A. tenuis Nees, isolated from decaying wood; A.
diantht Stev. & Hall, isolated from Dianthus leaf spot; A. iridicola
(E. & E.) n. com. (Macrosporium tiridicolum E. & E.), isolated from
iris leaf spot (Iris germanica L.); A. sonchi Davis, isolated from leaf
spot of Sonchus asper (L.) Hill; Macrosporium sarcinaeforme Cavara,
isolated from clover leaf spot (Trifolium pratense L.); M. sarcinula
Berk., isolated from decayed spot on Jonathan apple (Pyrus malus L.).

All studies were made with pure cultures originating from a single
spore of each species used, the spores being located on thin poured
plates and transferred before germinating to other plates. The in-
struments invented by Keitt! greatly facilitated this operation.

Whenever it was possible to do so, ten-day-old cultures were used
as the standard for comparisons. Measurements of spores and conidi-
ophores were made whenever present. In measuring spores a me-
chanical stage was used and every spore on the slide which came within
the field was measured for both length and width, thus eliminating
unconscious selection. Curves of the measurements of each hundred
spores were made for each culture. Since in a given species there is
frequent variation in the length of the beaks, from one third to six or
seven times the length of the spores, the HES were not in any case
included in the measurements.

Graphs giving the superimposed curves of the measurements of
one hundred spores of the same species under different conditions
were prepared for the sake of easy comparison of the variations due to
differences in cultural conditions (Graphs I-09).

Host RELATION AND PARASITISM

Inoculations were made on the natural hosts, on their near relatives,
and on other plants, when there seemed to be any advantage in so
doing. For example, Alternaria solani, which is morphologically
similar to A. brassicae, was inoculated on cabbage leaves. _

The plants were inoculated by needle pricks and by placing spores
in drops of water on the unbroken leaf surface. The percentage of
successful inoculations was estimated from the needle pricks, since
these were more easily located. An inoculation was considered suc-
cessful if the fungus to a notable degree invaded and killed the tissues

1 Phytopathology 5: 266, 1915.

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 45I

U. OF 1.8.8. FORM?

45

JOHN A. ELLIOTT

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 453

454 JOHN A. ELLIOTT

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 455

surrounding the punctures. Unless otherwise stated, successful inoc-
ulations also occurred on the unbroken leaf surface. To check the
results of the needle-prick inoculations, sterile needle pricks were
always made and all the plants kept under identical conditions. In
no case did the control plants show signs of fungous invasion around
the needle pricks. They will receive no further mention. Unless
otherwise stated, spores of the fungus used were always re-isolated
from the spots caused by the inoculations. In the following discussion
the terms ‘‘normal’’ and “characteristic’’ indicate that the spot or
fungus appeared the same as in the cases of natural infection.

Alternaria solani from potato leaves: Inoculations were made from
the first and tenth generations on agar and from spores taken directly
from potato leaf spots. The mycelium was used as the inoculum,
since very few spores were produced in culture. All of the inoculations
with mycelium were made by needle pricks on plants under bell jars.
The plants inoculated were: potato (Solanum tuberosum); Datura
stramonium; tomato (Lycopersicum esculentum |L.| Mill.) leaves and
fruit; Solanum nigrum L.; cabbage (Brassicae oleracea L.); iris (Its
germanica L.); Lactuca canadensis L.

All inoculations on solanaceous hosts were successful; about 50
percent of the inoculations on cabbage leaves failed; none succeeded
on iris or Lactuca. The development of the spots was less rapid on
Datura than on any other of the solanaceous hosts. Spots were all
of the characteristic ‘‘target board’’ type. Very few spores were
produced on any of the spots—none on the cabbage leaf.

A. solani from Datura leaves: Inoculations by needle pricks and
on the uninjured leaf surface were made with spores taken directly from
leaf spots and with spores of the first and tenth generations from agar
plates. All plants were kept under bell jars. The same plants were
used as in inoculations with the strain from potato, and the results
were similar except that all attempted inoculations on cabbage failed.
The fungus developed least vigorously on potato leaves. On the other
solanaceous hosts the spots were not different from those caused by
A. solani from potato. Very few spores were produced on any of the
spots.

A. brassicae var. nigrescens from cantaloupe leaves: Inoculations
were made with small pieces of diseased tissue, also with spores from
the first, sixth and tenth generations on agar, by means of needle pricks
and on the uninjured leaf surface. The plants were kept under bell

456 JOHN A. ELLIOTT

jars for the first few days following inoculations. Inoculations were
attempted on cantaloupe (Cucumis melo), cabbage, Datura, and Lac-
tuca. Characteristic spots were produced on cantaloupe in all cases;
25 percent of the inoculations on cabbage were successful; no spots
were formed on Lactuca or Datura.

The spots produced on cantaloupe leaves, appearing in 2 or 3 days,
developed rapidly. The spots on cabbage leaves were mere dots
where the spores had germinated on the unbroken leaf surface, but of
considerable size where punctures were made. No spores were pro-
duced on cabbage and very few on cantaloupe leaves.

A. brassicae var. microspora from cabbage leaf spots: All inocula-
tions were made with spores from pure cultures, by needle pricks or on
the unbroken leaf surface. Inoculations were attempted on cabbage,
radish (Raphanus sativus L.), Lobularia maritima (L.) Desv., Dianthus,
potato, Datura, tomato leaves and fruit. On all cruciferous plants the
fungus produced characteristic concentric spots with dark rings of
spores on the surface. Spores spread the fungus to other parts of
susceptible plants until they were entirely destroyed. On ripe to-
matoes dark spots were formed with narrow sharply defined zones of
spores. No spots were formed on the other plants used.

A. fasciculata from potato leaf spots: Inoculations were made with
spores by means of needle pricks and on the unbroken leaf surface.
The plants inoculated were: potato, tomato leaves and fruit, cabbage,
radish, Dianthus.

The only definite spots of parasitic appearance were from needle
pricks on etiolated or partially etiolated cabbage leaves. ‘The spores
and conidiophores on the cabbage leaf spots were very light amber in
color, instead of dark olive as is normal. In ripe tomatoes a hard
black core was formed in the interior as the mycelium invaded the
tissues. ‘This was surrounded by a soft decayed area. The inocu-
lations failed on the other plants.

A. tenuis from decaying wood: Only one series of inoculations was
attempted with this species. This was made on Dianthus leaves by.
means of needle pricks. No spots were formed. The inoculations
were made on Dianthus because A. fenuis spores cannot be distin-
guished from those of A. diantht.

A. dianthi from Dianthus leaf spots: Inoculations were made both
by needle pricks and on unwounded leaf surface with spores of the
first and fifth generations from agar plates. Dianthus and cabbage

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 457

were the only plants used. No spots were formed on cabbage leaves.
All inoculations on Dianthus were successful. In air of natural
humidity the spots remained very small but under a bell jar they
spread rapidly, produced concentric rings of dark spores on the surface
and finally killed the plants.

A. tridicola from iris leaf spots: Inoculations were made entirely
‘by means of needle pricks. Both spores and mycelium were used’
The plants inoculated were: Iris, Liliwm philadelphicum L., Datura,
potato, cabbage, Lactuca, cantaloupe, and onion (Allium cepa L.).
All inoculations except those on iris and onion failed. Spots on iris
extended slowly, on onion tops more rapidly, accompanied by the pro-
duction of a considerable number of spores. Inoculations were made
on Datura, potato, cabbage, and cantaloupe because of the general
similarity of the spores of A. tridicola and A. solant (Plate XX, Fig. 7).

A. soncht from leaf spots on Sonchus asper: The fungus could not
be isolated from the spots on Sonchus asper since the spores would not
germinate in culture.. Crude inoculations were made on Lactuca
canadensis, using as the inoculum diseased leaf fragments of S. asper.
These inoculations were successful and the production of spores on
L. canadensis abundant. Pure cultures were obtained from L.
canadensis leaf spots. Inoculations were made. by needle pricks and
on the unbroken leaf surface. Inoculations were made on Lactuca
canadensis, L. sativa L., Taraxacum officinale Weber, cabbage, tomato,
and Datura. Spots were formed only on the Compositae. On L.
canadensis very rapidly spreading characteristic dark brown spots were
formed in two days. On cultivated lettuce and dandelion, distinct
spots were formed around needle-prick inoculations, but soon ceased
to grow and no spores were produced. Spore production was abundant
on L. canadensis. (Plate XX, Fig. 8).

Macrosporium sarcinaeforme from leaf spots on red clover: Inocu-
lations were made with spores, both on the unwounded leaf surface
and by needle pricks. Red clover (Trifolium pratense), white clover
(T. repens L.), alfalfa (Medicago sativa L.), cow pea (Vigna sinensis
Endl.), Cucumis melo, and Allium, were inoculated. The fungus was
actively parasitic only on the clovers and alfalfa. Spots developed
most rapidly on red clover and spores scattered the disease over the
entire plant. On white clover and alfalfa the spots did not spread to
leaves that had not been inoculated. In all cases the spots appeared
within three days of inoculation. On onion, small white spots were

458 JOHN A. ELLIOTT

occasionally produced around needle pricks but no spores were pro-
duced and the parasitism was evidently very feeble.

M. sarcinula from apple fruit spot: This fungus was isolated by
Miss Jean MacInnes, of the University of Illinois, in her studies of the
rots of apples.? It appeared to be causing an apple rot. Small black
pycnidia-like bodies were scattered frequently over the rotted area
under the epidermis of the apple. These also appeared in the first
generation on agar. When broken they appeared to be merely
sclerotia. Spores from pure cultures were used in making inoculations
both by needle pricks and on uninjured leaf surface. The fungus
conformed to descriptions and figures of M. sarcinula. The same
plants were inoculated with M. sarcinula as with M. sarcinaeforme.
Spots were produced only on red clover and alfalfa. All attempts to
inoculate apples failed. The spots on red clover were lighter colored
but otherwise almost identical with those caused by WM. sarciniae-
forme. Few spores were produced on the spots. On the onion tops
the fungus produced small white spots surrounding the punctures but
no spores were formed.

The most noteworthy results of the inoculation experiments oc-
curred in respect to A. solant, A. sonch1, A. brassicae var. nigrescens,
and M. sarcinula. ‘Both strains of A. solani grew freely on all of the
solanaceous hosts, and one strain was feebly parasitic on cabbage.
Morphologically, judging from exsiccati, A. solani and A. brassicae
are identical, and the inoculation experiments might be considered
further evidence of this view. A. sonchi was actively parasitic on
‘Lactuca, a near relative of Sonchus, and was possibly very feebly
parasitic on other Compositae. A. brassicae var. nigrescens proved
slightly parasitic on cabbage although this would hardly be expected
from nearness of Cruciferae and Cucurbitaceae. M. sarcinaeforme
and M. sarcinula are entirely dissimilar except in general form of
spores, yet they appeared equally parasitic on red clover and alfalfa,
and both were feebly parasitic on onion tops.

CULTURES ON AGAR

For purposes of isolation and general study, standard lima bean
agar (beans 100 g., agar I5 g., water 1,000 g.) was used. ‘The cultures
were kept in darkness at 30° C., with the exception that A. soncit,
which would not grow at 30°, was grown at 20°. In testing the effect

3 Unpublished thesis.

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 459

of a lower temperature, cultures on standard bean agar were kept at
10° C. The effect of acidity, also alkalinity, was determined on stand-
ard bean agar, using 30 cc. and 20 cc. of normal hydrochloric acid,
or 20 cc. of normal sodium carbonate per liter. A standard synthetic
agar was used to study the effect of nutrition: 1.36 g. acid potassium
phosphate, 1.06 g. sodium carbonate, .5 g. magnesium sulphate, 5 g.
glucose, I g. asparagin, 15 g. agar, 1000 g. water. Variations from
this standard were made by omitting glucose, doubling the amount
of glucose, omitting asparagin, and by doubling the amount of as-
paragin. Plain agar, 15 g. per liter of water, washed for severai days
in distilled water, was also used.

Records of all cultures on the various media were made in tabular
form, but for the sake of brevity only the table for cultures on bean
agar is given here; the differences shown on other media being briefly
summarized. Variations in size of spores are given on graphs one to
nine.

The most striking characters brought out by the colonies on the
standard bean agar were the wide differences in the two strains of
A. solani which on their hosts are indistinguishable. The strain from
potato produced a pure white colony with marked red chromogenesis
in the medium, had straight colorless submerged mycelium and no
spores. On account of the abundant production of conidiophores and
spores the strain of A. solani from Datura formed a gray colony, it
produced no chromogenesis, and the submerged mycelium was dark
olive and torulose. A. fasciculata, A. tenuis, and A. dianthi produced
spores indistinguishable from each other. The conidiophores of
A. diantht were slightly larger in cross section than those of the other
two. In general appearance of the colonies and in the production of
aerial mycelium these three species were different. The other Alter-
naria species studied were quite distinct in most cultural characters.
The two species of Macrosporium were totally unlike except in general
form of spores.

Bean agar, 10°, 30 days: Zonation was absent or inconspicuous in
most of the colonies, especially in those of A. brassicae var. microspora
which at 30° had the most marked zonation of all the species. This
species also showed a marked alteration in color and form of its spores,
these being light amber, nearly colorless, and about half the width of
the normal dark olive spores. Another marked change occurred in the
size and color of the spores of A. dianthi, these being twice their normal

JOHN A. ELLIOTT

‘[euliou
‘Qaro ‘yuepunqy

*jewl1ou ‘aAT[O

ylep ‘yuepunqy|‘ysep

‘jew1ou uUeYyy
Joj[ews ‘sauoz
Terjao9Aw

“aAT[O F431]
‘fewsou ‘yuep
-unqe A[irey

‘[eur
-Iou ‘juepunqy
‘Jeurtou ‘jue

-punqe = <Ayjieyq

*[eulIoU
‘yuepunqge AoA

"[ew10U

‘yuepunqe Aro/\

*[eul1ou
‘AAl[O Yep ‘M7

*poqorsys
-uoo yonw nq
‘azis Ul [eur
-iou ‘yuepunqy

‘Teusou ‘SATO

WS ‘uepunqy| ‘3ysre1ys ‘ssap1ojoD

*jeuriou

‘jewitou ‘aAlo aed
Japun sdurnyj9o uy|A10A

‘sduinjo uy]
“JAT[O ajed A1aA
‘uepunqe <Ayjsey

“IAT[O
ys ‘Juepunqy

“JAT[O
1y431]‘Juepunqe JON

*JAT[O
sy “Juepunqy

‘yewsou ‘aATjoO
“WysI] ‘Juepunqy

‘[eulsou
‘QAlJO Yaep ‘May

*poqorys

-uod yonut JO
jewsou ‘juepunqy

dUON

so10ydorpruod

dso] N4to}

pue peoyoueiq

‘yuepunqy|yonw ‘aAlo yredq

"14 31e14s ‘ssaj1ojod

*Au0}}09 ‘a7TYM
‘suoy = =‘quepunqy

“yeom ‘ayy M
‘yzoys ‘ayy = AraA

*£U03}09 ‘dT M
‘yuepunqe Ala/\

*AUO}

"Wy Bre1ys ‘ssajiojop|-309 ‘ayYyM ‘su07

"74 31e14s
‘DATO F431]

“O7TyM

AI9A 10 ‘sSa[IO[OD| ‘suo ‘e[qesapisuo7

*7Yy31e14s

‘QAT[O suluin}

‘\sIy 3e ssa[IO[OD

"14.31e19s
‘938 YUM dIAT[O
suluiny

"yy S8re1ys ‘os

Ajieau 10 Ssaf1o[o7

"yy s1e14s
‘QAT[O JIC

“ISO[NIO}

pue payoueilq
yonut ‘aAljo y4r1ed

"yysrerjs ‘ssajtojo)

uinyaoAY] pediawiqns

‘SSd][1O[O7)

‘Suo]

‘Buo] “OUYM “O]IWT]

“Suo]
‘yuepunqy

“yeom ‘ouIyM
‘yroys ‘ayqqIy ATOA

‘o7TyM

ayy ay3H] AOA

“oq
‘quepunqy

*Au0}}09 ‘d}TYM

suo, ‘Juepunqy

uintasA I [Vay

‘aqiym ‘AU0}}03)

‘AUO[OO
29S Uo p “org
“SIsoUasOWO1YS
MoTPPA 1Y SITS
‘guoz [erpaoAul
asuap yqtM Auojoo
AW0}}09 ‘d}1YM 9INg

peuver
[[f worzeuoz ‘Au0}
-309 Avis 10 JY

*AUO[OO pauoz ‘Aeir)
spend
-op [JAM UOI}eUOZ
‘938pa 1ajno uo
ajed ‘19399 UI yAeq

‘pouoz opr ‘AeIn
‘uOI}eUOZ pauyep
Ajdreys MoJ1eu
yywm Auojoo yor[g

“194
-U99 dy} Ieau AT[e
-190dsa ‘yep Ala/A

“sISoUa SOUL
-o1y9 ou ‘Auoloo

pouoz apim ‘Airy
‘pol
youd pat0joo

uInI pau ‘pauoz
oprim ‘aUYM ing

Joyovieyy)
[elsuaxy) uol}eUu0Z

CI-OI

cI—-Oo!I

7,0 ‘Sz

Z1-S1

gI-SI

welts

Ass

OL—-9z

cv

ce-ce

ce

“WA

‘yIMoIy

DINULIADS * I

IUdofaDUut9
-4DS WNntd0gSodID

1YIUOS “PY

Djooupian “yy

AYE

sinua] *“y

DIDINIISDL * PV

ps404GSOdI
“UL * A ‘apIIsspléqg * VY

SUIISIOAB

-1u *A ‘apisspdg ‘VP

ein}
-eq WOl] tunj0s “Pp

0}e}0d wod]
WUDIOS DIADULIY PY

snsung

"gq ATAVL

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 461

dimensions and of much lighter color (Plate XIX, 10). A marked
change in the character of both the aerial and submerged mycelium
occurred in M. sarcinaeforme, which is normally black, but here was
white. All colonies were much retarded in rate of growth, the amount
varying with the species. Five produced no spores.

Bean agar + 30 Fuller's scale, 30°, 10 days: This medium did not
solidify and in many cases the whole colony was saturated with it.
There was little variation in the general appearance of most of the
colonies from that on standard bean agar. Two variations worthy of
note were the absence of zonation in colonies of A. brassicae var.
microspora, and the more than usual production of aerial mycelium on
colonies of A. brassicae var. nigrescens. The chromogenesis of A.
solani from potato was more marked on this than on any other medium.
Some species which usually produced spores in more or less abundance
produced few spores or none on this medium. They were A. brassicae
var. nigrescens, M. sarcinaeforme and M. sarcinula. The spores of
A. soncht and M. sarcinula were below normal size.

Bean agar + 20 Fuller’s scale, 30°, ro days: Except for a little
greater production of aerial mycelium by most species, the general
appearance of the colonies was like those on neutral bean agar. Chro-
mogenesis of A. solani was very marked, also the submerged mycelium
of this species was darker than normal. A. brassicae var. nigrescens,
M. sarcinaeforme, and M. sarcinula produced fewer spores than o
neutral agar. |

Bean agar — 20 Fuller’s scale, 30°, to days: On the alkaline bean
agar the most notable feature was the abundant production of spores
by A. brassicae var. nigrescens, which ordinarily produced few spores
on artificial media or on its natural host. In this instance the spores
covered all parts of the colony. A. solani from potato, which on
neutral and acid media produced marked chromogenesis, on the al-
kaline medium produced none, and the submerged mycelium, in the
older part of the colonies, became dark. A. diantht and M. sarcinula
produced spores smaller than normal on the alkaline medium. The
general appearance of the colonies was little different from that on
neutral bean agar with the exceptions already noted.

Standard synthetic medium, 30°, 10 days: Except for a general lack
of aerial mycelium and the appearance of less luxuriance, the colonies
were little different from those on the standard lima bean agar.

Standard synthetic agar minus glucose, 30°, to days: The appearance

462 JOHN A. ELLIOTT

of the colonies on this medium was very similar to that on washed
plain agar, although the appearance of starvation was not so em-
phasized.

Standard synthetic agar with double amount of glucose, 30°, Io days:
The aerial parts of the colonies, mycelium, conidiophores and spores,
were more abundant than on the standard medium. The size of the
spores of some species was much reduced (Graphs 4, 8, 9; Curve H).

Standard synthetic agar with double amount of asparagin, 30°, 10 days:
All colonies appeared less luxuriant than on the standard synthetic
agar or bean agar.

Standard synthetic agar minus asparagin, 30°, Io days: Partial.
starvation appeared in all species, and the submerged mycelium was
darkened in most cases. The latter was most noticeable in A. solani
from potato which usually had colorless mycelium. This same fungus
produced a few spores on this medium, which was also unusual.
Aerial mycelium was generally very meager in all species.

Plain washed agar, 20°: But for the mycelium of Macrosporium
sarcinaeforme, the only characters by which the different species could
be identified were those of the spores and conidiophores which were
natural -(Plate XIX., 1, 3,5, 7,/ 8,11, 16, 17; XX:°7, 0) terra
mycelium was lacking and the submerged mycelium colorless and
without distinctive features. Spores and conidiophores were charac-
teristic. This was the only artificial medium upon which A. solani
from potato produced spores in any abundance. A. solani, from Da-
tura, did not produce spores until after eighty days. The torulose
mycelium characteristic of this strain on most media was scarcely
noticeable on the washed agar. Macrosporium sarcinaeforme pro-
duced dark torulose submerged mycelium characteristic of the species.

VARIATIONS UNDER DIFFERENT CONDITIONS OF HUMIDITY

The effect of humidity upon the colony characters was studied by
keeping the cultures at 20° C. in saturated atmosphere; room humidity
(variation 45-70 percent); and in an atmosphere kept dry by the use
of calcium chloride. In order to prevent desiccation of the medium,
all the plates were coated with a thin layer of paraffin. After the
agar had hardened, hot sterile paraffin was poured over the agar and
the excess poured off. This left a thin film of paraffin effectively
protecting the medium from desiccation or contamination. The
paraffin was punctured in several places by needle pricks and spores

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 463

of the fungi planted in one of the punctures. The cultures were kept
with the plates inverted, without covers. In no case was there any
contamination after the plates had been coated and very little occurred
during the process. In the least humid atmosphere there appeared
to be no change in the water content of the medium even after sixty
days.

In all cases where aerial mycelium was produced, there was an
increase in its production in the saturated atmosphere over that under
ordinary conditions. Those species producing the greatest abundance
of mycelium produced proportionately fewer spores. On account of
the presence of the paraffin film, the usual colony form was lacking.
A. brassicae var. microspora and M. sarcinaeforme, both of which pro-
duced almost no aerial mycelium, produced an abundance of spores.
At room humidity little mycelium was produced, and that was in
tufts around the needle pricks. The relative production of spores was
greater than in the saturated atmosphere; the submerged mycelium
grew rapidly under the paraffin and covered the plate. In some cases
the conidiophores ruptured the paraffin coat. This was true of M.
sarcinaeforme to a greater extent than of any other species. In the
very dry atmosphere no aerial mycelium and no spores were produced
with the exception that M. sarcinaeforme produced a few spores in air
bubbles in the paraffin coating.

REACTION TO BACTERIAL INFLUENCE

One species of bacterium which will be designated as B. x, occasion-
ally occurred as a contamination on plates. It had a marked in-
hibitory action upon the growth of the fungi and was therefore used
for special study. Plantings of all the species of fungi that were being
studied were made in the usual way, then after four days the plates
were inoculated on four sides at about I mm. from the outer edge of
the colonies with cultures of B. coli, 2 strains, B. subtulis, and B. x.

In all cases the fungi grew through the colonies of B. colt without
showing any reaction. Occasionally in colonies of B. subtilis there
was slight darkening of the mycelium. On the sides of the fungous
colonies opposed to the colonies of B. x, however, there was in every
case a marked reaction by the submerged mycelium. The hyphae were
inhibited greatly in growth, were much darkened and extremely
torulose. Nodules were produced along the hyphal threads and at the
extremities of hyphal branches. The reaction by all the fungi was

464 JOHN A. ELLIOTT

essentially the same. In all cases there was no resemblance between
the normal and the abnormal mycelium (Plate XX: figs. 1-6). The
aerial parts of the colonies did not show reaction to the bacterial
influence.

CHROMOGENSIS ON RICE

On account of the chromogenesis exhibited by two of the species,
cultures were made on boiled rice (1 g. rice, 10 g. water autoclaved at
120° C. in test tubes). The colonies were allowed to grow for two
months before comparisons were made, although it was noted that
A. solani from potato, A. sonchi and M. sarcinula showed more or
less marked chromogenesis within three or four days after inoculation.

The most marked variations were those between the two strains
of A. solani, which throughout the other experiments had shown many
differences in reaction to media. A. solani from potato colored the
rice light orange-red at the bottom of the test tube, deep orange-red
or brick-red at the lower limit of mycelial growth. The part of the
rice through which the mycelium grew was dark-red-brown. The
aerial mycelium was also red-brown. A. solani from Datura did not
color the medium below the mycelium. The part of the medium occu-
pied by the colony was dark gray or black. The aerial mycelium was
white. A soncht colored the medium lemon-yellow; A. irdicola,
yellow-brown; A. fasciculata, A. tenuis, and A. dianthi light brown;
A. brassicae var. nigrescens gave a very faint brown; A. brassicae var.
micros pora gave no coloration. The medium was colored deep pink
by Macrosporium sarcinaeforme; blue-gray by M. sarcinula.

The remarkable thing about the reactions of these fungi to various
media was that it was impossible to predict from the reaction of one
species, what the reaction of another would be. A case in point was
the reaction to alkaline and acid media. A. brassicae var. nigrescens
showed a gradual increase in production of spores from high acid to
alkaline media, the production in the latter case being very abundant.
A. solani, from Datura, produced very few spores on the alkaline
medium while on the acid media the spore production was abundant.
Several species showed a lowered spore production in the acid media,
notably those which normally produced few spores. Macrosporium
sarcinaeforme, which usually produced spores in great abundance,
produced none on the higher acid medium.

Lack of nutritive substances affected all species alike, causing a

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 465

general lack of luxuriance, lack of color in submerged mycelium and
absence of aerial mycelium. This reached the extreme on the leached
agar, where the aerial mycelium was almost entirely lacking and the
submerged mycelium colorless and not at all characteristic, with the
exception of WM. sarcinaeforme.

The comparative studies noted above were made with cultures
which were ten days old, but all of the cultures were maintained thirty
days or longer—sometimes ninety days. In some cases marked
variations due to age were noted in the spores. In several species
secondary growth or development of spores occurred which entirely
destroyed their distinctive characters. This secondary development
consisted in a multiplication in the number as well as a great enlarge-
ment of the cells. This caused the spores to become very irregular in
form, much darker, deeply constricted, and very much larger than
normal (Plate XIX: 2, 6, 10).

As a result of the secondary development, the spores of Alternaria
assumed the form of Mystrosporium, as was pointed out by Constan-
tin (4). Corda (6) in his description of the genus Mystrosporium
made figures which are undoubtedly of Alternaria spores in an advanced
stage of secondary growth. Later (6) he figured the development of
M. stemphylium Corda from the regular form of a young Alternaria
spore through the various stages of secondary development to the
very irregular form called Mystrosporium. From the descriptions of
species of Mystrosporium there is no doubt that most if not all of them
belong to Alternaria. All specimens labeled Mystrosporium which I
examined in the exsiccati were Alternaria in various stages of secondary
enlargement.

Young colonies of all species of Alternaria produced smooth regular
spores which remained regular for a time interval which varied in
different species and on different media. The richer the medium the
earlier the secondary enlargement began. The first indications of
this secondary growth were multiplication of and deep constriction at
the septa. On leached agar no secondary enlargement occurred in
any species, even after a lapse of three or four months. There occurred
irregularly in all species a well-marked echinulation of the spores.
Beyond a deep constriction at the septa, no secondary change was
noted in Macrosporium, nor in A. sonchi, A. brassicae var. micros pora,
and A. solani from potato. This last, however, produced spores in
numbers sufficient for observation only on leached agar where secon-
dary enlargement did not occur in any species.

466 JOHN A. ELLIOTT

V. DISCUSSION

A general survey of the characters upon which specific descriptions
of Alternaria and Macrosporium are, or may be, based is fitting in
this place.

Broadly considered, the different bases for descriptions may be
divided into natural morphology, cultural morphology and media
changes, and host relations. In the present studies effort was made
to get a fair estimate of the value of each of these by studying them
under various conditions. In specific descriptions the shape, size,
and color of spores are generally given. Also septation, constriction,
and echinulation are more or less frequently mentioned. In many
instances one of these characters may furnish the only point of differ-
entiation of one species from another species in the description of
which that character is not mentioned. The conidiophores are gener-
ally more or less definitely described, the characters commonly given
being, length, width, septation, constriction, color, branching and
geniculation. As in the case of the spore characters, any one or all of
these characters may be neglected by one author or made the basis for
specific description by another. The nature of the submerged my-
celium has been made the basis for specific descriptions in some cases,
although in many cases it is not mentioned. In artificial cultures
where the mycelium is easily studied it is generally taken into con-
sideration. The characters noted are thickness, length between sep-
tations, color, constriction, fasciculation and rate of growth. Aerial
mycelium is not often mentioned in specific descriptions. Length,
width, color, abundance, and general habit are the characters con-
sidered.

MORPHOLOGY

1. Spore: The previously described experiments show that, within
the natural range, the size of the spores under natural conditions was, ~
in almost all of the species studied, quite constant, although under
ordinary cultural conditions the size of the spore on artificial media is
usually less than that on the natural host. However, in a relatively
short time—two to six weeks—in some species there is a marked
secondary development.

The most characteristic thing about the spores of these fungi is
their shape which is an immediate index to the genera, and generally
to the species. All obclavate, cuneate, ovate, pointed or beaked spores
belong to Alternaria and under suitable conditions form chains. ‘There

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 467

is variation in shape within a single species, but in most cases under
natural conditions the shape of the spore, combined with its size,
identifies the species. Parasites, as they are collected during the
period of active attack upon plants, are uniform in spore form. Some
species wintering on decaying leaves showed a secondary development
which quite destroyed their characteristic form and size and made
recognition impossible. On artificial media where secondary growth
occurred frequently, the shape of the spores was greatly changed.

Ail Alternaria spores have a more or less acute apex and some species
have beaks. In some cases the beaks are six or seven times the length
of the spore itself. This, however, occurs only when they are not
borne in chains, or in the terminal one of a chain. In the chains the
terminal spore is formed by budding out of the apex, or occasionally
from the side, so that the spore above takes the place of the beak of
the spore below. For this reason the length of beak alone cannot be
taken as of great value. Many species of Alternaria do not have
beaked spores, so to a certain extent the beak is a specific character.

There is marked difference between species in the number of septa.
Septation is dependent on the size of the spore in a single species, but
independent of the size of spore in different species. The amount of
septation increases with the age of the spore. The same conditions
which brought about changes in shape and size of spores, increased
the number of septa and constriction at the septa.

Echinulation occurred at times in all of the species studied. In
some cases part of the spores from a colony were echinulate while others
were smooth.

When the age of the spore is considered, color is an important
character of some species. Generally spores darken with age. Alter-
naria diantht and A. brassicae var. microspora grown at 10° showed
marked color change, both being very light amber instead of the
normal dark olive. A. fasciculata, forming spots on partially etiolated
cabbage leaves and also in growing on raw disinfected cabbage leaves
and petioles, showed a similar change.

2. Conidiophore: The form of the conidiophore is often quite
characteristic, although length depends upon the age and environment.
In Macrosporium sarcinaeforme the end of the conidiophore becomes
swollen before the spore appears. If this spore falls off another bud
forms at the end of the conidiophore, producing another ‘“‘joint”’
which in turn swells at the end and produces a spore. If the spore

468 JOHN A. ELLIOTT

does not fall off, the conidiophore may bud again at the side of the
swollen end and produce a second short “‘joint’’ which bears a second
spore. This process, repeated, causes the geniculation which is some-
times given as a specific character. Branching occurred occasionally
in all the species but was not often observed under natural conditions.
The width of the conidiophores was quite constant. Constriction
differs between species and with the age of the conidiophore. In
some species no constriction was noted while in others the constriction
was marked.

Color is as distinctive of the conidiophores as of the spores and is
usually concolorous with that of the spores. The only variations in
the color of conidiophores on artificial media corresponded with the
color change of spores.

3. Aerial mycelium: Parasitic species under usual conditions pro-
duce little or no aerial mycelium. Under humid conditions, a very
little aerial mycelium was sometimes produced by some species. ‘This
was always lax and white, and although in some species it was thicker
in cross section than in others, it was so nearly featureless as to be
thought worthless in classification. More or less aerial mycelium
was produced by all species under some conditions of artificial culture.
In some species, notably A. brassicae var. microspora and M. sarcinae-
forme, it could be found only with the microscope. Other species
produced aerial mycelium in abundance under humid conditions on
rich media.

4. Submerged mycelium: The characters of the submerged myce-
lium were the most variable of all characters studied. Under ordinary
cultural conditions on standard media the mycelium of some species
had distinguishing characteristics, but any change in the cultural con-
ditions might bring about a change in the characters of the mycelium
which would make its identity doubtful. On low-grade nutrient agar
the mycelium of most species was colorless and, with a few exceptions,
characterless. On some media the color, width, and general character
might separate certain species from certain other species, but under
the influence of strong inhibition, such as was brought about by bac-
terial influence, the changes entirely destroyed any distinguishing
characters. In most cases age greatly affected the mycelium in color,
septation, and constriction.

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 469

Host RELATION

Inoculations upon host plants show that some species are quite
narrow in their range; e. g., A. soncht was parasitic on Sonchus asper
and Lactuca canadensis but failed to maintain itself on cultivated
lettuce and other Compositae. Others may be broader in their range,
growing actively on many species of one family and less actively on
other families; as, A. solant which grew actively on all solanaceous
hosts used and less actively on cabbage. A. brassicae var. microspora
grew actively on all cruciferous hosts inoculated but failed in all inocu-
lations outside of that family. Forms morphologically inseparable
may grow on different hosts and fail to cross-inoculate, or of several
forms morphologically alike only one strain may be parasitic. On the
other hand, forms which are morphologically clearly distinct may be
equally parasitic on the same host.

The studies here presented would seem to show that the present
tendency toward limiting the distinction of species to differences in
morphology, regardless of host relationships, is commendable. They
would seem to indicate, further, that characters should be sought which
are least affected by cultural conditions, and in using such characters
due consideration should be taken of factors liable to make those
characters variable; such as, age or stages in development.

VI. GENERIC LIMITATIONS OF ALTERNARIA AND MACROSPORIUM
A, ALTERNARIA

In every case where species of Alternaria have been described, the
spores have been more or less elongated and pointed at one end. In
the present studies all spores of the obclavate, cuneate, or ovate form
produced chains of spores under favorable conditions. Since the
species studied in culture covered all the types of spores found in
exsiccati or described in the “Sylloge Fungorum”’ under the genera
Alternaria and Macrosporium, there is no doubt that all of the species
with these types of spores belong to the genus Alternaria, that most of
the species named under Macrosporium belong to this genus, and
can be recognized as such by the descriptions given of their spores.

The term “‘clavate,’’ as applied to the Alternaria spores, is a
misnomer arising from a misconception and carrying misconception
with it. Without exception the pointed end of the spore is the apex
and the rounded end is the base. Alternaria tenuis has been figured
with its spores attached by their apices, and as this error was widely

470 JOHN A. ELLIOTT

copied it undoubtedly went far toward spreading the misconception.
Other species of Alternaria, described as Macrosporium, have been
similarly figured and it is a common conception that the beak is a
pedicel. Many authors have made correct descriptions and figures
Alternaria. In such cases the term ‘‘obclavate”’ replaces ‘‘clavate.”’

The generic description of Alternariat should be emended as follows:

ALTERNARIA Nees. Conidiophores solitary or fasciculate, erect or
subdecumbent, simple or branched, generally short, colored. Conidia
muriform, often with few longitudinal septa, ovate, obclavate, or
elongate, always with more or less definitely pointed: apex, often

long-beaked, colored; under favorable conditions forming chains.
(Ex., A. tenuis, the type of the genus.)

B. MACROSPORIUM

Of the four species of Macrosporium described by Fries when he
created the genus, M. convallariae and M. cheiranthi undoubtedly be-
longed to Alternaria. Fries, however, had rejected the genus Alter-
naria and placed A. tenuis in the genus Torula as 7. alternata. In the
exsiccati studied, specimens labeled as M. convallariae and M. cheiran-
tht belonged to Alternaria. The third species of Fries, M. tenuissimum
(Helminthosporium tenuissimum Nees), is placed by Saccardo in the
genus Clasterosporium as C. tenuissimum. The fourth species of
Fries’s original publication, M. caracinum, is not mentioned by Saccardo.
The material for this species was supplied to Fries (7) by Schweinitz,
and the description of Fries corresponds exactly with the description
and figures of Schweinitz (18) for Clasterosporium caricinum, which is
the type of the genus Clasterosporium. There is no doubt, therefore,
that the types of Clasterosporium caricinum and Macrosporium carici-
num are from the same material.

Corda (6) placed Clasterosporium caricinum in the genus Sporades-
mium as S. closteriosporium, but the genus Sporodesmium was at that
time so heterogeneous that such a disposition cannot be considered
final.

Since Macrosporium caricinum Fries and M. tenuissimum (Nees)
Fries were published two years before Clasterosporium caricinum Schw.,

4 ALTERNARIA Nees, Syst. d. Pilze 11, p. 72 (Etym. alternus ob conidia
alterne crassiora et tenuata). Hyphae fasciculatae, erectiusculae, subsimplices,
breves. Conidia clavato-lageniformia septato-muriformia per isthmos (conidiorum
caudas) catenulatim digesta, mox vero secedentia.—Polydesmi, Macrospori et

Clasterospor1i species nonnullae obiter observatae forte huc spectant. (Saccardo,
Sylloge Fungorum IV: p. 545.)

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 47I

M. tenuissimum (Nees) Fries becomes the type of the genus Macro-
sporium, which will take over the species now given under’ the genus
Clasterosporium, Clasterosporium being antedated by Macrosporium.

All but twenty of the species of Macrosporium listed in the ‘Syl-
loge Fungorum”’ belong to Alternaria. These twenty are given as
having globular and sarcinaeform, or globular and clavate spores.
The latter species may possibly have both forms of spores but it is
much more probable that in most cases the “‘clavate’’ spores belong to
Alternaria and that two distinct forms were present when the species
were described. Such mixtures were found in the exsiccati in several
instances. Of the two species studied in culture M. sarcinula very
rarely produced a pointed spore bearing another spore and might be
considered a bridging species to Alternaria.

All of the twenty species given in the “Sylloge Fungorum”’ having
sarcinaeform or globose spores are one species in so far as spore meas-
urements are dependable. Of the thirteen specimens found in ex-
siccati, however, there were two distinct forms, and two distinct forms
were cultured during the present studies. Some of the names of the
twenty species in question have already been reduced to synonymy
and it is probable that only a few separable forms exist.

There is nothing in the morphology of the species of this group
which would exclude them from the genus Stemphylium.

The species now named Macrosporium which should be put into
Stemphylium, providing they are distinct species, are: M. sarcinula
Berk, M. sarcinaeforme Cav., M. abruptum C. & E., M. baccatum Ell.
& Kell., M. valerianellae Roum., M. elegantissimum Rabh., M. pelar-
gonu E. & E., M. globuliferum Vestergr., M. toruloides E. & E.,
M. nodipes Sacc., M. schemnitziense Baumel., M. puccinioides E. &
And., M. chartarum Pk., M. myrmecophilum (Fr.) Sacc., M. nitens
(Fr.) Sacc., M. subglosum Cke. & Rav., M. rosarium Penz., M. septo-
sporum Rabh., M. atrichum C. & E., M. parasiticum Thum., M.
commune Rabh.

Vi SPECIFIC LIMITATIONS “OF ALTERNARIA

Except for the species named above as belonging to Stemphylium,
all the species of Macrosporium in the ‘“‘Sylloge Fungorum”’ belong to
Alternaria, and can be recognized as Alternaria from their descriptions.

In the present studies enough variation due to age alone occurred
in Alternaria tenuis, A. fasciculata, and A. diantht, to fit these to a

472 JOHN A. ELLIOTT

large percentage of the species of Alternaria and. Macrosporium de-
scribed in the ‘‘Sylloge Fungorum”’ and found in the exsiccati. No
variations occurred in the forms in group 6, of the exsiccati material
studied except those which would occur due to age, and all in this
group might be considered A. tenuis and varieties of A. tenuis. Even
the narrowest limitation in the application of specific names would
preclude the retention of many of the present species.

No final disposition of the present specific names of Alternaria and
Macrosporium can be made without a study of authentic specimens
of each species. Most of the descriptions are far from being complete
or definite enought to permit their being used for this purpose. For
convenience a tentative grouping of similar forms which may be
identical, and which are undoubtedly closely allied, should be made.
These groups might well be retained to indicate the similarity of a
number of forms such as is exemplified in bacteriology in the B. cols
and B. subtulis groups. As in bacteriology, each group should be
designated by a typical species.

The groups suggested are as follows:

The A. tenuis group. This group is characterized by spores ranging
from 11-50 X 7-20. The spores are quite variable in form as well
as in size but are generally broad and muriform (fig. I, spores of the
A. tenuis group). All specimens mentioned in group 6, page 446,
belong here. :

The A. brassicae group. This group should contain regular, long,
tapering, acute-beaked spore forms with measurements ranging from
35-120 X 10-30 uw. The spores have few longitudinal septa and are
often long beaked (fig. 2, spores of the A. brassicae group). All of
group 3, page 444, belong in this group.

The A. herculea group. A. herculea (E. & M.) com. nov. (Macro-
sporium herculeum E. & M.) is the type of a group with spores similar
in form to those in the A. brassicae group but much larger (fig. 3, spores
of the A. herculea group). To this belong the specimens mentioned
in group 4, page 445.

The A. cucumerina group. A. cucumerina (E. & E.) com. nov.
(Macrosporium cucumerinum E. &. E.), syn. A. brassicae var. nigrescens
Peglion, is typical of a group similar in spore form to that of A. bras-
sicae but with the spores uniformly wider and more muriform and,
generally, shorter (fig. 4, spores of the A. cucumerina group). In this
group should be placed all in group 5, page 445.

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 473

The A. soncht group. A. soncht Davis is different from any of the
species in the above groups in having large-celled spores with a mark-
edly obtuse apex (fig. 5, spores of the A. sonchi group). The range in
size 1s 50-125 X 12-25 u. No similar form was found in exsiccati.

/oom /00M DOW Nickie OY [22 eencee aa
© > YP (sec
o> Cro
1 2 :
JOoo”mr 00 Joom :

5

TEXT-FIGS.I-—6.—For explanation see text.

The A. brassicae var. microspora group. The spores of A. brassicae
var. microspora (Berk.) Sacc. differ from those of A. tenuis in being
uniformly narrower and less muriform; longitudinal septa seldom being
formed (fig. 6, spores of the A. brassicae var. microspora group). Here
belong those of group 7, page 447.

A more complete study of exsiccati might show the necessity of
other groups for forms not included above.

Without doubt the number of valid species of Alternaria is a very
small percentage of the present named species of Alternaria and Ma-
crosporium, but it is outside of the limits of the present paper to reduce
these names to synonymy.

VIII. SUMMARY.

I. Spore shape in the genus Alternaria is correlated with catenula-
tion of spores.

2. All obclavate, ovate, cuneate, or elongate-pointed spores of the
Macrosporium-Alternaria type form chains and belong to Alternaria.

3. The acute end of the spore is the apex or beak, not the base or
pedicel.

A74 JOHN A. ELLIOTT

4. Of these spores, all globular, sarcinaeform, cubed, or oblong
spores without apex or beak belong to Stemphylium.

5. Macrosporium, by priority, becomes the name of the genus which
has been known as Clasterosporium Schw.

6. Of the four species described by Fries when he created the genus
Macrosporium, two belonged to Alternaria and M. tenuissimum
(Nees) Fries becomes the type of the genus Macrosporium.

7. In these genera, conidiophores and conidia possess the only
suitable characters upon which to describe species. .

8. Conidia in many species go through secondary changes which
destroy their distinctive characters.

9g. Secondary changes in conidia may be due to age or to abnormal
environment.

10. The genus Mystrosporium was described from Alternaria in
advanced secondary development.

11. Mycelial characters are too easily affected by external condi-
tions to be dependable in describing species.

12. Slight changes in media may cause great changes in the sub-
merged mycelium.

13. Identical changes of environment may bring about opposite
reactions in different species.

14. Morphology of the dependable stable characters under normal
conditions is the most useful basis for describing species.

15. Due consideration must always be taken of the age of the
organism in describing a species.

16. For convenience, the genus Alternaria should be divided into
groups of species having similar spores.

The writer takes this opportunity to express his thanks to Pro-
fessor F. L. Stevens for encouragement and many helpful suggestions
in carrying out this work; to Professor Wm. Trelease for aid in solv-
ing the taxonomic problems involved; and to others, mentioned or
not mentioned in this paper, who have given assistance in any way.

UNIVERSITY OF ILLINOIS, URBANA.

LITERATURE CITED

I. Bauk, H. Zur Entwickelungsgeschichte der Ascomyceten. (Pleospora her-
barum.) Bot. Zeit. 1877: 313-326.
. Berkeley, M. J. Notices of British Fungi. Ann. Nat. Hist. 1: 261. 1838.
Bioletti, F. T. Some Diseases of Olives. Calif. Agr. Exp. Sta. Rep. 1895: 243.
4. Constantin, M. J. Sur les Variations des Alternaria et des Cladosporium.
Rev. Gen. Bot. 1: 453-466. 1889.

WwW N

TAXONOMIC CHARACTERS OF ALTERNARIA AND MACROSPORIUM 475

mcorda, A.C. I)“ Icon, Pune..1; 12); Figit75;..2:13:, DabiX> . Fig. 61. 1836:

. Corda, A.C. I. Icon. Fung. 4: 4. 1836.

. Fries, E. Syst. Myc. 3: 373-375. 1819.

. Gibelli, G., e Griffini, L. Sul polimmorfismo della Pleospora herbarum Tul., in
Archivio Triennale del laboratorio di botanica crittogamica in Pavia 1: 53-
102, tav. V-IX. 1874.

g. Halsted, B. D. N. J. Agr. Exp. Sta. Rep. 13: 290. 1892.

10. Jones, L. R. Studies on Macrosporium solani. Vt. Agr. Exp. Sta. Rep. 9:
79. 1895. .

11. Jones, L. R. On Alternaria solani. Vt. Agr. Exp. Sta. Rep. 10: 45. 18096.

12. Kohl, F. G. Uber den Polymorphiamus von Pleospora herbarum. Bot.
Centralbl. 18: 23. 1883.

13. Mattirolo, O. Sub polimorfismo della Pleospora herbarum Tul. e sul valore
specifico della P. sarcinula e della P. alternariae di Gibelli e Griffini. Mal-
pighia 2: 357. 1888.

14. Miyake,G. Life History of Macrosporium parasiticum. Ann. Bot.3:1. 1889.

15. Nees, C. G. Syst. Pilze 2: 72. 1817.

16. Planchon, L. Influence de Divers Milieux Chimiques sur Quelques Cham-
pignons du Groupe des Dematiees. Ann. Sci. Nat. II. Bot. 8: 1-248. 1Igoo.

17. Prillieux et Delacroix. Bul. Soc. Myc. France 9: 201. 1893.

18. Schweinitz, L.D. Syn. Amer. Fungi No. 2998. Trans. Amer. Phil. Soc. n. ser.
42300: Fig. 4,-1832.

19. Tulasne,L.R.etC. Selecta Fungorum Carpologia, 2: 261. Plate 32: figs. 1-14.

Plate 33: figs. 11-14.

CON AU

EXPLANATION OF PLATES XIX AND XX
PLATE XIX

All drawings were made with camera lucida to scale shown on the plates.

1. Alternaria brassicae var. nigrescens on leached agar.

2. Aliernaria brassicae var. nigrescens spores showing stages in secondary
development.

3. A. solant from Datura on leached sugar.

4. A. solani from Datura, spores from bean agar culture.

5. A. solani from potato on leached agar.

6, A tenuis spores two weeks and six weeks old showing secondary development
due to age.

7. A. tenuis on leached agar.

8. A. diantht on leached agar.

g. A. diantht spores from colony on acid agar.

10. A. diantht spores from colony on bean agar at 10°, showing enlarged size
and secondary development.

11. A. fasciculata on leached agar.

12. A. fasciculata spores showing echinulation.

13. A. fasciculata showing development of a spore.

14. A. brassicae var. microspora on leached agar.

476 JOHN A. ELLIOTT

15. A. brassicae var. microspora on bean agar, 10°, showing reduced size of
spores and lack of color.

16. Macrosporium sarcinaeforme on leached agar.

17. M. sarcinula on leached agar.

PLATE XX

1. Mycelium of A. solani from potato; a, reaction to bacterial influence; 0,

2. Mycelium of A. brassicae var. microspora; a and b as in No. I.
3. Mycelium of A. brassicae var. nigrescens; a and 6 as in No. I.
4. Mycelium of A. solani from Datura; a and 0 as in No. I.
5. Mycelium of A. tenuis; a and b as in No. I.
6. Colony of A. brassicae var. microspora showing reaction to Colony of B. x.

7. A. tridicola on leached agar.

8. A. sonchi conidiophores and spores on Lactuca canadensia.

9g. A. sonchi conidiophores and spores on leached agar.

10. Spores of A. solani from Datura; A, from Datura leaf spot; B, from potato
leaf spot; C, from tomato leaf spot (artificial inoculations).

VOLUME IV, PLATE XIX.

AMERICAN JOURNAL OF BOTANY.

oe
Mise ey)

ELLIOTT: ALTERNARIA AND MACROSPORIUM.

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE XX.

Se

Od, “Pp

Si
he

[)

BIS

Pd

ELLIottT: ALTERNARIA AND MACROSPORIUM.

; ic F a PF es ty it P I ~ Ss
- ® * “wo 2
Ls iy i 7 :
' ; > : —— |
j - i x 7 .
7 | : , _ x Ms i . : ; :
7 * -“ he J 4 . ’
3 4 Naor ; ;
. 7 ¥ . -- : a
> H -
we : ‘ = . { - ‘
: , 5
: : 7 \
bad r ‘ > - -
= 3 : . “ : = 7 is 8 4
/ = : j = ; : : ; ‘
7 : 7 ‘ = id Md ‘ .
i ba) 7 i . . ~
we . 7 .§ i - . .
: *

r . ‘ , iy ®
< so | A 2
a : : Pa
é omy Os
; : o 7 a G 7 : :
; - : _ 7
A 7 . . 3 P 7
- , a -
m : 3 ~ . - : 7 . ole
. ‘I os -
‘ 2 e ‘ 7
A ; y A ; é :
Z . - . : ; } ’ “
. ’ = . i] - - = ye
: " . oe :
ad . fl a .
R , : ; Z z
: ~ o — ms . a =
: - : ; a ' ; tk a
Z Zz Fe hed . .
a ‘ p a : : - 7
: : . = - = : . ao > hi
: ec ; H
~ - ' ‘ : , ‘
y _ . ;. - 7 4 - ¢ 4 a :
j j ; , c ;
: i> ; 7 a a . , a

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES

J. G. GROSSENBACHER

INTRODUCTION

The term ‘“‘crown-rot’”’ is used to designate a bark disease of fruit
trees (chiefly apple—Pyrus malus and pear—Pyrus communis), occur-
ring in most of the tree-growing portions of the United States. The
affected bark eventually dies in various sized patches, and the sur-
rounding living tissues produce callus, which thus separates the living
tissues from the dead. This disease most characteristically affects
the bark of the lower trunk and that of the adjoining portions of the
upper roots. The location of affected patches of bark seems to depend
very largely upon the interrelation of growth and weather conditions;
in some cases the disease involves chiefly the upper roots, while in
other cases it occurs most frequently at the base of the trunk. An
affected patch of bark that dies to the wood decays more or less
rapidly, depending upon its distance from the ground or other sources
of moisture. The wood exposed by the decayed bark is usually dis-
colored at its surface but may be alive within and active in the con-
duction of water. When the crown-rotted patch extends around
three fourths or more of the trunk, the downward current of elaborated
food in the bark is interfered with to such an extent as to permit much
less than the normal amount of radial growth in the roots. The
enfeebled roots thus absorb less soil solution, and therefore smaller
leaves are formed. The wood under the wound dies in time, and
thus the water-conducting tissues are reduced. Such trees usually
die in a few years unless radial growth produces much new wood about
the wound in the meantime. In some instances the bark dies entirely
around the base of the trunk, and in many cases of this sort the width
of the dead girdle determines the length of time such ‘a tree will live.

Crown-rot has an important economic bearing upon the fruit
industry of this country, owing to the fact that it involves the lives
of trees and is therefore much more serious than fruit- and leaf-spotting
diseases which after all are essentially matters of a season.

Crown-rot and related bark diseases have been investigated inter-

477

478 J. G. GROSSENBACHER

mittently during many years; in the early days of phytopathology
considerable attention was given to these diseases and much valuable
information was accumulated. The subsequent enormous develop-
ment of mycology, in its relation to the diseases of plants in general,
has eclipsed and supplanted the interest formerly centered on bark
diseases, apparently because the mycological phases afforded more
tangible subjects for investigation. Most of the more modern at-
tempts at the study of crown-rot have been of a preliminary nature
and have led only to hazy or ill-founded conclusions.

Some of the apparently new ideas that occurred to me during
the earlier part of: this investigation were published in 1909,! while
others were stored away, embodied in the form of notes and photo-
graphs to be used later. Continued search of the older literature of
botany and forestry for observations upon bark diseases, as well as
with reference to the question of radial growth,? resulted in gradually
placing one after another of my supposed new ideas in the category
of confirmatory observations and conclusions.

The Literature-—The literature accumulated on crown-rot and
related subjects during the past six years has become very voluminous,
and to attempt a review seems rather discouraging. Most of the
important papers on radial growth, and on certain of the factors
determining its distribution, were reviewed some time ago in the
last cited paper. Many of the more general papers on this disease
and some of those dealing with the cause of the trouble were discussed
in my two former papers on crown-rot. There are still too many
abstracts of such papers on hand to be fully utilized in this connection,
and therefore only a few of the most pertinent ones will be mentioned
_later in the discussion of my results.

The Causes of Crown-Rot.—The common orchard bark-fungi are
evidently the causes of the rotting of the bark in crown-rotted trees,
but the cause of the initial injuries that led to the death of this bark
has not been experimentally determined, although some work upon
this problem has been done on citrus trees in Florida. Field observa-
tions in the north, together with a few experiments, have shown that
the manner and timeliness of radial growth, pruning, and the occur-

1Crown-rot, Arsenical Poisoning and Winter Injury, N. Y. Agr. Exp. Sta.
Tech: Bull. 12: 369.. 1900;

2 The Periodicity and Distribution of Radial Growth in Trees and Their Rela-
tion to the Development of ‘‘Annual”’ Rings, Trans. Wisc. Acad. Sci. Arts and Letters
TO) lew 1EO1 5;

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 479

rence of droughts or low temperatures are closely related to the
development of crown-rot there; and similar investigations in Florida
are revealing the prevalence of comparable adverse environmental
conditions, preceding the occurrence of foot-rot on citrus trees. In
the latter case it was also noticed that the leaf-surface area of a tree
exposed to sufficient light for active photosynthesis, when compared
to the surface area of bark on the branchless portion of the trunk, is
often rather small in trees which appear especially subject to this
disease; that is, it is suggested that a scarcity of elaborated food
coursing downward in the bark of the trunk below the branches may
also have a causal relation to the occurrence of crown-rot.

Methods of Work.—It was found possible to cut usable sections
from material of small diameter embedded in celloidin without the
previous use of softening agents. The series used in this study was
obtained chiefly in the manner described and from material collected
and fixed in spring and summer of I912. ‘The citrus material was
collected later in Florida and similarly treated.

Flemming’s strong solution and Carnoy’s mixture were used almost
exclusively for fixation. The former gave more satisfactory material
for staining, but the latter was more largely used because of its more
rapid penetration. Flemming’s triple stain and Heidenhain’s iron-
alum haematoxylin stain were most frequently used. The triple
stain was found to yield much quicker and more satisfactory results
when used after a mordant such as iron-alum, but for use in making
photographs iron-alum haematoxylin proved more desirable than
Flemming’s. The same was later found to be the case in sections
from foot-rot material of citrus trees.

The Early Stages of the Disease.—The first visible stages of crown-
rot consist of discolored and often ruptured tissues variously dis-
tributed in streaks and patches in the bark. In cross-section the
injured patches are often arranged more or less concentrically about
the wood cylinder, although they are usually most severe on one side
of the stem. In the mildest forms of the disease the medullary rays
of the inner phloem and groups of parenchyma cells about the
sclerenchyma strands and inner cortex are affected, although at times
only the one or the other of these tissues is involved. In more severe
cases much of the phloem and practically all the cambium may
be injured.

The severity and course of the disease following these evident

480 J. G. GROSSENBACHER

beginnings depends mainly upon the relative extent and number of
the injured or dead patches, upon the weather of the ensuing growing
season, and possibly upon the relative abundance of wound fungi.
In many of the milder cases, the injured and collapsed tissues are
merely more or less compressed by the subsequent growth of the
surrounding live parenchyma cells, and in late summer only the
presence of irregular formless dead masses among the living tissues
of the bark tells the story of the former trouble. In the more severe
cases, however, in which in addition to the medullary rays of the
phloem, the inner portions of the cortex and perhaps most of the
cork cambium have been much injured, the results are likely to be
more serious. In these cases, as in the milder ones, the resumption
of growth by the surrounding live parenchyma results in the com-
pression of the dead and dying tissues; but since the dead patches are
numerous, relatively large, and close together, the intervening live
parenchyma and ray-cells are insufficiently supplied with water and
nutrients and therefore cannot survive the drying weather of late
spring and early summer. During the latter part of this process a
new cork cambium is developed inside the dying cortical parenchyma,
resulting afterward in a rough, scaly bark. In cases where the
initial injury involves very large patches of outer phloem but leavet
the inner phloem and practically all the cambium intact, the resuls
is approximately the same, excepting that occasionally small patches
of bark die to the wood on account of the occurrence of coincident
injured patches in the cambium and inner phloem. It often happens
in instances of this kind that the cortex is affected but slightly and
that it retains its normal appearance until the internal trouble has
become far advanced; then it usually dies rather quickly and dries
out. However, none of the types of injury so far described usually
result in very serious trouble because at most only small areas of
bark are killed to the wood.

When most of the cambium and much of the phloem are initially
affected, the injurious results are usually much more evident; but
even in such cases the bark may survive if the weather is favorable
and if the area affected is not extensive as compared with the total
area of the bark of that portion of the stem. In case the injured
patches in the cambium and inner phloem are relatively large or fairly
close together, or if they form nearly continuous sheaths of affected
tissues, the regenerations from the living portions of the bark are

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 481

hampered by defective water conduction outward from the wood;
and unless the weather conditions are most favorable, so that the cells
proliferating from the bark are able to re-establish living connections
by fusing with the wood cylinder, these proliferating cells die, and
the result is the death of the entire bark. Although-such patches of
dead bark are produced in various ways and result from injuries of
varying degree, we term the wounds crown-rot if they occur on the
bases of trunks or on roots near the ground, and canker if they occur
on stems and branches above ground.

This histological investigation permits some inferences to be drawn
that support most of the important contentions advanced in my
former papers; as regards other contentions, however, the evidence
is not so convincing. My preparations, as will appear later, indicate
that both excessive tensions and certain degrees of immaturity of bark
tissues have a causal relation to the development of the initial injuries
that give rise to this bark disease. They also substantiate the results
obtained by the cultural tests; no fungi are usually in evidence until
the middle of May or even later.

THE DEVELOPMENT OF THE DISEASE

The first visible stage of crown-rot, as well as that of some other
bark diseases, consists in a discoloration and collapse, or even in a
rupture, of groups of tissues mainly of the inner bark. This stage is
usually found only in late winter and spring, and is generally not
evident to the ordinary observer unless the outer bark is conspicuously
cleft. From late spring to mid-summer, however, most of the severe
cases attract attention by the oozing of ‘“‘sap”’ or gum and by the
eventual discoloration of the outer bark. Such affected bark is most
commonly found on the trunk near the ground, in crotches, and at
the bases of small young branches arising from the large limbs of
heavily pruned trees. When at or near the ground, dead bark rots
quickly; above ground it usually dries and eventually scales off from
the wood.

The initially affected tissues are variously distributed in streaks
and patches, which in cross section usually appear in more or less
nearly concentric circles about the wood cylinder. In cases of slight
injury, the medullary rays of the inner phloem, groups of parenchyma
around the sclerenchyma strands or patches of cortical parenchyma
are affected. In more severe cases, much of the phloem and all the

482 J. G. GROSSENBACHER

cambium may be involved or the phloem and cambium injuries may
be accompanied by injuries in the cortex. The severity and course
of the disease following such initial injuries depend upon the size and
number of the affected patches, and upon the location of the most
severely affected portions of bark. |

The Initial Injuries.—Some of the common types of initial injuries
that subsequently give rise to bark diseases are shown on Plate X XI,
The group of figures shown on this plate does not, however, include
one of the kinds most frequently noticed: these somewhat concen-
trically arranged injuries often occur with conspicuous radial clefts,
as discussed in my former papers, and as indicated in Figs. 43 and 48, .
Plate VI. In many instances, however, the concentric injuries are
not accompanied by radial ruptures, and sometimes radial clefts
occur when other types of injuries are so slight as not. to hinder
subsequent normal bark growth. ‘The sections shown on Plate XXI
are all made from apple and pear material collected before growth
started in spring (April 17, 1912, at Madison, Wisconsin).

Fig. 1 shows a condition that is frequently found in injured bark.
Dead tissues are usually evidenced in these photographs by the occur-
rence of especially dark streaks or patches, by collapsed cells or by
both. Sometimes rupturesare much more prominent than discolora-
tions, as shown in Fig. 3, which shows a very common type of rupture
or separation in the inner phloem. In Fig. I may be seen a conspicuous
combination of the collapse of discolored tissues with ruptures in the
inner phloem and cambium. At the left of the section shown in this
figure the initial injury is confined chiefly to the cambium; on the
right the principal injury occurs in the phloem, only scattered cells
in the cambium being affected. A few groups of injured cells may
also be seen about the sclerenchyma strands as well as farther out
in the cortex. Fig. 2 shows a case from pear tissues in which the
initial injury is most pronounced in the inner cortex and outer phloem, ~
with only small groups of affected cells in the outer cortex and inner
phloem. Fig.3isfromapple. It shows a marked injury of medullary
rays in the phloem, and a rupture of the phloem.

The other figures on Plate X XI show, on a larger scale, small °
areas in typically injured bark. In Fig. 4 occurs a mixture of streaks
and patches of dead and living tissues present in the outer wood and
the inner bark of apple. In the center of this figure a large ray and
much of the surrounding tissue is dead and collapsed (appearing

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 483

black). A portion of the continuation of this ray in the wood 1s also
dead, although its cambial portion is still alive. In the next ray to
the right, the outer portion is dead and that in the inner phloem and
cambium is alive; but its extension into the wood cylinder is dead:
and discolored. The phloem and cambium tissues lying between
these rays are mostly dead, but scattered living cells occur singly
or in small groups. The phloem tissue between the two rays toward
the right of the figure are in much better condition, containing only
a few scattered groups of dead cells. It is worth noting that toward
the right of the figure the main line or sheath of injury runs through
the outer phloem, while toward the left it affects chiefly the cambium
and inner phloem. However, even the most severely injured sheath
has living cells interspersed among the dead and collapsed ones.
In Fig. 5 such is not the case; here the sheath of injured tissue involves
the cambium and inner phloem on the left; at about the middle of
the figure it bends suddenly and proceeds through the phloem, leaving
the cambium alive except for occasional groups of injured cells.
In this instance the sheath of most severe injury contains few living
cells, thus practically eliminating all living connection between the
wood, or its living fringe of cambium, and the bark proper outside
the injured sheath. Fig. 6 shows a comparable condition excepting
that more.dead cells are scattered among the living tissues. In this
instance the broad ray near the center is dead in both bark and
wood, and on the left.the entire cambium and the inner phloem, with
the exception of a few cells, are collapsed. The outer cortex shows
many dead cells. Fig. 7 gives a better idea than the others of the
injury occurring in the inner bark: here the cambium and inner
phloem as well as the rays and much other phloem tissue are dead
and more or less collapsed. In the cambial region near the left,
however, is a group of two or three living cells with irregular outlines.
These cells, as well as those of certain groups in Figs. 4 and 5, are
apparently enlarged, although the apple trees from which this material
was cut seemed perfectly dormant at the time. It appears likely that
this represents the beginnings of spring growth and regeneration,
brought on early as a response to the wound stimulus. In Fig. 8
occur two lines or sheaths of severe injury, one of which involves the
cambium and the other the outer phloem. Neither of these zones
is made up wholly of dead cells, so that the sheath of living phloem
between them is not entirely without living connection with the wood

484 J. G. GROSSENBACHER

and the outer bark. Fig. 9 shows a condition much like that in
Fig. 7, excepting that larger masses of inner phloem are alive, although
not evidently affording living links between the bark and the wood.
In both Figs. 7 and 9 the sheath of dead tissue seems complete, thus
isolating the bark from the wood cylinder at these places.

The initial injuries presented in Plate X XI are not shown because
they represent the most severe cases, but because the location of the
injuries is typical and yet they are not severe enough to prevent
proper handling of the sections. As noted above, material for sec-
tioning had to be taken from portions of trees where the areas of
individual injuries or dead patches were relatively small in order to
prevent the shattering of the blocks before they were imbedded.
When small blocks were cut from the very edge of one of the more
extensive injured areas, they frequently remained intact through the
imbedding and sectioning processes; if, however, the entire block
was within such an area, its different portions usually fell apart,
separating along the planes of severest injury. This falling apart
of the blocks was less troublesome in the material collected May 1
than in that collected April 17. The blocks cut on May 29 from the
more severely affected and larger areas were extremely fragile, while
those from regions of less injury were more stable than specimens
of the same degree of injury collected April 17.

Some Changes Due to Growth and Regeneration.—The figures of
Plate XXII are made from photographs taken of sections of apple
collected May 1, 1912. These show some interesting phenomena of
growth and regeneration, and among other things suggest how and
why it is that so few bark injuries give rise to dead patches of bark.

In Figs. 10, 11 and 12 are shown cases in which the initial injury
involved all or nearly all of the cambium and a portion of the inner
phloem, with dead streaks of less extent scattered in other parts of
the phloem. In all three of these cases subsequent regeneration
growth from the living portions of the phloem resulted in establishing
a more or less definite living connection through the zone initially
involved. As a result of this growth, the material in the scattered
dead streaks in other portions of the phloem has become compressed
into ragged plates with their edges directed toward the wood. In
other parts of the bark dead groups of cells are similarly compressed
by the more or less bladdery growths from the surrounding tissues.
In Figs. 10 and 11, comparatively few of the proliferating bark-cells

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 485

appear to have established a close contact with the old wood; but in
Fig. 12, representing a section in which the dead zone had not been
so wide and still contained a few living cells, the new growth is much
more firmly attached to the old wood. Fig. 13 again is more like
Figs. 10 and 11, although its dead or injured zones are less extensive;
in the section here shown, too, some living cells capable of further
growth appear, which are attached to the old wood and have grown
outward to meet the growth from the bark. In the tissues shown in
Figs. 10, 11 and 12, the initial injuries had been most severe in the
cambial and inner phloem regions and are of the type shown at the
left in Fig. 1, Plate X XI; while Fig. 13 shows a regenerated condition
of an initial injury more like that indicated in Figs. 7 and 9, Plate
XXI, where some living bark cells or cell-groups remain attached to
the wood. In Fig. 14 is shown an effort toward recovery that is
rather remarkable and far advanced for May 1. This represents a
reaction to wounds of the type shown near the right in Figs. 1, 3, 4, 5
and 6, Plate XXI. In addition to the general compression of the
dead tissues by the growth of the living cells around them, many
proliferating cells have pushed in among the portions of the dead
sheath, thereby facilitating the re-establishment of living connections
between the outer bark and the living inner phloem that remains
attached to the wood-cylinder. Considerable injury also occurred
in the cambium, although small portions of the latter appear to have
survived. A new cambium, however, is seen to be.forming outside
among the irregular cells arising from the wound growth. The figure
does not show it as clearly as the microscope; it is beginning to take
form in the line cc. Fig. 15 represents a similar instance, except that
the initial injury was more extensive and that larger groups of dead
cells resulted. A new cambium is forming at cc, though it is in-
complete and still has compressed fragments of dead tissue in its
course. Fig. 16 seems to be a later stage of a case something like the
left-hand portion of Fig. 8, where the cambium was only slightly
injured and the outer phloem rather severely, though in more or less
isolated streaks and patches. Some of the rays are dead, although a
few are practically normal, like that near the right of Fig. 16. The
new cambium is quite distinctly indicated by the dense band cc.
In Fig. 17, comparable but severer initial injuries obtained. The re-
established living connections between the growing phloem and the
wood are few and scattered, and the injury in the outer phloem forms

486 J. G. GROSSENBACHER

a nearly complete sheath, thus isolating the cortex considerably
from the phloem. A scattering of living cells occurs, however, in
this outer-phloem zone to afford water and nutrient transfer between
the outer and inner barks sufficient to permit the outer bark to endure
at least for a time.

The figures of Plate XXII likewise give only a few of the great
variety of the injuries that were mild enough to permit more or less
regenerative growth during the spring, although some of the figures
plainly indicate that only a most favorable summer would enable
the affected bark to survive.

Some Results Found at the End of May.—Plates XXIII and XXIV
give an idea of the great variety of results following some of the initial
injuries shown on Plate XXI. The low-power views collected on
Plate XXIII make it evident that regeneration and growth of the
living cells are not all that is required to sustain the affected bark and
to keep it from dying in early summer. Figs. 18 to 22 inclusive show
some of the milder forms, while Figs. 23 to 26 indicate various stages
and degrees of injury resulting in the death of patches of bark.

Fig. 18 shows an advanced stage of an initial injury of the type
shown in Figs. 7 and 9g, and, in later condition, in Figs. 10 and It.
The new cambium sheath arose much after the manner shown in
Fig. 17. The compressed fragments of dead tissues, present at the
time spring growth started, are noticeable in the new wood (nw)
as well as in the old phloem (op). The initial injury was so severe
that the old wood and bark-rays were not continued by the new
growth; new rays are just becoming differentiated on both sides of
the new cambium (nc). In this case the outer bark seems to have
established sufficient living connection with the regenerating inner
bark to continue its normal functioning, but the connection between
the old (ow) and the new wood (nw) seems to be insufficient in places,
for the new cambium (nc) has also developed in the inner phloem.
In this case the initial injuries in the outer phloem and cambium were
so extensive as greatly to delay the development of the new cambium
(mc), as seen near the middle of the figure. No definite new wood
cells have yet been formed at this point, although on both the right
and left sides a considerable layer of new wood has resulted and the
new cambium appears practically normal. Fig. 20 shows some inter-
esting irregularities in the distribution and configuration of the initial
injuries. They had evidently been of the type shown in Figs. 1, 5

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 487

and 9, and the new cambium apparently started to form in the inner
phloem just outside the dead cambium as shown by the light line near
the wood, at the lower left. But because of marked irregularities in
the distribution of the sheath of severest injury, the first new cambial
initial was supplanted by one farther removed from the wood and
overtopping the irregularities, as shown at uc. Toward the left of
this figure the injury had been so severe as wholly to inhibit the de-
velopment of a new cambium, although the cell proliferation had
occurred that ordinarily precedes the production of a new cambium.
Some of the small groups of living cells in that region had become
light brown in color and were evidently dying. Toward the left end of
the new cambium only scattered groups of new wood cells had de-
veloped. The portion of old inner phloem, separated from the rest
by the mantle of injured tissue (near the center), had grown some-
what, but no definite cambium had been evolved. In the section
shown in Fig. 21 similar isolated strips had developed into wood cells,
even though the new cambium, as in Fig. 20, developed farther out
in the phloem and produced a rather broad layer of new wood outside
the zone of greatest injury. This type of figure seems to have been
developed as the result of initial injuries like those shown in Figs. 5,
7 and 9. Much dead tissue was compressed into irregular masses
in the outer phloem and cortex. The new cambium (uc) has a brown-
ish tinge and seems to be much collapsed. It should be noted in
this case that the phloem left attached to the old wood was transformed
into wood without leaving a cambium. Fig. 22 is interesting chiefly
on account of the fact that injuries in the outer cortex resulted in the
development of a new phellogen layer or cork cambium within (ph).
Fig. 2, Plate X XI, shows an injury occurring mainly in the inner
cortex, that is often similarly cut off by a phellogen developing in
the outer phloem. Fig. 23 is somewhat comparable to the left portion
of Fig. 20, in that no new cambium has developed, although con-
siderable regeneration growth has occurred. The cortex is prac-
tically uninjured and therefore appears normal from the outside, but
both the outer and inner phloem are severely affected and the cambium
is entirely dead, except in isolated streaks like that shown near the
left. But even in this severely injured phloem occur groups of living
cells, though they are more or less completely isolated by dead tissues.
Many enormous bladdery outgrowths from the living cells are forced
into the dead masses. In some places the living and in others the

488 J. G. GROSSENBACHER

dead tissues predominate in the phloem. It may be noticed that on
the left the bark is thicker than it is on the right. This results from
the presence of larger groups of living cells in this portion of the
phloem; indeed, it appears that groups of wood cells (appearing in
the photograph as rather dim whitish patches) have in some way
arisen in this region. The process had advanced further in the
specimen shown in Fig. 24. Here some of the bark had died to the
wood, and, because of the presence of much dead tissue in the inner
phloem and the old cambium, the callus is a rather sickly affair.
It includes the repaired phloem considerably speckled with masses
of dead tissue, and the badly injured cortex. In the lower right corner
occurs a strip (white) where the phloem is being transformed into
wood cells, yet no definite cambium is in evidence. Although not
shown in this figure, the slide from which this photograph was made
shows abundant hyaline fungus mycelium in the dead bark, even in
the dead cambium and the old inner phloem between the callus and
the old wood. Fig. 25 shows a similar case in which the most severely
affected area was very narrow, thus permitting its use in this study
without its falling apart. In this instance the callus is much better
developed, having a definite cambium and a layer of new wood.
The old inner phloem and the old cambium were also dead for some
distance back of the nose of the callus. Some fungus mycelium was
present in the dead bark. Fig. 26 is made from the margin of a larger
area that had sunken in, like that shown in Fig. 41 of Plate XXV.
The marginal callus was much like that of Fig. 25, and the presence
of fungus is indicated by the pycnidia (of Sphaeropsis?) showing
under the periderm toward the left.

On Plate XXIV are brought together some higher-power views
giving greater detail, though in some instances cell outlines are
necessarily more hazy. Figs. 27 and 28 show clearly the remains of
the dead cambium and inner phloem; they also prove that even as
late as the end of May substitution growth is in progress along the
inner side of the repaired bark and has established better connection
with the old wood. This latter fact is indicated by the presence of
excessively large round cells, that appear to be filling the gaps left
by the shrunken dead tissues. Fig. 30 shows the development of a
new cambium (nc) between the dead inner phloem and sheets of dead
tissue in the outer phloem. In Fig. 31, groups of dead cells appear in
the former position of the inner phloem, around and among which

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 489

some of the new wood (nw) cells are becoming discolored. This
intermingling of living and dead groups of cells is most common in the
phloem. The position of the new cambium (nc) is occupied chiefly
by structureless masses having a yellowish brown color. The cortex
and outer phloem had only scattered groups of dead cells. No
fungous hyphae could be detected in this specimen. Fig. 32 shows
tissues which had sustained more severe initial injuries but which,
because of the small area of the most severely affected part, endured
quite well, while in the case of milder injury but of larger area shown
in Fig. 31 the tissues seemed to be dying. In the tissues shown in
Fig. 32, the outer phloem also is largely killed but the cortex is in fair
condition. A new cambium (nc) is forming near the old wood. At
this place no living connection appears between the bark and the
wood, but because of the smallness of the region so severely affected,
the necessary water and nutrients seem to reach it from the sides
of the injured patch where such connections do occur. Fig. 33 shows
a case in which much of the inner phloem had been killed and in
which very active filling-growth is occurring. A faint indication
of the formation of a new cambium can be seen here and there; a
spot of this kind appears near the center of the figure. The photo-
graph from which Fig. 34 was made shows a very large group of dead
phloem tissue which has been only partly permeated by proliferating
cells arising from living adjoining cells. In some cases the initially
killed strips of tissue form an anastomosing network lengthwise through
the bark; in extreme instances only anastomosing streaks of inner
bark may be alive while the greater mass of the bark is dead. A
cross section of such a living streak in great masses of dead tissue
may be seen in Fig. 35. In this instance large groups of dead cells
also occurred in the inner cortex, although when the specimen was
cut (May 29) its external appearance was practically normal.

From another type of initial injury enlargements as well as wood-
exposures occur on trunks of trees. Such cases of excessive enlarge-
ments on trunks may sometimes develop from initial injuries of the
type shown in Fig. 3, Plate X XI, in which a definite separation or a
concentric cleft has resulted, and where at the same time the repair
growth and connection with the wood are such as to prevent the dying
of the loosened outer bark. When radial clefts run through such
areas and the bark is otherwise sufficiently intact to withstand the
drying action of air, the repair growths may turn the edges of the

490 J. G. GROSSENBACHER

loose bark outward, while if no radial cleft occurs the enlargement
may look like that on a maple trunk shown in Fig. 40, Plate XXV.
The resulting repair growths are not always uniform even when they
follow the kind of initial injury that separates the bark from the wood.
In some cases the loosened bark has but few injuries (Fig. 3, Plate
X XI); in other cases, or perhaps even in other portions of the same
affected area, the outer phloem as well as the cortex may have so
many groups of dead cells scattered among the living parts that
the entire bark dies. That has been the case in the lower portion of
the stem part shown in Fig. 40. A section through the upper part
of this maple trunk reveals a condition like that shown in Fig. 37.
Here the discolored line oc represents the position of the cambium
when the injury occurred. The initial injury not only resulted in a
line of separation in the inner phloem like that shown in Fig. 3, but
involved cell-groups in the cambium proper as well as in the middle
and outer bark. As in the case shown in Fig. 3, however, the inner
phloem had sufficient living connection with the wood to permit the
development of a cambium that persisted, excepting in the bare
region shown toward the lower end of Fig. 40, where it died along
with the loosened bark outside. A cambium also developed in the
inner part of the loosened bark shown in Fig. 3, running through
the outer phloem. After the production of a sheath of new cells in
this new growing zone, the middle ones became wood and those along
both outer sides continued as cambial zones. In that way one growing
zone was converted into two, which separated more and more as the
older cells toward the middle were converted into wood. The sheath
of new wood (nw) just within the old outer bark (0b, Fig. 37) arose in
that manner, and has a cambium on each side. The low-power
views shown in Fig. 36 (a-e) are photographs of sections cut from
blocks obtained from the specimen shown in Fig. 37. Fig. 36a shows
the old outer bark (0b) of Fig. 37, with only a small portion of the
new cambium (nc) included. The new bark shows compressed inclu-
sions of dead tissues resulting from the initial injury. Figs. 36) and
36d are so tightly pressed against each other in Fig. 37 that the two
barks seem to be one. Fig. 36e is taken from the line oc in Fig. 37
and shows more clearly the similarity of this line to the figures ob-
tained from apple, such, for example, as Fig. 18. Fig. 39 is a more
highly magnified view of a portion of Fig. 36e. It shows the presence
of dead groups of old cambium cells adhering to the old wood (ow).

~

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 491

It is evident that in most places along the line oc of Fig. 37 the cam-
bium, giving rise to the new wood (nw) outside it, also arose much
like that shown in Figs. 14, 15, and 17, from cell divisions occurring
in the adhering layer of phloem (Fig. 3). Trécult seems to have
obtained very similar results by pulling loose and then tying back a
piece of bark.

Fig. 39 shows another fact of special interest, for it makes it
apparent that much of the new wood formed about the original line
of injury subsequently dies and becomes discolored. Even in portions
of the line cc of Fig. 37, where the new wood is practically continuous
with the old and where the rays are not even interrupted, there has
been more or less of this discoloration and dying of the new wood,
as shown in Fig. 42. Here the initial injury seems to have involved
chiefly the outer end of the wood rays and scattered, but small,
groups of cambium cells. Nevertheless a narrow, ragged-edged zone
of new wood died and became discolored, as Fig. 42 shows.

The Development of Crown-Rot from the Initial Injurtes.—As sug-
gested above, bark-injuries may or may not be accompanied by evident
radial clefts, and when’ they are not evident externally they are yet
often present, as may be gathered from Fig. 36a. The old bark (0d)
is shown to be run through by numerous small, radial rifts that repre-
sent incomplete ruptures which were afterwards filled by prolifera-
tions from adjoining cells. Fig. 38 shows a case of this kind, also
in maple, where apparently the bark was only separated from the
wood, and yet where so many of these tiny radial rifts occurred that
the bark involved is evidently dying. In this case the whole of
the cambium as well as much of the inner phloem died as a result
of the initial injury. The rest of the bark was still alive when cut
on May 28. Although no definite new cambium had yet developed
in this loosened portion of bark, the spring growth of wood is seen
to be considerable on both sides of the wound. Callus-roll formation
has made an evident beginning around the injury, even though the
bark involved is not dead.

On the other hand, Figs. 43 and 45, Plate X XVI, show apple
trees in which internal bark-injuries, resulting in a separation of the
bark from the wood, were accompanied by evident radial clefts.
The former is shown before and the latter after the loose bark was

4 Trécul, A., Production du bois par l’écorcz des arbres dicotylédonés, Ann.
sci. Nat. Bots IIP19o: 257: “1855.

492 J. G. GROSSENBACHER

removed. Fig. 44 shows the same tree as that in Fig. 43 with the
loose bark removed, making it apparent that the tree was half girdled,
though only a fairly narrow band (not exceeding 17 cm. in vertical
width) was involved. The loosened bark shown as removed in Fig.
45 had several short radial clefts, though otherwise the bark appeared
normal while it was still on the tree. When removed, the inner
surface of the bark and the wood thus exposed had a rusty brownish
color. On cutting with a knife it was found to contain numerous
closely scattered, small dead spots, which in some places had coalesced
to form ragged, dead patches as much as one or two centimeters in
diameter. These patches often involved all but the outermost layers
of the cortex and sometimes showed through the periderm or scaly
outer bark in the form of dead spots from one to five millimeters in
diameter. In places, however, this loose bark was found to have
developed a new cambium in addition to bladdery outgrowths from
the inner phloem, thus tending to re-establish connection with the
wood cylinder. Similar, though scattered, outgrowths had also de-
veloped from outside the wood cylinder, but the actual connection
established was evidently slight; for on June 24, when it was removed,
considerable areas of these proliferations had died and turned a rusty
brown. In fact, disorganization seemed to have set in over a large
part of the inner surface of this bark.

When, on May 7, the loosened bark was removed from the tree
shown in Figs. 43 and 44, only a slight discoloration was noticed on
the contact surfaces. The loosened bark appeared perfectly normal
on the outside, with the exception of the presence of a wide radial
cleft. Numerous scattered groups of dead tissue were found in the
older phloem and inner cortex. Proliferation growth had been abun-
dant, and in a few places it appeared that cambium was in process
of formation.

In cases of injury in which the affected bark does not die and where
but one substitute cambium develops, only the discolored line in the
wood afterward remains as a permanent record. This line is marked
oc in Figs. 18 and 19, Plate XXIII, and 37, Plate XXV. “Fig.%46,
Plate XXVII, is an especially clear illustration. It represents a
cross-section of the base of a large apple-tree trunk from an orchard
in which the initial injuries, leading to the development of crown-rot,
had occurred on many trees the same number of years back, as is
indicated by the radial-growth zones outside the conspicuous line of

‘

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 493

discoloration in the wood. It should be noted also that at least one
radial cleft occurred in the bark of the tree represented in Fig. 49:
the fusion is still incomplete. However, in the tree, a section of
which is shown in Fig. 50, a portion of the affected bark died and the
entire wood cylinder, up to the line represented by oc in Fig. 18
(the outer boundary of the wood at the time of injury), has completely
decayed. In a few points, just as in Figs. 3,39 and 42, some of the
wood produced afterward also died and decayed. This shows that
decay depends as much upon the death of the wood as upon the
presence of wood-rotting fungi. Bark’ and wood having groups or
streaks of dead tissue naturally contain relatively large quantities of
air, and sapwood dying from such bark injuries is full of both air
and stored food. The high air content of such wood led Miinch® to
conclude that the presence of the excessive air is the factor that permits
wood-rotting fungi to vegetate actively in such tree trunks. Based on
these conclusions of Miinch, Appel® has been led far afield in his
discussion of the factors governing the activity of wood-rotting fungi.
The fact that such wood is killed while it is young and full of stored
food makes it evident that it differs materially from ordinary heart-
wood that has become depleted of most of its stored food (including
the layers of hemicellulose usually present on the inside of its cell-
walls) before it became lifeless. It seems more likely that wood-
rotting fungi thrive uncommonly well in such wood because it contains
large quantities of stored food and masses of more or less disorganized
and therefore non-resistant protoplasm, rather than because of the
great abundance of air present.

The small apple tree shown in Fig. 46 and the large ones of Figs.
47 and 51 are examples in which the most,severely affected bark died.
In those shown in Figs. 26 and 51 a complete girdle is involved, while
in that of Fig. 47 only about three fourths of the bark succumbed.

Comparison of Effects on Large and Small Trees.—The initial
injuries, from which crown-rot and some other bark diseases arise,
are the same on large and small trees; the differences usually
noticed afterward result from subsequent changes owing to differ-
ences in the thickness of the bark and in the diameter of the

5 Miinch, E., Untersuchungen tiber Immunitat und Krankheitsempfanglichkeit
der Holzpflanzen, Naturw. Zeit. Forst. Landw. 7: 54; 87; 129. 1909.

6 Appel, O., The Relations between Scientific Botany and Phytopathology,
Ann. Mo. Bot. Gard. 2: 275. I915.

A494 J. G. GROSSENBACHER

stems or branches involved. It is also shown that if the area of the
bark most severely affected is large compared with the circumference
of the stem involved, the result is more serious than when the injured
patch is comparatively small. This holds for both large and small,
as well as for young and old stems. If the section from which Fig. 26
is made were photographed whole and magnified, a picture much like
Fig. 50 would result, although the wood cylinder within the injured
zone was not decayed or even entirely dead when cut on May 29.
The new growth of wood shown in Fig. 26 as well as that in Fig. 25
consisted of only a very thin layer, while the wood shown in Figs. 49
and 50 represented several years’ growth. Fundamentally, however,
these sections are not only comparable but very similar.

THE CAUSE OF THE INITIAL INJURIES

The work so far has clearly shown that the initial injuries in the
bark of trees that result in crown-rot arise during the dormant season,
but their cause has not been definitely established. The years of
observation and a few experiments together with the histological
study here reported, indicate the most probable factors, and thus pave
the way for an experimental study of the problem. In general terms
it may be said that these initial injuries are due to a lack of adjustment
between radial or bark-growth and the environment.

Some Facts about Bark-Growth as Related to the Development of
These Injuries —In the study of forest trees it has been shown that
growth and development proceed in a wave-like manner. The various
functions, the size of cells, and the amount of annual growth increase
to maxima in certain stages of a tree’s life, and decrease to minima
again at other stages. These periods or cycles are repeated at inter-
vals more or less characteristic of a species. Kapteyn’ calls attention
to growth cycles that may be traced in the wood and extend over
periods of 12.4 years, apparently independent of the species. Bailey
and Shepard® found that the length of coniferous tracheids varies
in more or less definite cycles usually ranging from 35 to 80 years, and
apparently differing in different species.

It is a well-known fact, for instance, that at a certain age of a

“ Kapteyn, J. C., Tree-Growth and Meteorological Factors, Réc. Trav. Bot.
Néerland. 11: 70. I914.

§ Bailey, I. W., and Shepard, H. B., Sanio’s Laws for the Variation in Size of
Coniferous Tracheids, Bot. Gaz. 60: 66. I9I5.

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 495

woody plant the development of primary bark is supplanted by the
formation of secondary bark, apparently because of the inability of
the primary tissues to continue their adjustment to the increase of
the stele. This transition stage in trees usually occurs at an age
characteristic for the species. Afterward the bark continues to in-
crease in thickness, and to be distended, growing somewhat radially
for a certain number of years; then a new phellogen is again formed
inside that portion of the bark that is no longer able to undergo suff-
ciently rapid tangential growth, and a fresh portion of lifeless bark
results which is more or less continuous around the trunk—the results
in this respect also varying with the species.

It appears that there is a close relation between the growth cycles
mentioned above and the periodicity that may usually be noticed in
regard to bark cycles. The growth and cell-size minima seem usually
to fall in the season just preceding a new period of bark-roughening,
while the maxima are usually reached during the second third of the
time elapsing between the occurrence of two minima. ‘The environ-
mental variations, and, in case of cultivated trees, the culture and
tilth given, have a marked influence upon the prominence of this
periodicity.

In a smooth-barked stem that portion of the bark outside the oldest
circle of sclerenchyma (the cortex) often undergoes much, although
limited, growth. Sections of apple and pear stems from material
fixed at different times of the year indicate that during seasons of
much or of late radial growth cortical growth sometimes continues
very late and is not completed by the time the period of dormancy
arrives. ‘The increase in diameter of wood necessitates and 1s followed
by an increase in the area of the bark. When an adverse change in
the weather conditions interferes before this cortical growth is com-
pleted, the dormant period must be passed with the outer bark in
this unfinished condition. In such a case the bark is often under
considerably higher transverse tension than it is in cases in which its
cortical growth has been finished. In instances also in which bark
growth has been very slight during some years, the cell walls of the
cortical tissues and those in the outer phloem are thickened to such
an extent that a rather rapid resumption of radial growth is not
immediately followed by cortical growth, and therefore high bark
tension ensues. If such hardened outer bark is eventually forced
into growth late in the season, some of the cells necessarily pass

496 J. G. GROSSENBACHER

through the dormant season in immature condition, and thus are
likely to become injured. N6rdlinger? found by peeling tests that
cambial activity precedes cortical growth and may continue after
cortical growth ceases. In some cases, however, cortical growth
continued later than cambial activity.

R. Hartig!® describes several cases in which the bark of a very
high percentage of forest trees was burst and injured at certain
places two years after those forests had been thoroughly thinned.
After thinning, the trees grew as much in one year as they had grown
before in many years. His conclusion, that the bark burst in early
summer owing to the rapid radial growth of the wood, can fortunately
be more carefully examined because he gives a photographic record of
cross-sections. These figures prove beyond question that the bark
was split and separated from the wood during the dormant’ season
preceding the growing season in which he assumes the splitting to
have occurred. From Fig..52, Plate X XVII, which is a reproduction
of one of Hartig’s figures, it is apparent that the bark injury occurred
between the growing seasons and not while growth was going on
because the lines of injury and separation coincide with the line
separating the wood of two growing seasons. Another case which
Hartig gives in some detail, in which a high percentage of the trees
in a thinned forest sustained bark injuries just above or at the ground
line a few years after thinning, is also of decided interest. In this
instance he concludes that the rank growth of herbaceous plants
developing about the tree trunks after thinning prevented proper
aeration, excluded light, and thereby injured the bark. But in this
case as in the former, cross-sections show that the injury occurred
during the dormant season, when aeration was probably good. The
chief difference between these two cases lies in the fact that in the
former instance the tension reached a high enough point to rupture
the bark as well as to loosen it, while in the latter the tension was
less. It seems possible that in one instance the bark was more
resistant to radial rupture than in the other, though it is likely that
some additional factors are involved in the occurrence of radial clefts.

° Nordlinger, H., Wann beginnt Bast, wann Lederschicht der Rinde sich zu
lozen? Centralbl. Gesamt. Forstwes. Wien. 5: 128. 1879.

10 Hartig, R., Zersprengen der Eichenrinde nach plétzlicher Zuwachssteigerung,
Untersuch. Forstbot. Inst. Miinchen 1: 145. 1880.

; Das Zersprengen der Hainbuchenrinde nach plétzlicher Zuwachssteiger-
ung, Untersuch. Forstbot. Inst. Miinchen 3: 141. 1883.

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 497

Environmental Factors Having a Causal Relation to the Injurtes.—
My work in northern orchards showed that bark injuries which are
caused by the combination of immaturity of tissues and the occurrence
of low temperature may give rise to crown-rot. In Florida the
occurrence of a temperature only about two degrees or less below the
freezing point of water, when certain bark tissues of citrus trees are
immature, may result in similar injuries and give rise to equally
destructive diseases. Severe droughts often cause similar injuries.
When the bark of citrus trees is dormant, it will endure temperatures
even below — 10° C. and severe droughts without serious injury.

It is still uncertain whether this bark injury is due chiefly to simple
physical causes, such as contraction or to chemical and physiological
changes induced in the protoplasm by low temperatures and drought,
or to both sets of factors acting together. Asa matter of fact, changes
of both kinds take place in plants subjected to low temperature and
untimely droughts, and we have fairly tangible evidence that both
may be injurious.

The photographs submitted with this paper give ample evidence
that high tensions, and even ruptures, accompany some at least of the
more severe bark injuries. Trunk measurements previously published
also show the occurrence of high tensions.“ In some cases, however,
no actual ruptures appear to result, and yet tissues become discolored
before the commencement of the next vegetative period. Injurious
low-temperature tensions of less degree than those required to rupture
the bark are evidently of frequent occurrence, and these are apparently
responsible for’much of the bark injury afterwards resulting in disease.
An extreme form of this effect is shown in Fig. 37a. Some of the
milder tensions are also shown in Figs. 3, 4 and 5 of Brown’s” recent
paper. Sorauer® has given much attention to this type of injury

11 Crown-Rot of Fruit Trees: Field Studies, N. Y. Agr. Exp. Sta. Tech. Bull.
232.30 1O12,

‘’Sorauer, P., Experimentelle Studien tiber die mechanischen Wirkungen des
Frostes bei Obst- und Waldbaiimen, Landw. Jahrb. 35: 469. 1906.

, Weswegen erkranken Schattenmorrellen besonders leicht durch Monilia?
Zeit. Pflanzenkr. 22: 285. 1912.
——,, Einige Experimente zum Studium der Frostwirkungen auf die Obstbaume,

Die Naturw. I: 1055; 1094. 1913.

#2 Brown, H. P., Growth Studies in Forest Trees. 2. Pinus Strobus, Bot. Gaz,
50: 10771915:
, Altes und Neues tiber die mechanischen Frostbeschadigungen, Zeit.
Pflanzenkr. 24: 65. I914.

498 J. G. GROSSENBACHER

and even applies his low-temperature tension hypothesis to cold
injury of herbaceous plants.

Some interesting advances have been made in recent years in the
study of the chemical and physiological side of this question, but
unfortunately the investigators interested in this phase of the subject
have thus far given no attention to the more simple physical con-
comitants presented by Sorauer in the papers just referred to. It is
in fact usually assumed that the earlier works had decided this matter.
Nageli,“ for example, made some studies of this type and concluded
that since walls of Spirogyra cells killed by low temperature are not
ruptured, death must be due to changes induced in the protoplasm.
Kunisch maintained that low temperatures induce harmful irre-
versible changes in certain components of the protoplasm that result
in the death of tissues; that in some plants such changes may even
occur above the freezing point,’ although in others a temperature
much below freezing is necessary to cause injurious effects. Fischer,!”
after very fully discussing the literature and giving the results of his
own extensive experimental study of the problem, concluded that the
low-temperature death-point of plants usuallv does not vary more than
two, though it may vary as much as ten, degrees. On the other
hand, Winkler'® found that the condition of the protoplasm at the
time of the occurrence of the low temperature has much to do with
the degree of resistance or injury.

Some have held that low-temperature injury results from ice-
formation; others believe that it is the withdrawal of water during
freezing that causes the injury.'® Apelt and others?® have brought

—, Uber Frostschorf an Apfel- und Birnenstammen, Zeit. Pflanzenkr. 1: 137.
189g.

; 14 Nageli, C., Ueber die Wirkung des Frostes auf die Pflanzenzellen, Sitzungsb.
Akad. Wiss. Miinchen 1: 264. 1861.

1 Kunisch, E. H., Ueber die totliche Einwirkung niederer Temperaturen auf
die Pflanzen, Inaug. Dissert. Breslau. 1880.

16 Molisch, H., Untersuchungen tiber das Erfrieren der Pflanzen. Jena. 189%.
, Das Erfrieren von Pflanzen bei Temperaturen tiber dem Eispunkt,
Sitzungsb. Akad. Wiss. Math. Naturw. (Wien) 105: 82. 1896.

17 Fischer, H. W., Gefrieren und Erfrieren, eine physico-chemische Studie,
Beitr. Biol. Pflanz. 10: 133. I9II.

18 Winkler, A., Uber den Einfluss der Aussenbedingungen auf die Kalteresistenz
ausdauernder Gewachse, Jahrb. Wiss. Bot. 52: 467. 1913.

19 Miiller-Thurgau, H., Ueber das Gefrieren und Erfrieren der Pflanzen, Landw.
Jahrb. 9: 133. 1880.

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 499

out some very pertinent reasons for their contention that low-tem-
perature injury is not due to the withdrawal of water.

Even in much of the older literature on this subject, one finds
interesting comments on the apparent changes that living substances
and stored foods of plants undergo on the gradual approach of cold
weather. It was noted, too, that plants thus modified are more
resistant to low temperature. The more modern work on this phase
of the subject, by Gorke*! and Lidforss,” as well as the work by
Bartelzsko,” formed a basis for the most interesting researches of
Maximow.*% The latter author showed that it is not the molecular
concentration of the cell sap that is most significant, but that the
physical nature of the solutes determines the degree of resistance
afforded a plant. It was found that the introduction into plant

Greeley, A. W., On the Analogy between the Effects of Loss of Water and
Lowering of Temperature, Amer. Journ. Physiol. 6: 122. 1902.

Matruchot, L., et Molliard, M., Modifications produites par le gel dans la
structure des cellules végétales, Rev. Gen. Bot. 14: 463; 522. 1902.

Bruijning, F. E., Zur Kenntniss der Ursache des Frostschadens, Forsch.
Gebiete Agr. Phys. 19: 485. 1896.

Chandler, W. H., The Killing of Plant Tissue by Low Temperature, Mo. Agr.
Exp. Sta. Research Bull. 8: 143. 1913.

20 Apelt, A., Neue Untersuchungen iiber den Kaltetod der Kartoffel, Beitr.
Biol. Pflanz. 9: 215. 1909.

Rein, R., Untersuchungen tiber den Kaltetod der Pflanzen, Inaug. Dissert.
Halle. 1908.

Voigtlander, Hans, Unterkiihlung und Kaltetod der Pflanzen, Beitr. Biol.
Pflanz. 9: 359. 1909.

Mez, Carl, Neue Untersuchungen tiber das Erfrieren eisbestandiger, Pflanzen,
Flora 94: 8. 1905.

21 Gorke, H., Uber chemische Vorginge beim Erfrieren der Pflanzen, Landw.
Vers. Stat. 65: 149. 1907.

2 Lidforss, B., Die wintergriine Flora, eine biologische Untersuchung, Lunds
Universitets Arsskrift, No 22nr. 53.) 1007:

*3 Bartelzsko, H., Untersuchungen tiber das Erfrieren von Schimmelpilzen,
Jahrb. Wiss. Bot. 47: 57. IgIo.

24 Maximow, N. A., Chemische Schutzmittel der Pflanzen gegen Erfrieren, I.
Bericht. Deutsch. Bot. Ges. 30: 52. 1912.
, Chemische Schutzmittel der Pflanzen gegen Ertrieren, II. Die Schutz.
wirkung von Salzlésungen, Bericht. Deutsch. Bot. Ges. 30: 293. IgI2.
, Chemische Schutzmittel der Pflanzen gegen Erfrieren, III. Uber die
Natur der Schutzwirkung, Bericht. Deutsch. Bot. Ges. 30: 504. 1912.
, Experimentelle und kritische Untersuchungen tiber das Gefrieren und
Erfrieren der Pflanzen, Jahrb. Wiss. Bot. 53: 325. I9gI4.

500 J. G. GROSSENBACHER

tissues of substances having relatively high cryohydric points gave
very little added resistance even though their molecular concentrations
were high, while the introduction of substances with very low cryo-
hydric points afforded much added resistance, even at fairly low con-
centrations. He concluded that since the low-temperature death-point
can be lowered by the introduction of substances of low cryohydric
points, protoplasm can have no specific death-point, but that the
death-point depends upon the temperature at which water and other
substances are crystallized out. Some interesting experiments by
Gassner and Grimme” also show that Maximow’s results have a wide
application.

The part played by enzymes in plants injured by low temperatures
is still rather uncertain, though they are probably involved in the
many protoplasmic changes that result. It seems very likely, too,
that some of the harmful changes that are caused by low temperatures
are due to the perverted action of enzymes no longer properly con-
trolled by substances that have been modified by the cold. Kras-
nosselsky”® found that an oxidizing enzyme evinced more activity in
sap expressed from a frozen plant than in that obtained from living
tissues. The browning of sap expressed from tissues injured by cold.
is suggestive of the brown-spotting of herbaceous plants obtained by
Molisch in experiments cited above, in which low temperatures above
the freezing point were used. Md6bius?’ obtained very similar results.
At any rate, it has been well established that the best known enzymes
present in plants are not destroyed by ordinary low temperatures,
for Palladin?® and his students use low temperatures to kill tissues
before extracting enzymes. Kovchoff?’ maintains that protein-split-
ting enzymes are very active in cold-injured plant tissues, though his
experiments seem to admit the assumption that perhaps the proteins
were split as a direct result of the low temperature and that the

25 Gassner, G., und Grimme, C., Beitrage zur Frage der Frostharte der Getreide-
pflanzen, Bericht. Deutsch. Bot. Ges. 31: 507. 1913.

°6 Krasnosselsky, T., Bildung der Atmungsenzyme in verletzten Pflanzen,
Bericht. Deutsch. Bot. Ges. 23: 142. 1905.

27 Mébius, M., Die Erkaltung der Pflanzen, Bericht. Deutsch. Bot. Ges. 25:
67. 1907.

28 Palladin, W., Uber den verschiedenen Ursprung der waihrend der Atmung der
Pflanzen ausgeschiedene Kohlensaure, Bericht. Deutsch. Bot. Ges. 23: 240. 1905.

29 Kovchoff, J., Enzymatische Ejiweisszersetzung in erfrorenen Pflanzen,
Bericht. Deutsch. Bot. Ges. 25: 473. -1907.

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 501

changes he records were mainly due to the subsequent increase in the
activity of the oxidizing enzyme present in the injured tissues. From
his experiments with low temperatures, Schaffnit®® holds that tem-
peratures a little above the freezing point induce chemical changes
in the protoplasm that convert labile into more stable compounds.

The discoloration noted in the new cambium of Fig. 31, Plate XXIV,
and in the new wood shown in Figs. 39 and 42, Plate X XV, is probably
due to the diffusion into the new tissues of some injurious by-products
from cells affected by the cold. This aftermath of low temperature
effects seems to account for the fact that bark-injured trees may be
sick for some time before they die or recover, and that some may
remain dwarfed for years.24. Mutinch*® found that the discoloration so
commonly present in the heart wood of some trees is not due to sub-
stances secreted by living cells, but to oxidation products arising in
dead cells. The presence of fungi increases the extent of the browning.
Goethe® has also made notes on this diffusion into living tissues of an
injurious substance from tissues injured by low temperatures. He
found that even in cases of severe bark-injury on the lower portions
of tree-trunks, the affected trees survived if this discolorizing substance
did not diffuse throughout the sapwood, while if its diffusion was
rapid and extensive the tree usually died in a fairly short time. The
same facts were found to hold regarding branches at the crotches of
which bark-injury had occurred. Sorauer** thinks it likely that the
injurious substance which diffuses from dead protoplasm into sur-
rounding cells is an enzyme which arose from the disintegrating proto-
plasm. Active growth is said to check this diffusion or to make it
harmless. Bailey® found, also, that oxidizing enzymes are largely
responsible for discolorations developing in new green lumber during
warm, moist weather.

30 Schaffnit, E., Studien tiber den Einfluss niederer Temperaturen auf die
pflanzliche Zelle, Mitth. Kaiser. Wilh. Inst. Landw. Bromberg 3: 93. I910.

- 81 Gutzeit, E., Dauernde Wachstumshemmung bei Kulturpflanzen nach voriiber-

gehender Kalteeinwirkung, Arbeit. Biol. Anstalt Lands. Forstwirts. 5: 449. 1907.

22 Miinch, E., Uber krankhafte Kernbildung, Naturw. Zeit. Forst-Landw. 8:
533; 553. I9Io.

83 Goethe, R., Die Frostschaden der Obstbaume und ihre Verhiitung. Berlin.
1883.

34 Sorauer, P., Was bringen wir mit den Samenriiben und Samenkndueln der
Zuckerriiben in den Boden? Zeit. Pflanzenkr. 24: 449. 1915.

% Bailey, I. W., Oxidizing Enzymes and Their Relation to Sap Stain in Lumber,
Bot. Gaz. 50: 142. I9gI0.

502 J. G. GROSSENBACHER

It seems possible, too, that certain degrees of severity in the en-
vironment disturb the equilibrium between the enzymes in cells that
are in a susceptible condition, and thus eventually lead to disintegra-
tion which may culminate in the death of the tissues. Such an assump-
tion might lead to the surmise that the disintegrations evident in the
cambial region shown in Fig. 31 are due to an excess of a hydrolyzing
enzyme or to the absence of factors that normally inhibit hydrolytic
action at a certain stage of growth, and permit the usual maturing
processes to go on to completion. Lepeschkin’s** studies of the effects
of high temperatures on protoplasm, as well as some of the results
noted by Overton*’ when using heat to kill portions of Cyperus stems,
are interesting in this connection because they suggest the possibility
that opposite extremes of temperatures may, after all, have some
parallel effects.

Although the researches that have been cited on the chemical and
physiological phases of low-temperature injury are apparently of
fundamental importance, they give only a very meager understanding
of what seems to be a small portion of the process. As already men-
tioned, some of the simpler physical effects of a lowering of the tem-
perature must also be brought into proper relation with the physio-
logical changes induced. After these simpler matters have been dis-
posed of and a fair understanding of the development of bark injury
has been attained, the practical phases of the problem will still be
unsolved. One who has given this subject much thought cannot
avoid the striking fact that 7m nature these injuries ordinarily occur
not so much on account of the degree of the low temperature reached,
as because of the condition of the bark at the time of its occurrence.

SOME OTHER BARK DISEASES RESULTING FROM INTERNAL BARK
INJURIES

In the course of my study of crown-rot some other bark diseases
were also traced to their origin in bark injuries very similar to those
often giving rise to crown-rot. The so-called ‘‘cankers,’’ “sun-scorch,”’
and the premature roughening of bark on smooth-barked apple and
pear trees were the types most commonly encountered. The latter

°6 Lepeschkin, W. W., Zur Kenntnis der Einwirkung supramaximaler Temper-
aturen auf die Pflanze, Bericht. Deutsch. Bot. Ges. 30: 703. 1913.

37 Overton, J. B., Studies on the Relation of the Living Cells to Transpiration
and Sap-flow in Cyperus, Bot. Gaz., 51:28; 102. I9II.

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 503

type of injury was studied by Sorauer®® about twenty-five years ago.
By comparing Fig. 29, Plate XXIV (copied from Sorauer), with Fig. 2,
Plate X XI, it is evident that the initial injuries were very similar in the
two cases, except that that shown in Fig. 2 is much more severe.
Sorauer found that this premature roughening is of especially frequent
occurrence on rapidly growing varieties of fruit trees when they are
from six to eight years old. The same cold spell that resulted in the
bark-roughening described by Sorauer had also caused the bark of some
trees to rupture and of others to ‘‘scald”’ or die to the wood in long
patches. Some cambium, medullary rays, protoxylem, and pith
tissues were killed and discolored; in the cortex the larger patches of
dead collenchyma cells were subsequently cut off by new phellogen.

This type of bark and twig injury of pear trees was apparently also
studied histologically by Miczynski.*® He shows the distribution of
dead and discolored tissues in a colored plate. In cases in which the
cambium had been killed, the new cambium developed in the inner
phloem much like that described in a former section of this paper.
Another good but general account of bark injuries of fruit trees is
given by Oberdieck.*® He gives many clear details regarding numerous
cases. :

Sun-scorch is usually confined to trees that have not yet reached
the rough-bark age, and consist of dead and discolored bark on the
trunk or main branches, usually (though not always) on the west or
southwest sides. Histologically its early stages are similar to those
giving rise to the premature bark-roughening described by Sorauer.
In many cases of the sun-scorch type, however, only the outermost
collenchyma cells are involved, and consequently the resulting new
bark surface looks only slightly frayed. Numerous interesting obser-
vations have been made on this bark disease, and in many of the
discussions one may find pertinent suggestions. Hess,*! for instance,
notes that this trouble develops on smooth-barked forest trees one or

38 Sorauer, P., Uber Frostschorf an Apfel- und Birnstammen, Zeit. Pflanzenkr

Bet 7) 189 ke
39 Miczynski, K., Ueber das Erfrieren der Gewebe des Birnbaums, Bot. Centralbl.
48: 228. 1891.

40 Oberdieck, J. G. C., Beobachtungen tiber Erfrieren vieler Gewachse und
namentlich unserer Obstbaume in kalten Wintern; nebst Erérterung der Mittel
durch welche Frostschaden méglichst verhtitet werden kann, pp. 108. Ravensburg.
1872.

41 Hess, R., Der Forstschutz. Leipzig. 1878.

504 J. G. GROSSENBACHER

more years after the forest has been thinned. Hartig® gives some
interesting data along this line and concludes that the injury results
from the contraction and expansion of the bark rather than from the
heating of the sun as is maintained by many, because he found frequent
cases of it on north slopes and on the east and north sides of trunks.
Many of the sections used in the present histological study of
crown-rot were made from the bases of shoots arising from large
branches of apple trees that had been pruned rather severely, and
they therefore also represent the initial injuries preceding the develop-
ment of crotch cankers as more fully discussed on pages 40-42 of
my paper written in 1912. Goethe published a paper in 1877, in which
he announced the conclusion that cankers are due to low-temperature
injury of the bark. When it was pointed out to him that in Italy
where the winters are mild cankers are equally prevalent, he reinvesti-
gated*® the matter and revised his conclusions to the effect that many
of the cankers are due to fungus parasites. It should be noted,
however, that his revised conclusion was based largely on the fact
that in the spring of 1878 he found new cankers even though no date
frosts had occurred. (The notion that only late frosts cause these
injuries has led many astray.) Fungi developed on cankers when
placed in moist chambers, but when spores were used on uninjured
bark no cankers resulted. In the following winter bark injuries were
numerous in crotches and other places where cankers usually occur.
Many of the wounds were carefully cut out in April and most of them
healed rapidly, although in a few instances the branches involved died.
Some years later, also, Goethe** made an extended study of winter-
injuries, giving particular attention to the aftermath or the results of
such injuries. A drop in mid October to — 2.5° C. and one to — 10°C.
in November very severely injured the pith and other tissues in
shoots and the bark of trunks just above the ground. High-headed
trees were found more subject to trunk injury than low-headed ones.
This is in agreement with what I found in western New York (Tech.
Bull. 23, pp. 18-20). Goethe described interesting cases in which
“ Hartig, R., Ueber den Sonnebrand oder die Sonnenrisse der Waldbaume,
Untersuch. Forstbot. Inst. Miinchen 1: 141. 1880.
“8 Goethe, R., Mittheilungen iiber den Krebs der Apfelbaume. Leipzig. 1877.
, Weitere Mittheilungen itiber den Krebs der Apfelbaume, Landw. Jahrb.
9: 837. 1880.

44 Goethe, R., Die Frostschaden der Obstbaume und ihre Verhiitung, Nach den
Erfahrungen des Winters 1879-80, dargestellt. Berlin. 1883.

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 505

the different buds and nodes on the same branch varied greatly in
their susceptibility to injury; some remaining normal while others
were entirely killed. :

Miiller-Diemits and Stormer-Halle* found that fruit trees are most
subject to bark diseases at the age when they first become profitable.
The bark at the crown, crotches, and various other places on trunks
and branches, according to these authors, dies; fungi and bacteria
enter the wounds and induce further injury and decay. The wood
becomes discolored, and the branch or tree involved dies.

As an illustration of the especial susceptibility of trees to bark
injuries and the resulting diseases, during certain stages in their life
history, especial attention may be called to the bark-roughening
discussed above as well as to this paper by Miiller-Diemitz and
Stormer-Halle. The older literature of forestry contains many items
of interest in this connection. Graebner,“ for example, described a
case of this kind, and in the writings of Hartig, Nordlinger, Hess, and
others, are to be found many further instances. Graebner found that
a high percentage of trees in a spruce forest had sustained bark injury
on their trunks. Very many of them died of crown-rot. The trees
had apparently been from 34 to 57 years old at the time the injuries
occurred. In an adjoining spruce forest, where the trees were under
20 years of age, no bark injury could be found. It appears from
various published statements that spruce trees in a forest stand usually
become rough-barked between the ages of 30 and 4o, depending upon
the rate of growth and thickness of stand.

CONCLUSIONS

The histological study here briefly reported, in connection with
my two former papers, throws enough light on the earlier stages of
crown-rot to permit more definite and general statements regarding
its development. It is shown that this and some related bark diseases
are not due primarily to the organisms usually found in such affected
bark in summer, but to injuries arising when adverse environmental
conditions overtake trees having immature bark in certain regions.
The rotting of the dead or dying bark is due chiefly to fungi which in

45 Miller-Diemitz, J., und Stormer-Halle, K., Das Obsthaumsterben, Deutsch.
Obstbauzeit. 56: 81. I9gI0.

46 Graebner, P., Beitrage zur Kenntnis nichtparasitarer Pflanzenkrankheiten
an forstlichen Gewachsen, Zeit. Forst. Jagdwesen. 38: 705. 1906.

506 J. G. GROSSENBACHER

some cases also kill living portions while vegetating in severely injured
bark.

On Plate XXI are shown some of the main types of injuries often
found in bark after unseasonably severe periods. This material was
collected before evident growth started in the spring and therefore
gives some idea of the actual distribution of the injuries. An examina-
tion of these figures makes it appear that injuries are of two types:
in Figs. 1, 2, 4 and 8 they are evidenced chiefly by a discoloration and
collapse of the affected tissues, whereas in Fig. 3 the injury consists
mainly of a tangential rupture with only a few of the groups of dis-
colored cells; in Figs. 5, 6, 7 and 9 there occur combinations of the
two types of injury. In the latter cases the tissues along the margins
of the ruptures are discolored and collapsed much as they are in Figs.
tand 2. Unfortunately the sections of the material having a combina-
tion of the types of injuries shown in Figs. 1 and 2, or of those in Figs.
2 and 7, turned out to be such poor preparations that no use could
be made of them.

Plate XXII shows comparable cases as they appeared about two
weeks later. This represents a stage of regeneration growth during
which living parenchyma cells surrounding injured or dead regions are
actively proliferating into spaces formerly filled by the shriveling
masses, and into gaps occasioned by ruptures. Figs. 10 and 12 are
especially interesting because dead tissues are compressed into more
or less radially arranged plates. The proliferating cells are seen to
penetrate many of the dead masses, and apparently make contact
with living cells beyond. This rapid early regeneration-growth in
injured bark is responsible for the fact that so few injured places
result in patches of dead bark.

Practically the final alignment of injured and living tissues, as
well as the locations of the new meristematic layers, is shown on Plates
XXIII and XXIV. From these figures it is evident that when
enough of the cambium and inner phloem are killed to form a fairly
thick dead layer only a few, or in some cases over considerable areas
no living connections are re-established between the old wood and the
bark. In some instances the most severely affected bark died early
(Figs. 24, 25 and 26), while in others (Figs. 18, 21, 22, 28, 31 and 32)
regeneration went on rapidly and the formation of a considerable
amount of new wood and bark resulted (Figs. 18, 22 and 31). There
are some in which the outer bark has remained alive but in which very

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 507

little effective regeneration has resulted (Figs. 23, 30, 32 and 35).
In most of the last type of cases the bark died to the margin of the
severe injury before mid-summer, so that they appeared like Figs.
24 and 26. In instances like those of Figs. 21, 30 and 31 the results
were sometimes most striking, because often a fair beginning had been
made on the new growth of wood when suddenly the bark died over
large areas. In case such an injury occurred on a small branch or
shoot, it usually died outright as shown in Fig. 39a.

The transition from stages like those shown in Figs. 23, 32 and
35 to those of Figs. 24, 25, 26, 41, 46, 47 and 51 seems fairly clear.
The associated micro-organisms evidently play an important rdéle in
the decay or disintegration marking the later stages of these bark
diseases, and in some types doubtless extend the injured areas by
their vegetative activities in the places initially killed (Figs. 27, 30,
33 and 35). Yet, this does not seem to be generally the case. The
tree shown in Fig. 50 had lived at least fourteen years after the occur-
rence of the injury, eventually resulting in ‘“‘heart rot.’’ It is evident
from this figure that the fungus rotting the wood present at the time
the initial injury occurred has not progressed far outside the last
layers of injured wood. In fact, it appears as though it may have
rotted only as much of the wood formed since the occurrence of the
injury as had been discolored by the diffusion of disintegration
products from the initially killed cells. This diffusion injury is shown
in Figs. 39 and 42. Yet, in looking over the figures of Plate XXIV
it becomes obvious that once a wound parasite, or even a saprophyte
especially adapted to a particular host, gains entrance to such an
admixture of dead and regenerated living tissues, some living portions
may be killed as a result of the vegetative activities of the fungus.

This record of low-temperature injuries occurring in the bark of
fruit trees, and of their subsequent development into bark diseases,
is of interest and value independent of the factors that give rise to the
initial injuries. The diseases in question are thus traced so much
nearer to their first causes. Both the macroscopic and the micro-
scopic appearances of much of the bark affected indicate that excessive
tensions are developed during the occurrence of the injuries. It
remains to be determined whether or not the tension-injury hypothesis
of Sorauer is applicable to low-temperature injury in general, in con-
nection with the physiological disturbances induced by the occurrence
of severe weather while some of the bark tissues are in certain stages

508 J. G. GROSSENBACHER

of immaturity or arrested development. It seems possible that, at
least in some cases, the presence in the bark of metabolized foods of
insufficient concentration to allow normal growth and maturation
is the most significant phase of immaturity; the occurrence of droughts
appears to have a significant relation. An adverse period in the
environment occurring at such a time stops the further accumulation
of the labile components of protoplasm, and a long retention of these
elementary constituents, together with the enzymes present, may
lead to catabolic processes that eventually result in the death of the
tissues involved.

s

ACKNOWLEDGMENTS

In this connection it is a pleasure to acknowledge the aid received
from Professors J. B. Overton, L. R. Jones, and E. M. Gilbert, of the
University of Wisconsin, during the winter of 1911-12, and especially
that of Professor C. E. Allen, of the same institution, for the many
excellent suggestions he made regarding the revision of this paper.

DESCRIPTION OF PLATES XXI-XXVII

PEATE Sock

Sections of material collected before growth started in the spring, showing types
of initial injuries. All from apple excepting Figs. 1 and 2, which are from pear.

Fic. 1. Injury mainly in cambium, phellogen, and phloem regions; scattered
cell-groups in xylem and cortex affected, as indicated by discolored places.

Fic. 2. Another section with severe initial injuries confined chiefly to inner
cortex and outer phloem; phloem rays and cambial zone also injured. Scattered
groups of dead cells in other parts of cortex and phloem.

Fic. 3. Section of apple branch with common type of injury not usually
accompanied by much discoloration, consisting principally of a rupture in the
inner phloem and only secondarily of groups of dead, discolored cells.

Fic. 4. Section of apple with most of the injury in the inner phloem, as indi-
cated by discolored streaks and masses. On the left the cambium is killed; on the
right it is alive and apparently normal.

Fic. 5. A condition comparable with that of Fig. 4, excepting that a portion
of the injury consists of ruptures as shown near the right.

Fic. 6. Some living cambium on the right, and a zone of severe injury in the
inner phloem above; on the left most of the cambium is killed. The rays are more
severely affected than in the section shown in Fig. 5.

Fic. 7. Collapsed tissue in the inner phloem and cambium is interspersed
with a few living cells. Phloem rays are dead, and scattered groups of dead cells
occur in the older portions of the phloem.

Fic. 8. Some groups of dead cells and some ruptures occur in the cambium;
the inner phloem has but few affected cells. The middle and outer phloem are
considerably injured.

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 509

Fic. 9. Much like Fig. 7, but with larger groups of living cells in the cambium;
more definite radial ruptures are evident in the inner phloem. On the right all of
the phloem and cambium between the clefts and the wood are dead.

PLATE XXTI]

Displacement of initially killed tissues by regeneration-growth. Collected two
weeks after the material used for Plate XXI; all from apple.

Fic. 10. A development from an injury like that shown in Figs. 6 and 7. All
living cells have grown and compressed the dead material. Living cells proliferate
outgrowths into the dead masses. Regeneration tissue is all parenchymatous;
it divides in all planes.

Fic. 11. Similar to Fig. 10; in neither of these instances are living connections
evident between the bark and wood.

Fic. 12. Different from the two preceding chiefly because some of the cambium
has survived. Even where the cambium is killed, frequent living cells are found to
bridge the gap. Regeneration has repaired many holes that resulted from the
shrinking of groups of dead cells.

Frc. 13. Seems to be a later stage of an initial injury of the type show in
Fig. 6; both at the right, where the cambium is involved, and at the left, where the
inner phloem is affected, proliferating cells connect the wood with the bark through

mantle of dead tissue.

Fic. 14. Though the old cambium (0c) was only slightly injured, a new one
has begun to develop (cc) in the regeneration tissue of the inner phloem. Initial
injury in the older phloem was very severe, yet the living portions of it are compress-
ing the dead masses by growth.

Fic. 15. From the type shown in Fig. 8; new cambium (cc) is forming in re-
generated tissue of the inner phloem, and the former cambial line is discolored.
The outer phloem and cortex are almost wholly isolated from the inner bark by a
ragged layer of dead tissue. Occasional living cells occur in the dead mass.

Fic. 16. Derived from a portion that was less injured than that shown in Fig.
6. Living tissues dominate, and new cambium (cc) has become established.

Fic. 17. Much like Fig. 15, excepting that the injury is much more severe in
the old cambium.

PEATE) XXII

Final disposition of much of the initially killed tissue. Collected a month after
that shown in Plate XXII; all from apple.

Fic. 18. Late stage of a type shown in Figs. 7 and 9. Former position of the
cambium (oc) and its present location (mc). A considerable layer of new wood
(nw) has developed, and a new phloem (zp) is also noticeable. In the phloem
the dead masses are localized near the boundary between the old and the new phloem,
some extending into the new phloem toward the cambium just as dead streaks
extend into the new wood from the injury outside of the old wood (ow).

Fic. 19. From the middle toward the right of the figure the new cambium (nc)
is only faintly indicated. Injury at the old cambium (oc) is not as conspicuous as
in Fig. 18, but it is of wider extent. Initial injury occurred throughout the old
phloem and was very severe.

510 J. G. GROSSENBACHER

Fic. 20. Mantle of greatest injury with irregular course. At the right the
new cambium (vc) has begun the development of a new layer of wood (nw), while
at the left no substitute cambium is yet visible.

Fic. 21. More irregularity in the course of the mantle of greatest initial injury.
Living phloem is left attached to the old wood (ow) and converted into wood without
leaving an active cambium. New cambium (nc) has developed outside the zone
of injury that produced the layer of wood. The cambium has become abnormal,
yellowish in color, and is partly disorganized. Discolored streaks extend from it
into both new wood (mw) and new phloem (np).

Fic, 22. Much like Fig. 18, showing in addition a new phellogen (ph) cutting
off the outer part of the cortex.

Fic. 23. Exaggerated form shown near the left end of Fig. 20. No new
cambium is in evidence, and dead matter predominates, although from the outside
the bark appeared normal.

Fic. 24. A case in which the most severely injured portions of the bark died
and a callus (cal) developed along its margin. New wood (zw) at the lower right
arose without leaving cambium.

Fic. 25. Severe initial injury confined to a small space. Although much
isolated from the old wood, the callus is normal, having an active cambium. As in
Fig. 24, fungus mycelium is present in the dead bark and in the dead mantle between
the callus and the dead wood.

Fic. 26. Bark half-way around stem is dead and sunken, much like the patch
shown in Fig. 41. In cross-section this looked like a miniature of the specimen
shown in Fig. 50. Mycelium and pycnidia of a bark fungus were present.

PLATE XXIV

Higher power views of some injured tissues in the stage shown on Plate XXIII.

Fic. 27. View of a region like that in oc of Fig. 18. Living connections
through the dead region are evidently few and imperfect.

Fic. 28. Like Fig. 27. Former phloem rays have become discontinued and
have undergone division and become converted into callus tissue.

Fic. 29. Copied from Sorauer’s paper on ‘‘Frostschorf’’ of apple and pear in
Zeitschrift fiir Pflanzenkrankhetten, 1: 137-45. 1891. Cortical injury that usually
precedes premature bark-roughening.

Fic. 30. Magnified view of the type shown in the center and right of Fig. 19;
substitute cambium (nc) developed in a meandering course. Much-injured bark
practically isolated from the old wood (ow).

Fic. 31. Detail of a case something like that shown in Fig. 22, excepting that
practically no new phloem has yet developed; new cambium (nc) is considerably
disorganized and discolored. Old phloem (0p) is permeated by initially killed
tissue, in direct contact with disintegrating new cambium.

Fic. 32. Much like Fig. 30; only a faint indication of substitute cambium (nc)
isin evidence. Injury in old phloem is more severe.than in Fig. 30; cells in regenera-
tion-growth are less affected by pressure than those in Fig. 30.

Fic. 33. Higher power view of case like that in left-hand portion of Fig. 20,
with uncommonly thick mantle of dead tissue. Living portions of former bark
rays are converted into ordinary parenchyma.

CROWN-ROT OF FRUIT TREES: HISTOLOGICAL STUDIES 511

Fic. 34. Cross-section of a large dead streak of phloem surrounded by modified
irregular parenchyma.
Fic. 35. Similar view of a living streak in the phloem surrounded by layers
of dead, collapsed cells.
PEATE XOXY

Initial injuries followed by another type of regeneration. Atl from maple
except Fig. 41, which is from pear.

Fic. 36. Series of magnified views of a portion of Fig. 37, from old bark (0b)
to old wood (ow): a, old bark (0b) run through by rifts, new bark (vd 1) with inclu-
sions of dead masses, and new cambium (vc I) just outside some new wood shown at
the outer edge of outer new wood (nw I) in Fig. 37; 6, from the inner edge of the
outer new wood (nw 1) of Fig. 37, showing the new cambium (nc 2) and some very
irregular new bark (nb 2); d, from the outer edge of the inner sheath of new wood
(nw 2) of Fig. 37, nb 2 and nb 3 together constituting the compressed new bark
between nw I and nw 2 of Fig. 37. e, higher power view of line oc of Fig. 37, showing
some detail.

Fic. 37. Cross-section of maple tree (Acer platanoides) with a season’s growth
added after the occurrence of the initial injury, that had been similar to that shown
in Fig. 3, and somewhat like that shown in Fig. 38. Three cambial layers have
developed in place of one. The tangential cleft left some living phloem adhering to
the old wood like that shown in Fig. 3. Substitute cambium arose in the strip of
inner phloem adhering to the wood, giving rise to c 3 of Fig. 37; then along the
inner surface of the loosened outer bark another cambium developed which gave
rise to new wood in its middle and was thus divided into two cambial sheaths (c I
and c¢ 2), each producing wood and bark. Activity of three cambial layers, as
detailed in Fig. 36, gives rise to unsightly enlargements like that shown in Fig. 4o.

Fic. 38. Section of box-elder tree (Acer Negundo) with portion of its bark
separated from the wood, though still living. Beginning of callus formation is
shown along the edges of the loose bark (May 28).

Fic. 39. Higher power view of a portion shown in Fig. 26e; considerable regen-
eration-growth of wood outside the zone of iaitial injury, which subsequently died
and became discolored.

Fic. 40. Trunk of a street tree (Acer platanoides) unduly enlarged near the
upper part of the trunk owing to the activity of three cambial zones developed after
the occurrence of some injuries initially much like those shown in Fig. 3.

Fic. 41. Trunk of a smooth-barked pear tree in early summer, with a sunken
patch over the places sustaining most severe internal injuries.

Fic. 42. Detail view of a section taken across the faintest portion of the line
oc in Fig. 37, showing that normal new wood (nw), arising outside such a line of
initial injury, may subsequently be killed and discolored.

PLATE XXVI

Collection of bark-injured and crown-rotted stems, in which the injury was
accompanied by radial clefts. All of apple except Fig. 48, which is of orange.

Fic. 43. Apple tree with nearly complete girdle of loose bark (one patch
opposite) and a radial cleft 17 cm. long. —

Fic. 44. Shows the extent of the loose bark of Fig. 43.

512 J. G. GROSSENBACHER

Fic. 45. Twelve-year-old apple tree which had a complete girdle of loose
bark from the ground up to the main branches. A radial cleft 25 cm. long occurred
in it near ground. .

Fic. 46. Apple tree with complete girdle of dead bark; thick callus along its
upper edge.

Fic. 47. Atypical case of crown-rot on apple.

Fic. 48. Stem of orange tree showing radial clefts in loose bark. Initial
injury occurred on the night of November 20, 1914, when the temperature sank to
a little below — 2° C. In the summer of 1915 many trees affected in this manner
died with symptoms of “‘ withertip.”

PLATE XS XVET

Crown-rot and other troubles of large trees.

Fic. 49. Section from near the base of a large apple-tree trunk (28 cm. in
diameter), showing a line of initial injury that occurred some fourteen years before
cutting; also showing that the bark sustained a radial cleft (upper side).

Fic. 50. Section of another tree of the same size and from the same orchard
as that shown in Fig. 49. The initial injury occurred in the same year as that in
Fig. 49. The wood cylinder subsequently died and rotted, and some of the wood
produced by the new cambium also decayed.

Fic. 51. Large apple tree with complete crown-rot girdle. Upper roots died,
but those under the center of the tree were alive.

Fic. 52. Section of a spruce stem, copied from Hartig (Untersuch. Forstbot.
Inst. Miinchen 1: 147. 1880). Included here to show that the initial injury from
which the trouble developed occurred during the dormant season and not during the
growing season as was maintained by Hartig.

VOLUME IV, PLATE XXI.

AMERICAN JOURNAL OF BOTANY.

ES.

-ROT OF FRUIT TRE

CROWN

SSENBACHER

GRO

a

—_

oS

VOLUME IV, PLATE XXII.

AMERICAN JOURNAL OF BOTANY.

GROSSENBACHER: CROWN-ROT OF FRUIT TREES,

VOLUME IV, PLATE XXIII.

CROWN-ROT OF FRUIT TREES.

GROSSENBACHER

AMERICAN JOURNAL OF BOTANY.

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE XXIV.

GROSSENBACHER: CROWN-ROT OF FRUIT TREES.

— +e, =e.

a of
~ LW,
a. .
© sy.
Fall i
a
;
:
‘\
S
}
—_—_—_— =
=

>

6

t
oi
* i
ety

nts a

le
i
Uc
-
cy ite
.
,
f
ol
4 3
a
“ .
+
x
4 :
ashlee nde cen Ge sare iV nied arnt REN WAC =
a. 7 a
> . é
cet 6 a enda ds Ae) Oe ess hore,
. =

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE XXV.

GROSSENBACHER: CROWN-ROT OF FRUIT TREES.

nD

¢

“SHHUL LINAA AO LOA-NMOU) *-AHHOVANASSONUL)

"IAXX 3Lv1d ‘Al 3WNI0A “ANV.LOG 4O IVNYNOPF NVOINSNY

a

wos

VoLuMmeE IV, PLATE XXVII.

i Sas
ea 4NSI—N

Sh

41 76 75

CROWN-ROT OF FRUIT TREES

GROSSENBACHER

AMERICAN JOURNAL OF BOTANY.

ee
Mi ay re

“Devorep TO > An Brancnrs or Botamteas ‘Scune! E :

aM

ms Ae, \
: eae een feroaat eau Business ete De
ee | Brooklyn. BULOnt: Garden’ Baa wien 4
tae a “Irvine W. Balke eS hee Pais) Sele
BENG oi Rs WRC RT Bussey: Institution Ae ERO OD Aa aa) a
oo Wa ie nen eo H. BARTLETT, ieee Bi Saige
awk, : University of Michigan sehiesan Pavan , i
NG ie gut cA Rae oy ee are. Jounson, University of Wielka n oy
ARN i a iso abtes _ CReprerenting American Phytopathological Secety) Ha
he te patie “The eaaat is publisnen, Jaen: cmeat hiine’ ‘Ageue ‘aa er a
aa ie teas price, $5.00.a year. Single copies 60 cents, plus. postage. Postage will
Pe ets ‘\be charged to all: foreign countries, except’ “Mexico, Cuba,” ‘Ri ne
SNe Semana Sr) eB Zone,’ Republic of Panama; ‘Hawaii, . Philippine telands, pe hoe n
‘ ‘Islands, and Shanghai. Postage to Canada 20 cents a volume on. annual peat
“tions; to all other SOLE IES in: a the one nom a nents a votume: on pe al suk
: eer ‘eae . bist RS Seen Gal sy ne Lead : ey
ag i id n all, bs
a a a te iets Gh BIA
ie ily Wiha * R t ES Xe j META ES hg fore age aie
Ae People! should. ibe oben Coed pen es on Peceige! ines
‘) | fosrel of Botany: peck Botanic en 1 een N. ee "A sa Ne

: ‘cents must be added to all chen not. drawn on b New York chy Sake eG yi
hl De: “<,\ Claims for missing numbers should be made. within 30. days: following ad
». > / “date of mailing.’ The publishers wll aupply missing numbers: free only: wher ae

Bk - eve been lost in the mails.” |
ia nae ee _ Correspondence concerning editorial otntiaun’ should be addressed to. Prof. ‘F,
ni lc Newcombe, 9 Geddes Heights, Ann Arbor, Mich. RDM gy Bn I Sie or

2 Fa - : Ns « Business correspondence, fncliitling: notice of siase of f aditvess, and directions iy yee
a Rane ee Mentarains reprints, should be addressed to American Journal of Botany, Brook-; |”
Ree Fhe eke rome Bottini: ie ae Brchye, N. ag or Ae North Cub basses) Lanca <

AMERICAN
JOURNAL OF BOTANY

ed

No. 9

VoL. IV NOVEMBER, IQI7

BEPECT OF SOIL TEMPERATURE ON THE GROWTH OF
BEAN PLANTS AND ON THEIR SUSCEPTI-
BherY TOA. ROOT PARASITE

DONALD REDDICK

The general opinion prevails that temperature plays an important
role in the infection of a host by a fungous parasite. The experi-
mental data showing just what this réle is, however, are very meager.
In the case of infection of aerial parts other factors are interrelated
with temperature, such as persistence of moisture for spore germina-
tion, rapidity of germination of spores, and so forth, but in the case
of infection of roots by organisms persisting in the soil these conditions
ordinarily do not enter. Apparently the soil-inhabiting parasites are
largely capable of saprophytic existence so that, given the requisite
amount of soil moisture to maintain plant development, the parasite
is able to grow and reach the roots of a susceptible host. Gilman! has
recorded observations on the relation of infection by Fusarium con-
glutinans Wr. on cabbage to soil temperature conditions and thinks a
high soil temperature favorable to infection. Gilman! continued this
work with F. conglutinans and appears to have established the point
just mentioned, although the control of conditions in some of his ex-
periments was not all that might be wished for. Tisdale? arrives at
similar conclusions in connection with the infection of flax (Linum
usttatissimum) by Fusarium Lini Bolley and states that the low critical
temperature is about 15°—-16° C.

Gilman, J.C. The relation of temperature to the infection of cabbage by
Fusarium conglutinans Wollenw. (Abstract.) Phytopathology 4: 404. 1914.
Cabbage yellows and the relation of temperature to its occurrence. Ann. Mo.
Bot. Gard. 3: 25-82. 1916.

? Tisdale, W. H. Relation of temperature to the growth and infecting power
of Fusarium Lint. Phytopathology 7: 356-360. 1917.

[The Journal for October (4: 439-512) was issued October 2, 1917.|
os

annson 1a Ng]

‘C

«

flonal Ni 1} 2 e\) :

514 DONALD REDDICK

The writer undertook an investigation of somewhat similar nature,
using the bean, Phaseolus vulgaris, as host and Fusarium martu phaseoli?
for the parasite, a fungus that has been shown by Burkholder? to be
the cause of a serious disease of beans in New York.

The work was performed in the laboratory of plant physiology of
the Johns Hopkins University, Baltimore, Maryland, where the writer
was fellow by courtesy during the academic year 1916-17. His
thanks are due the authorities of that institution for the facilities af-
forded him and special acknowledgment is made of the critical advice
given him by Prof. B. E. Livingston, under whose immediate guidance .
the work was done. While the investigations are by no means com-
pleted, some of the physiological features of the results thus far ob-
tained are of sufficient interest to warrant this note.

_ The plants were grown in cylindrical vessels of tinned sheet-iron,
17 cm. tall and 15 cm. in diameter, which in turn were placed in a
water bath. The garden soil used was first heated in an autoclave
for one hour at a temperature of 110° C. and it was then made uniform
by repeated sifting. The culture vessels were filled and nearly uni-
form packing was obtained by letting the soil fall into place always
from the same height. Water was supplied by means of the Livingston
auto-irrigator,* two cylindrical porous clay cups being used, each with
an exposure to the surrounding soil of approximately 121 square
centimeters.

The irrigation water was drawn directly from the water of the bath
and care was taken to have the supply uniform, so as to avoid difference
in soil moisture content that might influence the growth of the plants.

Since the water level was nearly as high outside the cylinders as was
the level of the soil within, it was necessary, while the plants were

3 Burkholder, W. H. Some root diseases of the bean. (Abstract;) .Phy-
topathology 6: 104. 1916. Bean diseases in New York State in 1916. (Ab-
stract.) Phytopathology 7: 61. 1617.

Burkholder states that Fusarium martit Ap. & Wr. does not produce infection
on the bean but that the fungus from bean is nearly identical with this species.
The name martii phaseoli has not been used previously and is only introduced here

as a matter of convenience.
_4 Livingston, B. E. A method for controlling plant moisture. Plant World
Il: 39-40. “1908.
Hawkins, Lon A. The porous clay cup for automatic watering of plants.
Plant World 13: 220-227. IgI0.
Livingston, B. E., and Hawkins, Lon A. The water-relation between plant and
soil. Carnegie Inst. Wash. Publ. 204: 3-48. 1915.

EFFECT OF SOIL TEMPERATURE ON BEAN PLANTS BLS

small, to retard the flow of water into the cups. This was accom-
plished by introducing mercury columns of equal heights into all the
supply tubes. Later the mercury was not needed and was removed.

The water baths employed were three in number, each 60 cm. in
diameter and 25 cm. deep (ordinary galvanized iron laundry tubs)
thus giving space for seven culture vessels each. A wood grating at
the bottom supported the culture vessels and allowed them to be sub-
merged to within one centimeter of che top.

Three temperatures, 34°, 22°, and 15° C., were arbitrarily decided
upon, but this choice was governed somewhat by the facilities avail-
able. The highest temperature was obtained by means of an electric
heater under thermostatic control, and was maintained uniformly
throughout the course of the experiment.

The medium temperature followed that of the culture room; there
was no special control in this bath. Because of the great bulk of
water, the fluctuation in temperature was not very great. The range
was from 20° to 23° C. (usually 21° to 22°) whereas the diurnal varia-
tion in the temperature of the greenhouse room was large, 12° to 28°C.
In our present state of knowledge of the influence of soil temperature
on host or parasite this fluctuation is to be regarded as of little con-
sequence but obviously some constant temperature might have been
maintained with very little difficulty, by employing such an outfit as
was used for the highest temperature.

A constant water level was maintained in each of the two warmer
baths by means of a Mariotte flask.

The lowest temperature was obtained by passing a continuous
stream of tap.water through the bath. When the water flowed at the
rate of 1,500 cc. per minute a temperature of 15° was maintained,
during the winter months. On very warm days a rise of two or three
degrees sometimes occurred. ‘The total range was from 14° to 18°.

The surface of the water in the baths was covered with a thick
paraffin oil to reduce loss of heat by evaporation and to eliminate the
vapor blanket chat would otherwise have been present over such an
exposed body of water. Later it was found that a covering of ordi-
nary paraffin (melting point about 50°) was very much better for the
purpose. This was melted and poured on the water, where it was
allowed to spread and harden.

The soil in four of the culture vessels of each series was contami-
nated by sprinkling in it, when nearly full, some soil heavily laden with

51 6 DONALD REDDICK

viable spores and mycelium of Fusarium marti phaseoli from culture.
The fungus, which was supplied through the courtesy of Dr. W. H.
Burkholder, had been maintained for several months in pure culture
but the medium (bean-pod decoction agar) was uniform throughout
the period, and a sub-culture had been made every ten days.

Pure-line seeds of a pea bean,® were disinfected externally with a
I to 1,000 solution of mercuric chlorid, after which they were sprouted
in a moist chamber. They were planted on January 10, 1917, six®
seeds in each culture.

When the cotyledons had broken through the ground all plants
were inoculated with B. radicicola by injecting into the soil about the
roots one cubic centimeter of a heavily laden water suspension of this
organism taken from bean nodules, and the number of plants per pot
was reduced to four.

After twelve days the plants with soil temperature at 34° were
developing the first trifoliate leaf; those at 22° had just spread the
first pair of true leaves and those at 15° were not all through the soil
surface. On the forty-fifth day the plants at 34° were beginning to
blossom while those at 22° began blossoming eleven days later. The
plants at 15° were either dead or very poor and none developed satis-
factorily. A single one of these cold-soil plants finally reached a
height of about 15 cm. and produced one blossom but did not set a pod.
It is to be borne in mind that the air temperature here was practically
the same as that of the plants with soil temperature of 22°.

Unfortunately some of the plants in the control cultures became
infected, the contamination apparently being carried by numerous
small insects that were abundant on the plants. In the cultures at
22° five of the twelve plants were diseased and in those at 34° eight of
the twelve plants were affected. All of these plants were affected
relatively late as compared with the inoculated plants, so that it is
impossible to judge what amount of damage may be attributed to the

> The seed was supplied through the courtesy of the Department of Plant Breed-
ing, Cornell University and is maintained under the department number 1986-2.

6 From the outcome of this experiment and numerous others subsequently
performed with beans of this and another pure line, and with beans secured on the
open market, it is very evident that not enough seeds were used at the outset. After
seeds of uniform size and appearance are selected it is safe to allow for only about 25
percent as likely to yield plants entirely free from defect and of perfectly uniform
appearance. Weak plants frequently cannot be detected for ten days or two weeks
after the plants emerge from the soil.

EFFECT OF SOIL TEMPERATURE ON BEAN PLANTS Sl7

disease, but it was obvious at harvest time that the plants in two of the
control cultures of the series of 24° were severely injured.

In addition there were ‘‘weak’’ plants in nearly every culture.
These could not be detected as such for two weeks or more after the
plants were up and it was then too late to correct for the trouble. In
fact it was thought for some time that some of these plants were ones
on which infection had been particularly severe. As there were plants
of varying degrees of “‘weakness’’ it is not possible to throw the poor
plants out of consideration.

Furthermore, it is not possible to make a comparison between the
cultures grown at the two temperatures because of the fact that the
plants grown at high temperature developed more rapidly from the
very beginning and thus matured under a different set of air condi-
tions. In this experiment this meant that the plants grown at the
highest temperature had very much less sunshine than those grown at
22°. The difference is noticeable in part in the total dry weight of
seed, but some of the difference is attributable to a more severe in-
fection on control plants grown at the high temperature.

Finally the difference in growth at the two different temperatures
might have been due in part to a difference in air temperature. Ther-
mometers suspended over the water baths at a distance of 15 cm. from
the surface showed constantly a higher temperature over the bath at
34° than over the ones at 22° and 15°. The difference varied from
.5° to 4.5° and averaged from 3° to 4° higher.

With these four considerations in mind it may now be stated that
the average yield per plant for ‘‘healthy’’ plants in the series at 34°
was I.451 grams of air-dry seed. For the infected plants the average
was 1.081 grams. ‘Thus the presence of this Fusarium on the roots of
beans under the conditions stated resulted in a direct loss of 25.5 per-
cent. For the cultures grown with a soil temperature of 22° the aver-
age yield per healthy plant was 2.361 grams. For the inoculated
plants it was 1.557 grams. Here the reduction in yield on account of
disease was 34 percent.

The most interesting feature of the experiment is the fact that
these beans grew faster and matured a crop earlier with the higher soil
temperature. The relatively small difference in air temperature may
account for some of this difference in growth but certainly cannot
entirely account for the results obtained. Wholly aside from its scien-
tific interest the question may have an important practical bearing for

518 DONALD REDDICK

those engaged in the production of flowers and vegetables under glass,
and from either standpoint is worthy of further attention.

The idea of supplying bottom heat has been used extensively by
florists’ for starting cuttings, but not for growing crops. Plant phys-
iologists do not seem to have studied the problem, judging by the
absence of literature on the subject, but this experiment with beans
and some trials with radish (Raphanus sativus), cucumber (Cucumis
sativus), and tomato (Lycopersicum esculentum) indicate that root
temperature and foliage temperature are readily separable as condi-
tions influencing the growth of plants.

With respect to susceptibility due to environmental changes it
would seem that in the case of temperature as applied in this experi-
ment the relation between host and parasite cannot be analyzed
readily. The experiments show that the host is influenced markedly
by a change in soil temperature so that it is impossible to make a
direct comparison of various temperature conditions because of the
slow action of the parasite. If the parasite made a rapid invasion and
killed the host outright within a few days there would be an oppor-
tunity to grow all plants under identical conditions until the day of
inoculation but even then the sudden change of soil temperature might
have an even more marked effect on the physiological condition of the
host, perhaps changing its susceptibility in a very pronounced manner.
In the case of this disease, and of the majority of root diseases, prompt
death of the host does not follow because some water continues to
enter even after the roots have been killed and especially because on
most plants new roots generally push forth above the point of infection.

In will be necessary to study under controlled conditions the be-
havior of the uninfected host when subjected to certain changes in
this one environmental condition, and that of the parasite in the same
way, in order to determine the true relation of host and parasite. |
This involves the control of all the known conditions affecting the
growth of plants, including light, a method for doing which has only
been hinted :at® to: date:

The physiology of the fungus here used has not been the subject of
investigation as yet, but in some preliminary experiments on the rate
of growth of the fungus at different temperatures it was found that
the diameters of the thalli on bean broth agar in petri dishes varied

7 White, E. A. The principles of floriculture. p. 162-164. New York. I915.
® Eivingston, B. LE. Plant World 20: 11,4! 1917.

EFFECT OF SOIL TEMPERATURE ON.BEAN PLANTS 519

with the temperature. In one instance, at the end of five days, the
diameters of the thalli in millimeters for the stated temperatures are
shown in the accompanying table.

TABLE I

Diameters in Millimeters of Thallt of Fusarium marti phaseolt when Grown for Five
Days at the Temperatures Indicated

Diameter of ’
Temperature Thallus =
i OF Millimeters
Ld Nina SR Bl nae a a 8
eS cede a agate pe ec eae S28 Ee
1 fori ol uty chin ce eee ee Cee Oe Ones ae ea 15
INO P 22 Ey Biss ona le ral a RSS ee a ge ee 17
Dae Set ee tee a. iene diac sce tie wee ey et 28
DOO ee CE AE Ae eyes Rhee aye AY. Kewl 28
BOP Sera ie ete ee a a ote Eee hee an
mr AER ecm e eS Sraees aae emule ch clara i avnven 12
Bona SORT eee sabe eto el eheeeaak Phe al Es No growth

It appears from the table that the highest temperature selected for
the experiment was one near, but perhaps slightly above, the optimum
for the growth of this fungus, but it is to be noted that growth takes
place at a temperature much below the lowest temperature selected
and infection occurred on inoculated plants in the cultures at all three
soil temperatures employed.

It is unfortunate that a low temperature was not selected that
would at least have permitted the growth of beans even though poorly.
It is well known® that beans require a warm soil for their best develop-
ment. In acold soil presumably bean plants would not have as great
vitality and might have proved particularly susceptible to this hem1-
parasite. Likewise, in the case of cabbage it is well known that the
plants do well in a cool summer and poorly in a warm one. At the
higher temperatures the plants may possess a lower degree of vitality
and hence should be more susceptible to facultative parasites. This
point Gilman passes over lightly in his work.

CORNELL UNIVERSITY,
ITHACA, NEW YorK.

9 Reynolds, J. B. Temperature in relation to seed. Ont. Agr. Col. Rept.
29 (1903): 9-II. 1904.
Sevey, Glenn W. Bean Culture, p.7. New York. 10914.

THE DEVELOPMENT OF CORTINARIUS-. PHOLIDEUS
W. H. SAWYER, JR.

INTRODUCTION.

Cortinarius pholideus is characterized by the peculiarly strong
development of dark, pointed, erect scales on the pileus and stem.
This feature is unusual in Cortinarius, but is very striking in certain
species of Pholiota, so that Cortinarius pholideus in its general aspects
suggests Pholiota, the spore color in the two genera being the same.
Since I have recently studied three scaly species of Pholiota (12), it
therefore occurred to me that it would be extremely interesting to
study the development of this species, which I found in all stages of
development in the same region in which the Pholiotas! were collected.
Especially would it be interesting to determine the formation of a
cortina in a species where such a prominent, coarse, universal veil is
Present.

PRIMORDIUM OF THE BASIDIOCARP

The very young fruit-body is elongate, composed of slender,
closely interwoven hyphae, with numerous interhyphal spaces. These
hyphae are, in general, parallel with the long axis of the basidiocarp;
they have abundant protoplasm and are active in growth, as indicated
by their deeply staining property and long slender cells. The periph-
eral threads, however, take the stain poorly or not at all. They
turn outward on all sides, and in an extremely early stage the outer-
most cells are enlarged, dead, and brown in color.

This outer zone of differentiated hyphae forms a loose-meshed
envelope for the entire plant, and is a universal veil, or blematogen, in
the sense in which this structure has been interpreted by Atkinson (4).

Figure I represents a median longitudinal section through a very
young fruit-body, which is about one millimeter long and half a
millimeter in width. The hyphae are slender and very uniform in
size, averaging about 3 win diameter. The loose peripheral threads
belonging to the blematogen, however, are enlarged, many of them
being 10 u in diameter, dead, and brown. A conspicuous feature of

1 Woods in vicinity of Seventh Lake, Adirondacks, N. Y.
520

THE DEVELOPMENT OF CORTINARIUS PHOLIDEUS 521

the young fruit-bodies of this species are the numerous interhyphal
spaces scattered throughout the basidiocarp.

DIFFERENTIATION OF THE STEM FUNDAMENT

Very early in the development of the fruit-body the hyphae in the
basal part increase in number and show evidence of more rapid growth,
so that this portion of the basidiocarp becomes more dense in struc-
ture. This new growth is the primordium of the stipe, which, by
progressive growth and differentiation, finally reaches the apex of the
fruit-body. From the beginning, the structure of the basidiocarp is
compact, ‘and the gradual progressive differentiation of the stem
fundament does not at first produce any marked change in appearance.
In figure 2 the fundament is well differentiated nearly to the stem
apex; the latter is still in the primordial condition. In figure 3 at the
left is shown the compact cortex of the stem fundament, from which
the looser hyphae of the inner portion of the blematogen radiate out-
ward and upward. The cortex of the stem fundament is somewhat
more dense and deeply staining than the inner portion, as shown in
figure 5.

A similar, but more marked origin of the stem, has been shown to
occur in two species of Lepiota (6), in Rozites gongylophora (11), in
the three species of Pholiota already mentioned, and in five species of
Cortinarius (8).

DIFFERENTIATION OF THE PRIMORDIA OF PILEUS AND HYMENOPHORE

After the organization of the stem fundament the hyphae in its
apical end take on more rapid growth and branch freely, as indicated
by their deeper stain and by the fact that they traverse the interhyphal
spaces at this time. This new interstitial growth causes a bulbous
expansion of the stem apex, which marks the young primordium of
the pileus (fig. 6). At the same time the peripheral hyphae of this
apical region, instead of growing, in general, straight upward, as
they have done heretofore during the development of the stem
fundament, now grow outward in all directions (fig. 7). On the lateral
surfaces they become subject to epinastic influence and turn strongly
downward, forming the pileus margin, as shown in figure 8 by the
small deeply stained area on either side of the pileus, beneath the
blematogen layer.

522 We -H. SAW YER, 7 PR.

Almost simultaneously with the formation of the pileus margin the
hyphae of its under surface begin to grow outward and downward very
rapidly. These hyphae are slender, very rich in protoplasm, crowded
together, and with terete ends. Their outward growth while under
the influence of epinasty causes them to curve strongly, so that the
ends point downward, or even inward toward the stem. This ring of
new growth surrounding the stem apex is the hymenophore primordium.
A median longitudinal section at this stage shows it as a deeply stained
region on either side, as in figure 10. Since the primordium of the
hymenophore is formed from hyphae of the pileus margin, and at
practically the same time with the latter, it is extremely difficult to
point to the exact stage at which it originated. As has been stated,
the hyphae of the pileus margin stain deeply and by new growth in-
crease the density of its structure, and the beginning of this period of
increased activity probably marks the origin of the hymenophore
primordium.

The appearance of this new fundament definitely marks off the
pileus area from the stem primordium. As development continues,
the pileus broadens centrifugally and becomes more compact by in-
terstitial growth. At the same time the hymenophore primordium,
by the intercalary growth of new hyphae from the pileus, and by the
increase of its own elements, likewise develops centrifugally, and keeps
pace in its growth with the pileus margin.

FORMATION OF THE PALISADE LAYER

For a time the growth of the hyphae composing the hymenophore
is very rapid and uneven, the pointed ends of some of the threads grow-
ing down beyond the others, so that the surface is rough and jagged
(figs. 12 and 13). Gradually, however, the hyphae acquire a more
uniform rate of growth, and the ends reach the same level, becoming
clavate and crowded. This condition of the hymenophore in which the
hyphal ends form an even, compact surface is the palisade stage. Such
a condition is shown in figure 23. In this species, as in others pre-
viously investigated, its development is centrifugal, from the stem
toward the pileus margin. Here however it develops very uniformly,
so that at one time in the same fruit-body the palisade occupies the
whole area of the hymenophore except the extreme margin. In all
the species. that have had this phase of their development described,
the formation of the palisade is more gradual, so that in the same fruit-

THE DEVELOPMENT OF CORTINARIUS PHOLIDEUS 523

body there is present at the same time the palisade condition of the
hymenophore and a considerable amount of primordial tissue.

THE ANNULAR PRELAMELLAR CAVITY

When the pileus and stem areas become differentiated from the
primordial tissue of the basidiocarp, in the angle formed by the junc-
tion of these two structures a small amount of primordial or ground
tissue is left. The primordia of pileus, stipe, and hymenophore grow
more rapidly than does this ground tissue and as a result tensions are
produced which cause it to become loose in texture and to tear apart.
This results in the foimation of a cavity in the form of a ring around
the apex of the stem, beneath the surface of the hymenophore. In
this species the formation begins very soon after the origin of the
hymenophore, as shown in figures Io and 11, where the tissue immedi-
ately below the hymenophore primordium is becoming loose through
the lateral and upward pull exerted on it by the margin of the expand-
ing pileus. In figures 14 and 15 the development has proceeded
further, so that an actual cavity is formed, although still weak and
spanned by hyphal threads. By the time the level palisade stage is
reached the gill cavity is well differentiated and entirely free from
intervening tissue, as shown in figures 16, 17 and 23. Like the pri-
mordium of the hymenophore and like the palisade, its development is
from the stem toward the pileus margin, so that its earliest appear-
ance is close to the stem. ‘The two tangential sections represented by
figures I2 and 13 show this fact; in figure 13 the cavity, while not com-
plete, is more strongly developed than in figure 12, a section of the
same fruit-body nearer to the pileus margin.

THE ORIGIN AND DEVELOPMENT OF THE LAMELLAE

At the time of the beginning of gill formation the pileus and stem
are completely formed and the gill cavity is well defined. The hy-
menophore is in the stage in which there is an even palisade layer extend-
ing from the junction of the stem and pileus nearly to the pileus margin;
near the latter the palisade grades off into the primordial condition of
the hymenophore. The palisade is composed of small hyphae with
blunt and crowded ends (fig. 23). The continued growth of these
hyphae and the intercalation of new elements from the hymenophore
above gives rise to sufficient lateral pressure to throw the palisade

524 / W. H. SAWYER, JR.

surface into downwardly projecting folds (fig. 24) which are the first
gill salients. At the same time a more rapid growth of the hymeno-
phore downward in radial, regularly spaced areas directs the formation
of the folds, as described for Agaricus rodmant (5) and species of
Coprinus (7), so that the gills are radially symmetrical with reference
to the stipe.

The origin of the lamellae is next the stem, and by continued
growth and differentiation the lamellae develop toward the pileus
margin. This centrifugal manner of formation enables one to study
their development by means of serial longitudinal sections from the
pileus margin toward the stem, since they are youngest near the former
and become progressively older as they approach the latter. Figure
23 represents a section near the pileus margin. The even palisade
occupies the greater part of the hymenophore surface, with a little of
the primordial tissue on either side. The gill cavity is well formed.
The tissue below the latter belongs to the stem cortex and universal
veil, together with some ground tissue belonging to the partial veil.
Figure 24 represents a section a little nearer to the stem. The palisade
is no longer level, but has an undulating surface, with two slight, very
broad folds. In figure 25, still nearer to the stem, these two folds are
more pronounced, and at the right the beginning of a third may be
noted. The breadth of these folds, and their distance apart, can
leave no doubt that they are the first salients of the lamellae them-
selves.

The trama of the mature gill (fig. 34) has its origin in the hyphae
beneath the palisade layer which grow down into the young gill salient.
Further growth takes place by the elongation and enlargement of
these hyphae. Throughout the center of the lamella they are com-
pactly interwoven, with their general direction of growth toward the
edge (fig. 34). Laterally, however, they turn outward and form the
hymenial layer of the lamella.

The primary gills, because of their radial arrangement with the
stipe as a common center, diverge as they approach the pileus margin.
Continued growth of the hymenophore results in the production of
shorter secondary lamellae between their outer ends in the same way as
that in which they were formed. Figures 28-32, however, show the for-
mation of two gills in a somewhat different manner. In figure 28, a
section near the stem, it will be noted that a gill salient occurs that is
unusually broad. In the successive sections it can be seen that this

THE DEVELOPMENT OF CORTINARIUS PHOLIDEUS 525

broad salient, by branching, forms two salients, each of which there-
after develops into a lamella in the usual way. This probably illus-
trates the method of origin of the dichotomous or forked lamellae
characteristic for Cantharellus and certain species of Russula.

Figures 35-37 illustrate a condition due to the strongly inrolled
margin of the pileus. Figure 35 is of a section tangential to the pileus
margin. In the center a lamella appears with a cavity on either side.
Figures 36 and 37, respectively, nearer the stem, show the same con-
dition; the lamellae appear as bars continuous from the upper to the
lower part of the pileus, with separate cavities between them. The
gills, however, have not become continuous with the tissue below by
growing down and uniting with it. This tissue belongs to the hymeno-
phore of what is morphologically the under surface of the pileus. The
inrolling of the margin of the latter, however, has reversed the posi-
tion of the hymenophore. The presence of the salients of secondary
lamellae on this lower surface serves to make this more clear. The at-
tachment of the gills below, as well as above, represents their point
of origin. The secticns are not cut perpendicular to these points, but
are tangential to the “‘backs”’ of the lamellae; their direction of growth
is not in the plane of the section, but at right angles to it. The spaces
between the lamellae are extensions of the general annular cavity
nearer the stem.

THE BLEMATOGEN

Before any internal differentiation takes place the young basidio-
carp is completely enveloped by a universal veil or blematogen. The
hyphae comvosing this outer layer are differentiated from the other
elements of the fruit-body by the fact that their cells are short and
enlarged, the outer ones being dead, with thick brown walls and
scanty content. ‘The direction of these hyphae is outward and up-
ward. They diverge at the ends, forming a loose structure easily
rubbed off during growth or in the manipulation of the fruit-bodies
preparatory to study.

The blematogen has a very striking appearance at about the time
of the formation of the hymenophore primordium (fig. 10). The
large hyphae stand straight out from the pileus surface, their clear
vellow-brown walls, which do not stain at all, contrasting sharply with
the deeply stained and closely interwoven elements of the pileus. At
this same time the weft of hyphae between the pileus and blematogen,

526 W. H. SAWYER, JR.

ieft when the former became differentiated from the primordial tissue,
loosens, probably because of partial cessation of growth, and forms
a thin layer with many interhyphal spaces extending over the surface
of the pileus (figs. 17 and 19). As the plant approaches maturity the
erect hyphae of the universal veil (figs. 18, 19) become aggregated into
little tufts or clumps that form the erect, dark scales covering stem and
pileus, so characteristic of this species.

A peculiar and interesting feature of the blematogen is its double
character over the margin of the pileus, as shown in the left side of
figure 14 and in figure 15. The outer layer is characteristic of the
universal veil elsewhere on the plant, being composed of large, thick-
walled cells that radiate outward in loose arrangement. ‘The inner
portion, however, is very different in appearance. The hyphae are
slender, with abundant protoplasm and thin walls. Instead of grow-
ing outward in loose structure they lie closely side by side and passing
up over the edge of the pileus margin become ingrown with the pileus
surface an appreciable distance above its free edge. Kniep (10) has
demonstrated that in hyphal threads bearing clamp connections, the
growing end always lies in the direction in which the obtuse angle,
formed by the junction of the cross wall of the hypha and the cross
wall of the clamp, opens. These inner blematogen hyphae bear nu-
merous clamp connections, whose walls all form angles opening upward;
therefore these hyphae could not have grown down from the pileus,
but must have had an upward direction of growth. Furthermore, in
the section shown in figure 16, the free ends of some of these hyphae
may be seen interlacing with the threads of the pileus surface just above
the margin of the latter. It is probable that the growth of this inner
layer is slow, and its union with the pileus is due to the active outward
and downward growth of the hyphae belonging to the latter, which
interweave with the threads of the former. A duplex blematogen has
been described by Miss Douglas in Cortinarius anfractus and C. ar-
mullatus (8), differing, however, from the condition here in that the
outer layer in these two species is thin and compact, while the inner
part is loose and floccose.

THE: MARGINAL. VEE

The marginal veil is very poorly developed in this species, as com-
pared with Agaricus rodmani (5), Armillaria mellea (3), Agaricus
comtulus (2), species of Hypholoma (1), and other species. After the

THE DEVELOPMENT OF CORTINARIUS PHOLIDEUS 527

differentiation of pileus and stem, some ground tissue is left in the
angle between them. | This is nearly all broken away in the process of
formation of the gill cavity, but a small amount may remain attached
io the pileus margin, beneath the blematogen. This is increased by
the downward growth of a few hyphae from the extreme margin of the
pileus, and in figures 14 and 15 it is probable that the inner layer of the
veil described above does not belong entirely to the blematogen, but
has on its inner surface some hyphae belonging to the marginal veil
proper, as limited from the universal veil or blematogen.

THE CORTINA

The name ‘“cortina’’. is a term applied especially in the genus
Cortinarius to the veil composed of delicate silky fibrils stretching
from the pileus margin to the stem. It is usually evanescent, although
in a few species, like C. armullatus,it may persist for a long time in the
form of rings about the stem. In C. pholideus it breaks away early,
leaving a very slight ochraceous annulus around the top of the stipe
that disappears with age. Occasionally a half-grown plant is found
with the arachnoid veil still intact. It is light in color, almost white,
and stretched tightly over the gills. The fibers composing it are very
slender, and this character, together with its lighter color, distinguishes
the cortina from the brown-walled, larger hyphae of the outer blema-
togen layer, which is external to the cortina in varying amount, de-
pending on how much has been rubbed off during the development of
the plant. The cortina is composed of the hyphae of the inner layer
of the blematogen, together with whatever marginal veil may be pres-
ent. In figures 20 and 21 it may be seen extending from the pileus
margin to the stem. Outside is tissue belonging to the outer layer of
the blematogen. Figure 22 shows a condition so common as to be
almost characteristic in this species, in which the pileus margin is so
strongly inrolled that it has become free from the cortina, which is
attached to the pileus surface above its margin, thus showing that
the cortina represents here the inner zone of the duplex blematogen.

Fries (9) muse have regarded the cortina as a structure distinct
from the universal veil or blematogen, for, although its presence was
used by him for a generic character, only two of his six subgenera,
namely Myxacium and Telamonia, are said to possess a universal veil.
C. pholideus, however, put by Fries in the sub-genus Inoloma, has a
universal veil, and the same has been found by Miss Douglas in species

528 W. H. SAWYER, JR.

representing two other Friesian subgenera. No generalizations can
be made until the development of many more species is known, but
the evidence at hand indicates that the presence of a universal veil
(blematogen) is constant for the genus. If so, it is probable that it
plays some part in the origin of the cortina, as in the species studied
by Miss Douglas, and as it does in this species.

In conclusion, I wish to acknowledge my indebtedness to Professor
George F. Atkinson, under whose direction the greater part of this
work was done at Cornell University, for his helpful interest and
kindly criticism.

SUMMARY

1. The primordium of the basidiocarp of Cortinarius pholideus is
composed of slender hyphae interwoven into a compact structure
with numerous interhyphal spaces, and enveloped in a layer of dif-
ferentiated hyphae.

2. These enveloping, radiating hyphae form the blematogen or
universal veil. They are loose in their arrangement, with large,
thick-walled cells. Soon after pileus formation the blematogen shows
a double character over the pileus margin and gill cavity. The inner
layer has an upward growth direction and the hyphae of the pileus
surface interlock with its upper portion.

3. The appearance of the stem fundament is the first differentiation
to take place within the basidiocarp. It is formed in the base of the
fruit-body, and advances to the apex by progressive growth and dif-
ferentiation.

4. The pileus is formed by the expansion of the stem apex, due to
interstitial and divergent growth. The lateral hyphae of the pileus
fundament by epinastic growth form the pileus margin.

5. Perpendicular downward growth of hyphae from the under
surface of the pileus, beginning in the angle between stem and pileus,
forms the primordium of the hymenophore as an annular zone of new
growth surrounding the stem apex. At first, because of unequal
growth of its hyphae, the primordium is uneven and jagged, but later
the ends of the hyphae grow down to the same level, forming the even
palisade zone.

6. The annular prelamellar cavity is formed by the breaking away
of ground tissue left in the angle between stem and pileus after their
differentiation, due to the growth and expansion of these parts. A
small amount of this ground tissue may remain attached to the edge
of the pileus and form a slight element of the cortina.

THE DEVELOPMENT OF CORTINARIUS PHOLIDEUS 529

7. The lamellae originate as downward-growing folds of the level
palisade zone, through.the influence of lateral pressure in the palisade,
and, more particularly, by downward growth of hyphae from the hy-
menophore in radial, regularly spaced areas. Their differentiation is
centrifugal, from the stem toward the pileus margin. The first folds
or ridges in the hymenophore are the salients of the lamellae them-
selves. The gill trama is formed by the downward growth of hyphae
from the hymenophore into the gill salient, and increases by inter-
stitial growth.

8. The cortina is the silky veil stretching over the gills, attached on
the one hand to the surface of the pileus margin and on the other to
the stem. It is composed of the hyphae of the inner layer of the blema-
togen, together with fragments of the ground tissue below the hymeno-
phore. It is covered externally by remnants of the outer layer of the
blematogen, as indicated by the dark patches that may be attached to
its outer surface.

BOTANICAL DEPARTMENT,
CORNELL UNIVERSITY.

LITERATURE CITED

1. Allen, Caroline L. The Development of Some Species of Hypholoma. Ann.
Mycol. 4:.387-394. Pls. 5-7. 1916.

2. Atkinson, Geo. F. The Development of Agaricus arvensis and A. comtulus.
NiCr L Our DO. L322. “PIs. 1,2. IOLA.

3. —— The Development of Armillaria mellea. Mycol. Centralbl. 4: 112-121.
Pisei2 |) OTA:

4. —— Homology of the Universal Veil in Agaricus. Mycol. Centralbl. 5: 13-19.
elsiak—3.. \LOT4:

&, —— The Morphology and Development of Agaricus rodmani. Proceed. Amer.
Phil Soc. §4:: 300-342. . Pls: 7-13. 19015.

6. —— The Development of Lepiota cristata and L. seminuda.

7. —— Origin and Development of the Lamellae in Coprinus. Bot. Gaz. 61:

89-128. Pls. 5-12: 1916.

8. Douglas, Gertrude E. A Study of Development in the Genus Cortinarius.
Amer. Journ. Bot. 3: 319-335. Pls. 8=15.- 1916.

g. Fries, E. Hymenomycetes europaei. 1874.

10. Kniep, Hans. Beitrage zur Kenntnis der Hymenomyceten. Zeit. Bot. 5: 615,
COA else Ah LOLS:

11. Moéller, A. Die Pilzgarten einiger sttdamerikanischer Ameisen. Bot. Mittheil.
Trop6:-1—127. ibis. 17.0 21803.

12. Sawyer, W.H., Jr. The Development of Some Species of Pholiota. Bot. Gaz.
64: 206-230. Pls. 16-20. 1917.

530 W. H.. SAW VERY PR:

DESCRIPTION OF PLATES XXVIII-XxIx

The following microphotographs were made by the author, some with the Spen-
cer Lens Co.’s horizontal camera with Zeiss lenses, the others with a Bausch and
Lomb microscope equipped with Zeiss lenses and a Bausch and Lomb vertical camera
attachment.

PLATE XXVITI

Fic. 1. Primordium of the basidiocarp. The hyphae are closely interwoven
to form a compact structure. Many interhyphal spaces occur, scattered through the
fruit-body. On the outside are a few blematogen hyphae. X50.

Fic. 2. A fruit-body somewhat older than the preceding. The stem fundament
is differentiated to near the apex, where the hyphae are still in the loose primordial
condition. On the sides may be seen the basal parts of blematogen hyphae, their
outer ends having been lost. X30.

Fic. 3. An enlargement of the right side of the preceding figure. The loose
blematogen threads are shown growing out from the compact stem cortex. 215.

Fic. 4. A young fruit-body after formation of the stem primordium. The
blematogen may be seen enveloping the stem fundament. X30.

Fic. 5. Aslightly older stage, showing the blematogen radiating from the stem
cortex. The latter is compact; the apex is slightly more deeply staining, showing
that growth is more active at that point. X30.

Fic. 6. <A state slightly older than the one shown in fig. 5. By interstitial
growth the stem apex has become bulbous, forming the pileus primordium. Outside
is the blematogen. X50. |

Fic. 7. A fruit-body in median longitudinal section that is a very little older
than the one shown in Fig. 6. The pileus fundament is a little larger and the hyphae
in the peripheral zone are growing straight outward over its entire surface. 650.

Fic. 8. Median longitudinal section. By-epinastic growth the outer, lateral
hyphae have grown downward to form the pileus margin, which appears as a small,
deeper stained area on either side. X30.

Fic. 9. An enlargement of the right side of Fig. 8. Near the center can be
seen the pileus margin, whose hyphae extend in a downward and slightly outward
direction. Immediately below it the ground tissue is beginning to break away to
form the gill cavity. On the outside, at the right, is the loose universal veil. XII5.

Fic. 10. An older stage; the hymenophore primordium appears on either side,
on the margin of the pileus. Over the pileus the brown, enlarged hyphae of the
blematogen stand straight outward. X30.

Fic. 11. An enlargement of the right side of Fig. 10, showing details of above
mentioned structures and early indication of formation of the gill cavity. X1II5.

Fic. 12. Tangential section, showing the uneven condition of the hymenophore.
Below the ragged surface of the latter the ground tissue is beginning to break away
to form the gill cavity. 650.

Fic. 13. Tangential section of the same fruit-body, but nearer to the stem.
The hymenophore is becoming more even and the gill cavity is much better formed.
X 50.

Fic. 14. Median longitudinal section. On the left may be seen the duplex
nature of the blematogen over the pileus margin. X30.

THE DEVELOPMENT OF CORTINARIUS PHOLIDEUS 531

Fic. 15. Left side of preceding figure, more highly magnified. On the outside
are the large hyphae of the outer blematogen layer, with an outward and upward
direction. Within is the inner blematogen layer, composed of slender parallel hyphae,
interlacing with the surface of the pileus above its margin. Below the hymenophore
is the gillcavity. XII5.

Fic. 16. Right side of median longitudinal section of a fruit-body a little older
than the preceding one. The palisade zone is differentiated and the gill cavity is
completed. Extending up over the edge of the pileus margin may be seen hyphae
of the inner part of the blematogen. XII5.

Fic. 17. The right side of median section, showing the palisade layer and the
universal veil, with the loose area between the latter and the pileus surface. X25.

Fic. 18. Median section of fruit-body after the gills are formed. The pileus
and stem have become very compact through branching and growth of their hyphae.
The pileus margin is strongly inrolled, as in all the fruit-bodies sectioned at this
stage. The cells of the blematogen hyphae are collapsed and shrunken and show
signs of aggregation into the tufts that later become the erect squamules on the sur-
face of pileus and stem. The narrow area of loose tissue between blematogen and
pileus surface may be seen here. X15.

Fic. 19. A fruit-body a little younger than the preceding, showing well the
radiating hyphae of the blematogen. X15.

PLATE: XXX

Fic: 20. The section represented by this photograph is from a fruit-body at a
stage after the gills are well formed. The margin of the pileus is strongly incurved
and the edge is free from the veil. X30.

Fic. 21. This section shows well the structure of the cortina. It is composed
of the slender hyphae of the inner part of the blematogen and the marginal veil. On
the outside are remnants of the outer layer of the blematogen. Since the pileus
margin is not strongly inrolled in this particular fruit-body, the edge has not become
free from the veil. X30.

Fic. 22. A portion of the right side of a section showing the pileus with its
edge entirely free from the cortina, due to its inrolled character. X30.

Fics. 23-27. A series of sections showing the origin and development of the
lamellae. In Fig. 23 a section near the pileus margin is shown. In the center is the
_ well developed gill cavity. Above it is the even palisade area of the hymenophore,
bordered on either side by primordial tissue. Below is loose ground tissue on the
stem. In Fig. 24 the first gill salients are seen as two slightly downwardly projecting
broad folds in the palisade. Fig. 25 shows these two salients better developed,
toward the stem, and at the right a third is appearing. Fig. 26 represents a section
of the same fruit-body tangent to the stem. On the left is the level palisade and on
the right development has proceeded further, so that a gill salient has been formed.
Fig. 27 is of a section nearer to the center of the fruit-body, showing the same features
as in the preceding section. XII2.

Fics. 28-32. Stages in the origin and development of a dichotomous gill
through the branching of a gill salient. The section shown in Fig. 28 is near the
stem, and the following figures are nearer the pileus margin in order. X50.

532 W.-H: SAWYER, -jiR-

Fic. 33. A section tangent to the stem at its junction with the pileus, showing
two adnate gills. X30.

Fic. 34. Structure of the mature gills. The hyphae of the trama turn out-
ward on the sides and contribute to the formation of the hymenium, which shows as
the deeply stained layer covering the gill. Xg3o.

Fics. 35-37. Tangential sections in the inrolled margin of a nearly mature
fruit-body. In Fig. 35 the section is tangent to the “‘backs”’ of the gills. In Fig. 36
the lamellae appear to grow across the gill cavity as bars, with a separate cavity be-
tween each two gills. In reality it is the backs of the gills seen in section, and their
attachment below, as well as above, represents their point of origin. Due to the
inrolled pileus margin the tissue at either end of the ‘“‘bars’”’ belongs to the same mor-
phological under surface of the pileus. Fig. 37 shows the origin of secondary gills
between the primary lamellae, both above and below. X15.

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE XXVIII.

SAWYER: DEVELOPMENT OF CORTINARIUS PHOLIDEUS.

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE XXIX.

SAWYER: DEVELOPMENT OF CORTINARIUS PHOLIDEUS.

Pent oLRUCTURE. AS RELATED? 1O7 ENVIRONMENT

HERBERT C,. HANSON

INTRODUCTION

This investigation was begun during the summer of 1915 and
carried on for about a year. Preliminary observation showed that
the leaves growing in the sun at the south periphery were thicker than
the leaves growing in the shade at the center of the same tree. The
purpose of the investigation was to find out the exact differences in the
structure of the leaves from the two positions, and to compare by
measurements of the factors the environments in which the leaves were
growing. ‘The factors measured were light, evaporating power of the
air, temperature, humidity and wind. A study of transpiration was
also made during the summer of 1916.

H1storical

The relation of leaf structure to factors of the environment has
been studied by numerous investigators. Although very few works
give factor measurements, some of the more recent investigations on
leaf structure are here reviewed. One of the most detailed works is
by Mrs. Clements (12), which shows the differences in the structure of
numerous hydrophytic, mesophytic and xerophytic leaves, and of
sun and shade leaves on different plants of the same species. The
physical factors, light, water-content of the soil, humidity and tem-
perature, were measured for most of the habitats in which the plants
studied were growing. Good reviews of earlier articles on leaf struc-
ture, and of explanations concerning the structure and formation of
palisade cells, are given by Mrs. Clements.

Experiments performed by Eberhardt (14) showed that humid air
caused an increase in the size of the leaf, in the amount of chlorophyll
and in root development; while dry air caused an increase in the thick-
ness of the cuticle, in the number of stomata, and in the amount of
sclerenchyma, woody tissue and palisade.

Brenner’s (9) experiments on various succulent plants are interest-
ing and important. Plants grown in moist air showed the following

5168)

534 HERBERT C. HANSON

changes as compared with plants growing in the normal environment:
storage tissue, the fibro-vascular bundles, air-spaces, dry weight, ash
and acid content of the leaves all decrease; chlorophyll tissue increases;
chlorophyll cells become more isodiametric; walls of epidermal cells
become wavy; number of stomata per leaf increases, although the
number for a unit area may be the same; and epinasty replaces hy-
ponasty, so that the leaves grow at right angles to the stem. He be-
lieved the air factors, not the soil factors, determined these changes,
but that physical explanations do not always seem adequate to account
for the changes.

Brenner (10) concludes from his anatomical and experimental
study on Quercus leaves that modifications caused by the environment
are hereditary and may develop into new species.

Bonnier (7) selected about fifty species of perennial plants at
Fontainebleau. Each plant experimented upon was split, one part
planted at Fontainebleau, the other at Toulon. The plants set out
at Toulon became like the wild plants surrounding Toulon in leaf and
and wood structure and in external characters.

Hansgirg (18) gives descriptions in detail of more than fifty types
of leaves, with a discussion of their ecological advantages.

Chrysler (11) compared the leaf structure of nine strand plants
from the Atlantic coast near Wood’s Hole, Massachusetts, and from
the vicinity of Chicago. The leaves of the maritime plants were from
less than once to a trifle over twice as thick as the inland plants. The
increase in thickness was mostly due to increased palisade develop-
ment. Greater compactness in tissue and’ increased thickness of the
outer epidermal wall were found in certain plants. The amount of
salt in the soil is given as the probable cause for the variations.

Boodle (8) in his experimental study on the leaves of Pieris aquilina
Linn. found that leaves grown in dry and exposed situations had a
xerophytic structure while those grown in sheltered positions were
more mesophytic. The former leaves possessed hypoderm, the latter
had none and the palisade was poorly developed, or entirely missing.
The same differences were found on leaves of the same plant, or on
different parts of the same leaf. A plant that had been producing
shade leaves in a moist greenhouse, produced sun leaves when placed
in a garden. The mature structure of the leaf is not determined at
an early stage in the leaf’s growth.

Copeland (13) found great variations in the shape and thickness of
leaves on the same branches of various plants.

LEAF-STRUCTURE AS RELATED TO ENVIRONMENT 535

Hesselman (23) brought out many important facts in his very de-
tailed and quantitative work. Leaves formed in the stronger light of
the forest in the spring developed more palisade than leaves that de-
veloped later in the weaker light. The trees that had high light re-
quirements, Betula, Fraxinus, produced sun-leaves, while the trees of
low light requirements, Quercus, Corylus, produced both sun and
' shade leaves. The trees in the first group produced starch in all the
leaves, while those in the latter group produced no starch in the inner-
most leaves. Shade leaves make more starch than sun leaves of the
~ same species when the light is equal. The production of starch de-
creases from spring to summer in the forest more in sun plants than in
shade plants. The respiration of sun plants is far greater than of
shade plants. If the leaf surface is equal, transpiration increases with
the amount of palisade. Sun plants in the sun transpire much more
than shade plants in the shade. The work also contains good repre-
sentations of ecological structures of plants.

Bergen (4) compared the transpiration, color, size and the general
structure of sun and shade leaves of the same individuals of the fol-
lowing: Olea europea sativa, Pistacia, Lentiscus, Quercus, Ilex, Rhamnus
Alaternus. He found the ratio in thickness of sun leaves to shade
leaves to be 1.5-3.7 to 1. The sun leaves had thicker cuticle, more
palisade, smaller air spaces, greater bundle development, 15 percent
more stomata as determined by two observations, greater scaliness,
darker color and smaller area. The greater transpiration in the sun
leaves was due to their greater activity, because their larger stems and
bundles transfer the water more quickly, and because the greater
thickness of the leaves afforded a larger interior evaporating surface.
In another article (5) Bergen compares the thickness and transpiration
of new and old leaves. From the result of another study (6) he states
that ‘it is undoubtedly a fact that the great majority of woody dicoty-
ledons have leaves which when freely exposed to the sun are concave
on the upper surface and that this concavity usually lessens or disap-
pears in the case of much shaded leaves on the same plant.”

Oltmans (26) noted that the leaflets growing on the south periphery
of Robinia Pseudo-Acacia trees were concave, while those on the north °
periphery were flat. Wiesner (32) also observed that while the upper
surface of peripheral leaves was concave, the leaves in the shade of the
same tree were usually flat.

Herriott (22) gives the frequency and violence of the wind and the

536 HERBERT C. HANSON

peaty soil as causes for the xerophytic leaf structure of some: New
Zealand plants.

Raunkiaer (27) found that palisade tissue was equally well de-
veloped in the leaves of certain plants above the water and to a depth
of twenty centimeters. Below this the length of the palisade cells
gradually decreases. No palisade tissue was distinguished in shade
plants under water, nor above, to a height of about thirty centimeters.
His discussion of the causes of palisade development and the orienta-
tion of the palisade cells is valuable.

Lubimenko (25) determined that the chloroplasts of certain
ombrophilous plants, as Tilia and Abies, were greater in size and in
sensitivity than the chloroplasts of ombrophobous plants, as Pinus and
Betula. The pigment was more concentrated in the former group.

Baumert’s (3) work contains a good review on the literature of
structures protecting leaves from light. From his own experiments
he found that a thick white coating of hairs as in Centaurea candidissimu
reduced the heat in the leaf 37.5 percent, shininess reduces the heat
30 percent, and a wax coating up to 13.6 percent. Wiegand’s (31)
experiments show the efficiency of hairy and cutinized coverings in
reducing the water loss by transpiration.

Areschoug (2) maintains that well-developed, compact palisade
tissue reduces transpiration, despite Hesselman’s experiments to the
contrary.

Ewart (15) experimenting upon 7 ilia europaea found that mature
leaves do not increase in size when most of the leaves are removed from
the tree. The increased size of the new leaves which replace those
defoliated is due to the increase in the number of cells.

Sampson and Allen (29) found that the sun leaf transpired on
two to four times as much as the shade leaf of the same species whether
the leaves were placed in the sun or shade. This was explained by
the greater number (20 percent—60 percent) of stomata in the sun leaf.

Harshberger (20,19) investigated the leaf structure of strand plants
in New Jersey, and sand dune plants in Bermuda. He states that the
xerophytic structures are due to intense light, strong winds, and in
rare cases to salt spray. The unequal illumination of the two sides of
the leaf causes the formation of palisade and sponge tissues.

Renner (28), in discussing the relation of wind to transpiration,
_ says that the transpiration of small mature leaves is increased to a far
greater extent by the wind than that of large leaves. This is ex-

LEAF-STRUCTURE AS RELATED TO ENVIRONMENT Oy

plained by the vapor cap which is thicker about the larger leaves. If
the air were absolutely quiet Renner believed the thickness of the
vapor cap would vary with the diameter of the leaves.

Adamson (1) found the xerophytic structures in leaves of certain
species of Veronica to consist in reduced leaf surfaces, reduction in
intercellular spaces, and an increase in the thickness of the cuticle.

Livingston and Brown (24) in their study of the daily march of
transpiration showed that the water content of leaves falls during the
day and rises during the night.

Starr (30) compzred the structures of stems and leaves of plants
on dunes and on flood plains. She discusses the ecological factors of
the dunes, but no measurements are given. The leaves of the dune
plants owe their greater thickness to increased palisade tissue chiefly.

Ganong (16) states that one of his students found that the petioles
from the exposed part of a tree were larger than those from more
sheltered positions.

Haberlandt (17) says that a comparison of the vigorously trans-
piring sun leaves and the feebly transpiring shade leaves of the same
plant shows an increase in the linear dimensions of the vascular system
in the sun lzaves.

Hasselbring (21), working on tobacco plants growing in the sun
and in the shade, showed that the proportions of dry matter, and the
production of plant substance for equal areas of leaf surface were
greater in the sun plants. The shade plants transpired 186.99 cc, of
water in producing one gram of plant substance, while the sun plants
transpired 241.72 cc. He found the water content of leaves from sun
plants to be 81.39 percent and of leaves from shade plants to be 83.68
percent. The sun plants transpired .412 cc. while the shade plants
transpired .224 cc. per square decimeter of leaf surface per hour.

METHODS

The readings of the environmental factors were made in the sun
among the leaves on the south periphery of untrimmed isolated trees
and at the apex of trees growing in the forest. At a height corres-
ponding to the sun readings on isolated trees, readings were taken in
representative positions within the crowns; and for the forest trees
readings were taken among the lowest leaves. For a given species the
readings were taken upon the same individual in the forest or in iso-
lated positions. Care was taken so that the various factors were

538 HERBERT C. HANSON

measured in the same positions in the crown or at the south periphery
of each tree. Cytological material was collected from leaves growing
in these positions respectively. Pieces from about twenty leaves from
each position were killed in chromo-acetic acid and in Juel’s killing solu-
tions. The material from Juel’s solution gave the best specimens for
study. The ordinary paraffine method was used in preparing’ the
material.

The response of the leaves to the environmental factors was shown
in transpiration; in the green and dry weights and water content of
given leaf areas; in the thickness of the leaf and its parts, palisade,
sponge, upper and lower epidermis and the cuticle; in the compactness
of the tissues; in the structure of individual cells; and in the macro-
scopic characters, as area and lobing.

THE PHYSICAL FACTORS

1. Light

The light was measured by means of the Clements’ photometer
between 11 A.M. and 2 P.M. in August, 1915. Four readings were
taken within the crown of isolated trees and among the lowest leaves
of the forest trees for four or five individuals of ten species. From
these 16 to 20 readings for each species the light values, as arranged in
the following table, were averaged.

TABLE I

Showing the Light Values in the Crown of Isolated Trees and at the Base of
Forest Trees

Position of Light Value in Crown
Species Trees* or at Base
Acer sdccharium Naish... nee Isolated 0.0175—crown
Forest 0.0076—base
TAO OMErICONG Nek. 2) ha ee eee Isolated 0.0688—crown
Isolated (L.) 0.1000—crown
: Forest 0.0086—base
Quercus macrocarpa, NIGH. .2.7gee Isolated 0.1132—crown
Forest 0.0754—base
Owercius PUOrd Lig ak aie ee ee Forest 0.0425—base
Quercus clog Wine hao aa ee ee Forest 0.0100—base
Acer sacchariniun oie ee eee Isolated 0.0406—crown
Isolated (L.) 0.0380—crown
Acer negundo L..........0.00% AIS Ie ay, Isolated 0.0979—crown
Olnus americana Lian Isolated 0.0770—crown
Fraxinus pennsylvanica Marsh............ Isolated 0.0497—crown
Celis Occraentalis li, sass. <d) ees eee ee Isolated (L.) 0.059I—crown

* Trees marked (L.) in Lincoln, all others in Minneapolis.

LEAF-STRUCTURE AS RELATED TO ENVIRONMENT 539

The order of the light values in isolated trees, beginning with the
lowest, is seen to be Acer saccharum, Acer saccharinum, Fraxinus
pennsylvanica, Celtis occidentalis, Tilia americana, Ulmus americana,
Acer negundo, and Quercus macrocarpa. In the forest the order is
Acer saccharum, Tilia americana, Quercus alba, Quercus rubra, and
Quercus macrocarpa. Comparing the isolated and forest individuals
of the same species it may be noted that the trees in the forest have
the lower light intensity.

2. Evaporating Power of the Aur

Livingston’s standardized porous cups were used for measuring the
evaporating power of the air. The cups were placed in the trees be-
tween 8 and 10 A.M. and taken down about 5 P.M. The racios of the
amounts of water evaporated from the crown to that from the south
periphery in isolated trees, or from the base to the apex of the crowns
of forest trees were lowest in isolated trees of Ulmus americana and
Acer negundo (1:1.44, 1: 1.48) and greatest in isolated Acer saccharinum
and Tilia americana (1:2.3, 1:2.2). The ratios of five other species
ranged between these. The ratios in the isolated trees were always
less than the ratios in forest trees of the same species, for example, in °
the isolated tree of Quercus macrocarpa the ratio was 1: 1.3, while in the
forest tree it was 1:2.2: This is due to the greater humidity at the
base of the forest tree as compared with the humidity within the crown
of the isolated tree.

The temperature and humidity data indicate that the differences
in the amounts evaporated in the sun and in the shade positions cannot
be explained by these two factors alone. The movement of the air
seems to be the controlling factor. As the water evaporates from the
cups in the exposed situations it is rapidly carried away from the
vicinity of the cups by air currents. As the wind data show, there is
much less air movement within the crown than at the south periphery,
so it is very probable that a vapor blanket is formed about the cups
within the crown. This blanket would very likely be formed also
about each leaf. Renner (28) concluded that the thickness of the
vapor blanket varies with the leaf size. In this way the saturation
deficit would be decreased, and evaporation and transpiration lowered.
Radiant energy may also play an important rdle in causing the dif-
ferences, because it is greater in the exposed than in the sheltered
positions. As the readings were not taken upon all of the trees at the

540 HERBERT C. HANSON

same time, comparisons between species cannot be made except in a
very general way.
3. Temperature

Numerous simultaneous readings of temperature were made in the
shade within the crown and in the sun at the south periphery upon
several isolated trees of Acer saccharum, A. saccharinum, Tilia amert-
cana and Fraxinus pennsylvanica. The greatest differences between
the two positions were 2.8° C. at 2:50 P.M. on Acer saccharum and
1.8° C. at 1:00 P.M.on Tilia americana. Usually the difference was
about 1° C, This difference was caused by the stronger light at the
south periphery.

4. Humidity

Humidity readings were made with cog psychrometers simulta-
neously within the crown and at the south periphery of the same trees
that were used for temperature readings. The greatest differences
were found on Acer saccharum, 16 percent, at noon. The greatest
difference in the case of Tilia americana was 9 percent at2 P.M. The
usual differences in the humidity of the two positions was from I per-
cent to 2 percent. The greatest factor in causing these differences
was the movement of the air. Dense crowns impeded the free move-
ment of the air far more than open crowns, so that trees with dense
crowns, as Acer saccharum, showed greater differences than trees of
open crowns. The greatest saturation deficits were always found on
the exposed parts of trees from 1 P.M. to 3 P.M.

5. Wind

Numerous readings of wind velocity were taken by hand anemom-
eters, operated simultaneously in the two positions. The ratios of the
velocity within the crown to that at the periphery have been deter-
mined from several readings on each tree. In Ace; saccharum the
ratio was 1:2.2; in A. saccharinum, 1:2.1; in (Quercus macrocarpa,
1:2.0;in Ulmus americana, 1: 1.6;in T1l1a americana, 1:1.4;in Fraxinus
pennsylvanica, 1:1.4;1n Acer negundo, 1: 1.3.

The amount of air movement within the crown depends upon the
openness of the crown. The two extremes in crown density in the
above series were Acer saccharum and Acer negundo. As seen from the
figures the least air movement within the crown in comparison to that
at the south periphery was found in the former, and the greatest in the

LEAF-STRUCTURE AS RELATED TO ENVIRONMENT 541

latter. When the wind velocity was high or if it came in gusts there
was less difference in the air movement within the crown and at the
south periphery than when the velocity was low. This is explained by
_ the fact that the leaves and branches were more effective in keeping the
wind out of the crown when the wind velocity was low than when it
was high. As the wind velocity was not the same when all the read-
ings were made, the ratios of the various trees cannot be directly com-
pared with each other. It is highly probable that if readings were
taken on Acer negundo under the same velocity as on Acer saccharum,
1. €., 68.9 meters per minute instead of 58.6, the ratio of wind in the
crown to that of the south periphery would be less than 1:1.3. In the
case of Acer saccharum as high a ratio of crown to south periphery as
I: 8.1 was found when the wind velocity was regular and fairly
low.

Readings were taken with two cup anemometers on a Fraxinus
pennsylvanica near Lincoln. The readings were made every half
hour from 9:45 A.M. to 3:45 P.M. As the wind came from the north-
west, readings were made at the northwest periphery and within the
crown. The wind velocity within the crown was found to vary from
32 percent to 52 percent of that at the northwest periphery. At 9:45
A.M. the velocity within the crown was 3.7 kilometers per hour and
at the northwest periphery 11.6. At 3:45 P.M. the velocity within
the crown was 6.6 km., and at the periphery 12.7 km.

EFFECTS OF THE PHYSICAL FACTORS
1. Transpiration
Method

Branches from the center of the crown and from the south periph-
ery, bearing numerous leaves, were cut under water. These
branches were securely fastened by rubber tubing to the base of bu-
rette tubes graduated to tenths of cubic centimeters. The amount of
water in each tube was about equal throughout the experiment, so
that there would be no error due to unequal heads of pressure. The
tubes, with the branches attached so as to be upright, were fastened in
representative situations in the center or at the south periphery.
Readings were usually taken simultaneously in both situations every
15 minutes. The leaf area was measured from prints on Kresko paper
by means of the planimeter.

542 HERBERT C. HANSON

Results

On July 11, 1916, four sets of branches were arranged on a Fraxinus
pennsylvanica growing isolated in a pasture. The tree had a well-
shaped, fairly dense crown. One potometer containing south periph-
ery leaves and one containing center leaves were placed among , the
leaves at the south periphery about Io feet above the ground. Two
more potometers, one containing south periphery leaves, the other
center leaves, were placed within the crown at about the same height
as the other two potometers. ‘There was very little wind and the sky
was very hazy. ‘The transpirations for periods of 15 minutes and for a
total of three hours are given in the following table:

‘TABLE 2

Transpiration of Fraxinus Pennsylvanica on July 11, 1916

| Time Total | Loss per ; Time Total | Loss per
p Pat |S! 83,0 PM, Gee
1:08 | [22
1323-10.5,. 2) /0,0212 1:38 | 0.10 | 0.0054
| 1538) O19: «is20883a 1:53 20: | YOREO
b 2353) \<d.2 ie eO5 12. 2:08 .20 .O1IO
2108-17 1.3 0 FeO554 2:23 35 .O194
Leaves from 2228 VAL Asis SeOOL Leaves from 2:38) ia 5 .O194
south periphery | 2:38 | 1.45 | .0617 | center of crown | 2:53 35 | .0194
placed at south | 2:53 | 1.4 | .0596/| placed at south 3:08 204), 30166
periphery 2:08 | 40-455) 0017 periphery 3628 35 | .0194
2222 i Tsd5 4) 220017 3:38 55-0304
3:38 [11.35 | 20574 | 3°53 45| .0248
3°53 | 1:3 P0554 4.08 35 | .0194
AtOS "i TAS .0488 |
Total | au hrs) | 14.G0')) 6340 lt 2% hr| 3.55 | 0.1962
1.08 |
fet s23 | ered > 1 ROrOl 22
1:38 55 .0216 1:38 | 0.20 | 0.0087
1353 | 205 .0254 1:53 25 .O1IO
22084) te7i5 ci) 20208 2:08 25 .O1IO
Leaves from | 2:23 250). F202 Te Leaves from 2:22 35 .0155
south periphery | 2:38 | .90 | .0352 | center of crown | 2:38 303) Ole2
placed at center |..2:53.| ».95 |. .0372)| placed at center || 2:5374| aos ona,
of crown | 3:08 | 91:00 <|-* 20262 3.08 .40 10177
2223)4).1.05 .O4II 2°22 40 | .O177
| -33384| "1.10. |) RoAet 338 .40 50177
3:93 0:90 “1-1-0302 3:53 -35° | 0155
| 4508! 0.90) 1) 15,0352 4:08 35.) (sOL55u
Total | 3 hrs'| 10.00.12 22013 234 hr | 3:65 |. some

LEAF-STRUCTURE AS RELATED TO ENVIRONMENT 543

The losses of the south periphery leaves were from 3 to 6 times as
great as those from the center leaves when placed in their native
situations. It is interesting to note that the south periphery branch
in the center of the crown lost more water than the center branch at
the south periphery. This may be due to several causes; probably
the scomata in the leaves of the center branch closed when exposed to
the sun (but other experiments indicate that this was not the case),
or the center leaf may have been transpiring up to its full capacity.
The south periphery leaf had greater capacity for transpiration than
the center leaf, because of its greater amount of solid matter and
chloroplasts. Graphs for the two potometers of center leaves give
evidence by their parallelism for this view, indicating that the center
leaves placed in the center are transpiring at almost full capacity.
Sampson and Allen (29) account for the greater transpiration of the
sun leaves, because they have from 20 to 60 percent more stomata.
Hesselman (23) accounts for the increase in transpiration in his state-
ment that the leaf surface being equal, plants transpire more as they
have greater development of palisade. Bergen (4) explains the
greater transpiration in the sun leaves by their greater activity, by
their greater thickness affording a larger interior evaporating surface,
and by their larger bundles and stems which would transfer the water
more quickly. These graphs also show by their abrupt changes
that the center leaves are more responsive to the environmental
factors than the south periphery leaves. This is explained by
the greater amount of protective material, as thicker cuticle, greater
thickness of the leaf and more solid material in the south periphery
leaves.

On July 22, 1916, acclear, hot day, with a light south breeze, three
sets of branches were used in an experiment on an isolated Ulmus fulva
Mich. Each set contained one potometer of a south periphery branch
placed at the south periphery, and one potometer of a center branch
placed at the center. The sets were run from 114 hours to 214 hours,
readings being taken every 15 minutes. The greatest differences in
the water loss between the south periphery and center leaves occurred
in the first set where in 114 hours the south periphery leaves lost 1.956
cc. per square decimeter, while the center leaves lost 0.1614 cc. per
square decimeter, a ratio of about 12: 1. The least differences were
in the third set, where the south periphery leaves in 214 hours lost
3.10 cc. per square decimeter, while the center leaves lost 0.638 cc.

344 HERBERT C. HANSON

per square decimeter, a ratio of about 5:1. As. the readings for the
third set were made in the morning and those of the first set in the
afternoon, the difference in the ratios of the two sets are most likely
due to the lower humidity, higher temperature and stronger light,
causing greater differences between the center and south periphery
during the first experiment.

The differences in the transpiration of Ulmus fulva leaves is about
twice as great as the differences in Fraxinus pennsylvanica leaves of
July 11. This greater ratio between the exposed and sheltered leaves
in Ulmus is partly due to weather conditions. On July 11 the tem-
perature was lower, the humidity higher, and the sunlight was less
bright than on July 22. The physical factors in the center and at the
south periphery were therefore more alike.

On June 23, 1916, a cloudless, warm day with a light breeze, an
experiment was performed on a well-formed isolated Acer saccharinum.
Two potometers of south periphery leaves were prepared, one was
placed at the south periphery, the other at the center. A potometer
of center leaves, also, was placed at the south periphery. At the end
of 50 minutes readings were made and the positions of all potometers
were changed from south periphery to center or vice versa, and allowed
to run 50 minutes after about 5 minutes for adjustment had been
allowed.

The potometers, containing south periphery leaves, placed at the
south periphery, lost 10.5 cc. and 11.95 cc. When these were moved
to the center of the tree the losses were 3.95 cc. and 2.85 cc., respect-
ively. The center leaves lost 5.4 cc. at the south periphery and 1.2
cc. in the center. The temperature at the periphery of the tree was
practically the same during the 105 minutes. The evaporation from
Livingston’s porous cups was 4.6 cc. in 50 minutes at the south peri-
phery and 3.4 cc. in\the center.

The amount of water lost by transpiration is increased from about
3 to over 4 times when the potometer is changed from center to the
south periphery. The small differences in temperature and evapora-
tion in the two positions compared to the great differences in trans-
piration show that plants, compared with mechanical apparatus, are
more sensitive to environmental factors. Comparison cannot be
made between the three potometers in this experiment as the leaf area
was not measured. Comparison can be made only between the posi-
tions of the same potometer.

LEAF-STRUCTURE AS RELATED TO ENVIRONMENT 545

2. Surface Area and Lobing

In nearly all cases the leaves from the center or from the base of
the crown were larger than those from the south periphery or apex.
The greatest differences were found in the second crop of leaves during
the season on forest forms of Tilia americana, in the first crop leaves of
isolated Fraxinus pennsylvanica and in the first and second crops of
leaves of isolated and forest individuals of Quercus macrocarpa. The
production of a second crop of leaves was due to the warm weather in
June following a cold spring. The least differences in surface extent
were found in the isolated Tilia americana and Acer saccharum. Us-
ually the leaves from the exposed positions were more deeply and
narrowly lobed, and more prominently toothed than the leaves from
the protected positions. The lobing of the south periphery and the
apex leaves of Acer saccharum was less deep than in the center or base
leaves.

3. Green and Dry Weights: Water Content

Two methods were employed in determining the green and dry
weights and the areas of the leaves weighed; one by means of the
Ganong leaf area cutter, the other by weighing entire leaves and then
determining their area by means of the proportional weights, using
solio paper for the leaf prints. Sufficient material was used in each
method so as to render the error negligible. The leaves were always
collected late in the afternoon. The green and dry weights per square
decimeter.of leaf surface and the water content are given in the fol-
lowing table:

In every case the green and dry weights were lower, and the water
content higher, in the shaded than in the sunny positions. The ratio
between the weights is greater in the dry weights than in the green
weights, showing that more solid material was laid down in the leaves
where the light and other factors were more intense. The highest
water content was found in the shade leaves of Acer saccharum, A.
saccharinum and Fraxinus, while the lowest was in Quercus alba and
Q. rubra. The greatest differences in water content between the sun
and shade leaves were found in Fraxinus and Acer saccharum, while
the smallest occurred in Quercus rubra and Q. alba. According to the
amount of dry material in the leaves at the south periphery or apex
the trees fall into the following order: Fraxinus 1.639 g., O. macro-
carpa 1.272 g., Q. rubra 1.190 g., QO. alba 1.173 ¢g., Taha americana

546 HERBERT C. HANSON

TABLE 3

Green Weight, Dry Weight and Water Content of Leaves from the Center and South Pert-
phery of Isolated Trees, and from the Base and Apex of Forest Trees

= Ss —
Green Weight | Dry Weight |
Si Pease SIAC eee ee

| Water Content

Position | lear
of lee Grams | Grams | | Grams | % of | % of
| perSq.| % | perSq.| % | perSq.|Green| Dry
| Dem. | Dem. | Dem. | Wt. Wt.
Acer saccharum (isolated)... .| Center. .| 0.835} 46) 0.354! 38 0.481 74.6| 135.9

| So. Per..4 1.760 |--100 | 0.937 | 100 |' 0.823 | 52.1) 87.9
Acer saccharum (forest)... . SlsBasen ty. 0.878 | 47 | 0.361] 35] 0.517 | 58.8) 143.2
| Apex. . .| 1.882 | 100 | 1.029 | 100 | 0.853 |'45.4| 82.9
Tilia americana (isolated)...| Center..| 1.078} 58 | 0.038] 51 | 0.698 | 64.8] 183.7
So. Per..| 1.861 | 100 | 0.745 | 100 | 1.116 | 59.9| 149.8

* Tilia americana (forest)....| Base....] 0.745 | 29 | 0.223| 20] 0.522 | 70.1| 234.1
Apex. . .| 2.589 | 100 | 1.114 | 100 | 1.475 | 56.9] 132.4

Quercus alba (forest)........ Base... ..| 1.354 | 74 | 0:817 | -70:/0,537 5 39u7z ones
Apex. . .| 1.825 | 1007 1.173 |. 100:),01052+)- 35.74) 355.0

* Quercus alba (forest)...... Base...:| 1:276| 58 | 0.467'| 4'7,| 0,809) 63:4) 17372
| Apex. ..| 2.186 |. 100'|.0.997)| 100 | 1:18@7 547 4) mere

Quercus rubra (forest)....... Base... .|:1.458 | 63 | 0.699 58 | 0.759 | 52.1) 108.5
Apex. . .| 2.331 |.I60 | 1.190 | 100 | 1.141 | 48.9] 95.9

Quercus macrocarpa Centers | 1.469 | 66) 0.579| 46 | 0.890 | 60.7) 153.8
(isolated) ii ak. Wie. ee So. Per...| 2:227 | 190'| 1.272 | 100)|0.055)|"42-0lae sem

Ulmus americana (isolated). .| Center. .| 0.906; 71 | 0.309| 55 | 0.597 | 65.9| 193.2
So. Per..| 1.274 | 100|.0.560 | 100s] 0:744'| 56.0) 127.5

Fraxinus pennsylvanica | Center. ./ 1.428] 59 | 0.467 | 29] 0.961 | 67.3) 205.8
(isolated) s<.. 4s wee ee So. Per..| 2.440 | 100 | 1.639 | 100 | 0.801 | 32.8] 48.8
Acer saccharinum (isolated). .| Center. .| 0.723 | 60/| 0.229] 42] 0.494 | 68.3] 215.9
So. Per..| 1.211 | 100 0.546 100 | 0.665 | 54.9) 121.8

* Second growth leaves of the season.

I.114 g., A. saccharum 1.029 g., Ulmus americana 0.560 g., A. sac-
charinum 0.546 g. The relation of this sequence to tolerance is note-
worthy. The green weights of the leaves from the same position do
not show so much relationship, but Fraxinus, Q. rubra and Q. macro-
carpa weighed most, while Ulmus and A. saccharinum weighed least.

Two crops of leaves were produced by many trees during the
growing season of 1915. The first crop appeared under cold and
humid conditions. The second crop developed about June 25 when
the warmer and drier summer weather had arrived. In the latter
part of August the first crop of leaves showed a much lower water
content than the second crop. The total green weight of the base
leaves of the second crop was less and of the apex leaves greater than
that of the first crop. The explanation of this probably is that the
new base leaves were developed under lower light intensity caused by
the shade of the first leaves; while the new apex leaves appeared when

LEAF-STRUCTURE AS RELATED TO ENVIRONMENT 547

the light intensity and the evaporating power of the air were greater
than when the first crop developed. In the former case less solid
material, and in the latter, more solid material would form in the leaf
cells.

4. Comparison of Leaf Structure

The detailed microscopical study and measurements of the leaf
structure of various trees are summarized in the following observations
for each species. ‘This detailed study was made upon cross sections
of from five to ten representative leaves from each position of the trees.

All figures are made from photographs of camera lucida drawings of sections of
leaves. The leaves were taken from the various positions indicated below. Except
as indicated, all material was from the vicinity of Minneapolis, Minn.

Figs. 1-4. Acer saccharum. Fig. 1, Isolated tree. Leaves from south periphery.
Fig. 2, Isolated tree. Leaves from center of crown. Fig. 3, Forest tree. Leaves
from base. Fig. 4, Forest, seven-year-old tree. Leaves from base.

The trees were selected carefully so as to secure specimens typical of
the species. Permanent mounts of these sections were made by the
paraffin process. Care was taken so that the sections were cut from

548 HERBERT C. HANSON

the same part of leaves of like age and that the measurements were
made in typical parts of the sections so that no error would be caused
by the thickening due to fibro-vascular bundles.

Acer saccharum.—The study of an isolated tree showed that the
center leaves were on an average 38 percent as thick as the south
periphery leaves. This increase was caused mostly by the great
palisade development, the thickness of the palisade tissue in the center
leaves being about 25 percent that of the south periphery leaves. In
the center leaves the palisade made up 38 percent of the total thick-
ness, while in the south periphery leaves it made up 58 percent. The
thickness of the sponge tissue and upper epidermis was about one half
as great, and the lower epidermis three fourths as great in the center
leaves as in the south periphery leaves.

Great differences in structure were found in the leaves from the
two positions. The south periphery leaves had two layers of palisade
and these layers were far more compact than the single layer in the
center leaves. The sponge tissue was more compact, the bundles
and water storage tissue more abundant, the cells in the upper and
lower epidermis more regular, and the number of crystals greater in
the south periphery leaves than in the center leaves.

The weight of the green leaves per given area and the weight of
the water-free leaves in the center were 46 percent and 38 percent of
the weights at the south periphery. -The water content based on
green weight of the center leaves was 75 percent, and of the south peri-
phery leaves 52 percent.

Factor measurements showed that the amount of evaporation and
the rate of the wind in the center of the tree were respectively 67 per-
cent and 28 percent of the amounts at the south periphery. The
light within the crown was 0.0086, while at the south periphery it
was 1.00.

Less pronounced differences were found in the leaf-structure of
trees growing in the forest. The thickness of the leaves growing at
the base of the trees were 44 percent the thickness of the leaves at the
apex. The apex leaves were 77 percent the thickness of the south
periphery leaves of the isolated tree and the base leaves 92 percent the
thickness of the center leaves. The weight of the green leaves and the
weight of the water-free leaves of the base leaves were 47 percent and
35 percent the weights of the apex leaves. The light intensity of the
base leaves was 0.0076. Leaves were found growing in a light in-

HERBERT C. HANSON 549

tensity of 0.0024, fifteen centimeters above the ground. The thickness
of these leaves was 30 percent the thickness of the south periphery
leaves of the isolated tree. The palisade and sponge tissue were very
loose.

Tilia americana.—The total thickness of the center leaves of
isolated Tilia americana in Minneapolis was 52 percent the thickness
of the south periphery leaves. This difference was caused chiefly by

tog 8 ys
Of

Figs. 5-8. TJuzlia americana. Fig. 5, Isolated tree. Leaves from south peri-
phery. Fig. 6, Isolated tree. Leaves from center of crown (Lincoln). Fig. 7, Iso-
lated tree. Leaves from center of crown. Fig. 8, Forest tree. Leaves from base.

the great increase of palisade tissue, the center leaves having only
22 percent as much palisade as the south periphery leaves. The
palisade tissue composed 34 percent of the thickness of the center
leaves, and 81 percent of the other. The sponge tissue is changed to
palisade in the south periphery leaves. There are four layers, usually,
of palisade in the latter leaves and only one in the former. The cells
in the center leaves are larger, more irregular, more often funnel-
shaped, the air-spaces are larger and more numerous, the bundles and

550 HERBERT C. HANSON

water storage cells are more poorly developed, the crystals are less
numerous, and the side walls of the epidermal cells are more wavy.
The weight of the green leaves from the center was 58 percent the
weight of the south periphery leaves, the weight of the water free
leaves from the center 51 percent. The water content of the center
leaves was 65 percent, of the south periphery leaves 60 percent.

The amount of evaporation and the wind in the center were 77
percent and 65 percent of the amounts at the south periphery. The
light intensity in the center was 0.0353.

The center leaves of an isolated tree examined in Lincoln had
thicker leaves than the isolated tree in Minneapolis, although the
south periphery leaves were thinner. As the light intensity in the
center was 0.115, and the evaporation 59 percent that of the south
periphery; it seems probable that the increase in light intensity ac-
counts for the increase in the leaf thickness.

As in Acer saccharum the leaves on forest individuals of this species
are thinner than on isolated individuals. The apex leaves of forest
trees are about the same thickness as the south periphery leaves of
isolated trees, while the base leaves of the forest trees are 69 percent
the thickness of the center leaves of isolated trees. The decrease in
thickness may be accounted for by the increased humidity and the
lower light intensity, 0.00865, in the forest, as compared with the
isolated tree.

The second crop leaves from the base of the forest individuals were
thicker (31 percent) than the first crop while the apex leaves were
thinner (17 percent). The weight of the green leaves and the weight of
the air-dried leaves at the base were 29 percent and 20 percent that
of the apex leaves per given area. Although the difference in the
thickness of the first and second crop apex leaves was only 17 percent,
the structure of the second crop leaves was far more mesophytic as
seen in the amount of air space, number of bundles, and water storage
cells.

Quercus macrocarpa.—The center leaves of a well-formed, typical
isolated tree were 61 percent of the thickness of the south periphery
leaves. The increase in thickness was due chiefly to the increased
development of palisade tissue. The amount of palisade in the center
leaves was 37 percent of that in the south periphery leaves. The
amount of sponge in the center leaves was over twice as great as in the
south periphery leaves, showing that most of the sponge tissue had

LEAF-STRUCTURE AS RELATED TO ENVIRONMENT 551

become palisade. In the center leaves the palisade made up 46 per-
cent of the total thickness; the sponge, 34 percent; the upper epidermis
13 percent; the lower epidermis 7 percent. In the south periphery
leaves the palisade made up 75.8 percent of the total thickness; the

va)
ae } fo bo tag4

a aN ace
= KP a ee

Fig. 9. Quercus macrocarpa. Isolated tree. Second growth leaves from south
periphery.

Figs. 10-11. Acer saccharinum. Fig. 10, Isolated tree. Leaves from south
periphery (Lincoln). Fig. 11, Isolated tree. Leaves from center of crown.

sponge 7.8 percent; the upper epidermis 10.6 percent; the lower epi-
dermis 5.8 percent. The weight of the green leaves and the weight
of the water-free leaves in the center were 66 percent and 46 percent of
those at the south periphery. The water content of the former leaves
was 61 percent, of the latter leaves 43 percent.

The amount of evaporation and the rate of the wind in the center
were 75 percent and 50 percent of the amounts at the south periphery.
The light intensity at the center was 0.148. Another isolated tree,
having a more open crown, showed an increase in the south periphery
leaves of about 3 percent and in the center leaves of about 12 percent.

552 HERBERT C. HANSON

Both the south periphery (22 percent) and center (13 percent) second
crop leaves of this tree were thicker than the first crop leaves. The
increase in the compactness of the palisade tissue in the south periph-
ery leaves was especially noticeable.

The base leaves of forest individuals were thinner than the center
leaves of isolated trees. The light intensity in which these leaves
grew was 0.075 and the amount of evaporation 46 percent of that at
the apex. The apex leaves were slightly thicker than the south
periphery leaves of isolated trees. The thickness of the base leaves
was 47 percent that of the apex leaves and the thickness of the palisade
in the former was 30 percent that of the latter.

Quercus rubra.—An individual of Quercus rubra growing in an
Acer saccharum and Tilia americana forest was studied. The lowest
leaves, 7.3 m. high, were 67 percent the thickness of the apex leaves,
12.8 m. high. The difference in the amount of palisade in the leaves
from the two positions was not so great as in the trees so far noted.
The thickness of the palisade in the base leaves was 57 percent that of
the apex leaves. In the base leaves the palisade made up 44 percent;
the sponge, 30 percent; the upper epidermis, 16 percent; the lower
epidermis, 10 percent of the total thickness. In the apex leaves the
palisade made up 52 percent; the sponge, 26 percent; the upper epi-
dermis, 15 percent; the lower epidermis 7 percent.

The chief differences in structure were that the apex leaves often
had three layers of palisade cells, while the base leaves had two; the
palisade was more compact and composed of longer cells, and there
was a decrease in the air-space and an increase in the water storage
cells and fibro-vascular bundles in the apex leaves.

The weight of the green leaves at the base was 63 percent that of
the apex leaves, the weight of the water-free leaves 58 percent. The
water content of the former was 52 percent, of the latter 49 percent.

The amount of evaporation at the base was 58 percent that of the
apex, and the light intensity at the base was 0.0425.

Quercus alba.—The Quercus alba studied grew in the forest very
near the Quercus rubra. ‘The base leaves, 3 5 m. high, were 64 percent
the thickness of the apex leaves 9.2 m. high. The thickness of the
palisade in the base leaves was 38 percent that of the apex leaves.
The thickness of the sponge and the lower epidermis was less in the
latter than in the former. In the base leaves the palisade made up
33 percent; the sponge, 41 percent; the upper epidermis, I5 percent;

LEAF-STRUCTURE AS RELATED TO ENVIRONMENT 553

the lower epidermis, II percent of the total thickness. In the apex
leaves, the palisade made up 57 percent; the sponge, 25 percent; the
upper epidermis, 12 percent; the lower epidermis, 6 percent of the
total thickness.

Vy x a ha e : yh

is fem yor
og 1 tf [asus ]as

EET VOSA ue

Figs. 12-13. Quercus macrocarpa. Fig. 12, Isolated tree. First growth leaves
from south periphery. Fig. 13, Figure represents first and second growth leaves
from isolated tree, center of crown; and forest tree, base of crown.

Figs. 14-15. Acer saccharinum. Fig. 14, Isolated tree. Leaves from center
of crown (Lincoln). Fig. 15, Isolated tree. Leaves from south periphery.

As in the leaves already studied the chief differences in structure
consisted in the increase in palisade tissue and the greater compactness
of the tissue in the apex leaves. From two to four layers of palisade
were found in the apex leaves, while only one was found in the base
leaves. The cells in the former leaves were more prolate in shape
than those in the latter. The apex leaves had greater bundle de-
velopment than the base leaves. The weight of the base green leaves
was 58 percent, the weight of the water-free leaves 47 percent the
weights of the corresponding apex leaves. The water content of the
former leaves was 63 percent, of the latter 54 percent.

554 HERBERT C. HANSON

The amount of evaporation at the base was 57 percent that of the
apex, and the light intensity at the base was 0.010.

Acer saccharinum.—lIsolated individuals of Acer saccharinum were
studied in Minneapolis and Lincoln. The center leaves of the tree in
the former place were 66 percent the thickness of the south periphery

eee ee
LN Aa aseat a
tJ ie 3 Oe i

Figs. 16-17. Fraxinus pennsylvanica. Fig. 16, Isolated tree. Leaves from
south periphery. Fig. 17, Isolated tree. Leaves from center of crown. Fig. 18.
Quercus rubra. Forest tree. Leaves from base.

leaves. Most of this increase was caused by the palisade as the thick-
ness of the palisade in the former was but 48 percent thai of the latter.
In the center leaves the palisade made up 39 percent; the sponge, 32
percent; the upper epidermis, 16 percent; the lower epidermis, 13 per-
cent of the total thickness. In the south periphery leaves the palisade
made up 53 percent; the sponge, 26 percent; the upper epidermis,
II percent; the lower epidermis, 9 percent of the total thickness.

The center leaves had one layer of loose palisade while the south
periphery leaves had one or two layers composed of larger and more
compactly arranged cells. The entire structure of the south periph-
ery leaves was more compact and there was greater development of

LEAF-STRUCTURE AS RELATED TO ENVIRONMENT 555

bundles and water storage cells. The weight of the green center leaves
was 60 percent the weight of the south periphery leaves, the weight of
the water-free center leaves 42 percent. The water content of the
center leaves was 68 percent, of the south periphery leaves 55 percent.

Figs. 19-20. Platanus occidentalis. Fig. 19, Isolated tree. Leaves from south
periphery (Lincoln). Fig. 20, Isolated tree. Leaves from center of crown (Lincoln).
Fig. 21. Quercusrubra. Forest tree. Leaves from apex of crown.

The evaporation at the center was 65 percent that at the south
periphery, and the light intensity .038.

The studies of several trees at Lincoln showed that the south
periphery leaves of these trees were from 13 percent to 42 percent
thicker than the south periphery leaves of Minneapolis trees. The
increase was due to the increase of palisade chiefly. The structure of
the south periphery leaves from Lincoln was more xerophytic. The
center leaves were about the same in both places.

Acer negundo.—The thickness of the center leaves of an isolated
Acer negundo tree at Minneapolis were 79 percent that of the south

5506 HERBERT C. HANSON

periphery leaves. Most of this increase was due to the palisade as it
more than doubled in thickness. In the center leaves the palisade
made up 33 percent; the sponge, 45 percent; the upper epidermis, 12
percent; the lower epidermis, 10 percent the total thickness. In the
south periphery leaves the palisade made up 57 percent; the sponge,
24 percent; the upper epidermis, II percent; the lower epidermis,
8 percent of the total thickness. The south periphery leaves had two
layers of compact palisade, the center leaves one layer. The amount
of evaporation at the center was 67 percent that at the south periphery,
and the light was 0.082.

Ulmus americana.—The thickness of the center leaves of an isolated
Ulmus americana was 64 percent the thickness of the south periphery
leaves. The palisade in the former was 45 percent the thickness of
that in the latter. In the center leaves the palisade made up 35 per-
cent of the total thickness, the sponge 38 percent, the upper epidermis
16 percent, the lower epidermis I1 percent. In the south periphery
leaves the palisade made made up 50 percent of the total thickness,
the sponge 26 percent, the upper epidermis 15 percent, the lower
epidermis 9 percent.

The south periphery leaves were more compact in structure, the
cells were narrower and longer, the upper epidermis more regular, and
two layers of palisade were developed as compared with one in the
center leaves. The weight of the green center leaves was 71 percent
that of the south perinhery leaves; the water-free leaves, 55 percent.
The water content of the center leaves was 66 percent, of the south
periphery leaves 56 percent. :

The amount of evaporation at the center was 69 percent; the wind
57 percent the amounts at the periphery. The light intensity at the
center was 9.084.

Fraxinus pennsylvanica.—The thickness of the center leaves of an
isolated Fraxinus pennsylvanica was 63 percent that of the south
periphery leaves. The palisade in the former was 38 percent the
thickness in the latter. In the center leaves the palisade made up 35
percent the total thickness, the sponge 50 percent, the upper epidermis
8 percent, the lower epidermis 7 percent. In the south periphery
leaves the palisade made up 58 percent the total thickness, the sponge
30 percent, the upper epidermis 6.4 percent, the lowér epidermis 5.6
percent.

The south periphery leaves were frequently entirely palisaded;

LEAF-STRUCTURE AS RELATED TO ENVIRONMENT 557

the upper part consisting of three or four layers of very compact cells,
the lower part of four or five lavers of irregular cells; while in the center
leaves there was usually but one layer of palisade. The bundles and
water-storage cells, and the crystals were more numerous in the former
also.

The weighce of the center green leaves was 58 percent that ot the
south periphery leaves, the weight of the water-free leaves 28 percent.
The water content of the former leaves was 67 percent; of the latter,
33 percent.

The amount of evaporation at the center was 53 percent that at
the south periphery; the wind, 65 percent. The light at the center
was 0.015.

Celtis occidentalis L.—The thickness of the center leaves of an
isolated Celtis occidentalis at Lincoln was 63 percent the thickness of
the south periphery leaves. All of the sponge tissue in the center
leaves was palisaded in the south periphery leaves, so the palisade in
the former is only 31 percent that in the latter. The differences in
structure were again found in the compactness of the cells, the shape
of the cells, and the cystolithic cells were more abundant in the south
periphery leaves. The light in the center was 0.059.

Platanus occidentalis L.—The thickness of the center leaves of an
isolated Platanus occidentalis at Lincoln was 61 percenc the thickness
of the south periphery leaves. The palisade tissue in the center leaves
was 46 percent that of the latter. The palisade made up 32 percent
the thickness of the center leaves, the sponge 38 percent, the upper
epidermis 17 percent, the lower epidermis 13 percent. The palisade
made up 42 percent the thickness of the south periphery leaves; the
sponge 38 percent, the upper epidermis 12 percent, the lower epider-
mis 8 percent. The south periphery leaves had more compact tissue,
the cells were more prolate, although there was but one layer as in the
center leaves. The scalloped appearance of the cross section of the
south periphery leaves was caused by the greater bundle and water
storage tissue development as compared with the center leaves.

SUMMARY

1. The light intensity, as measured by the Clements photometer,
within the crown of 10 common broad-leaved trees was found in
August to vary from .0076 of full sunlight in Acer saccharum to .1132
in Quercus macrocarpa.

558 HERBERT C. HANSON

2. The evaporation, measured by the Livingston porous cup at-
mometers, was found to be from 1% to 244 times as great at the south
periphery as within the crown.

3. The temperature at the south periphery was usually but one or
two degrees higher than within the crown. }

4. The humidity, measured by cog-psychrometers, was usually
from I percent to 6 percent higher within the crown.

5. A wind of low velocity caused greater differences in the air
movement between the center and the periphery of the crown than a
strong wind. The wind was found to be from 114 to 8 times as strong
at the periphery as within the crown.

6. Transpiration experiments showed that the south periphery
leaves lose more water per unit area than the center leaves. In
Fraxinus pennsylvanica the south periphery leaves lost from 3 to 6
times as much as the center leaves; in Ulmus americana about 12
times as much. Even when the potometer containing south periphery
leaves is placed under similar conditions with the potometer contain-
ing center leaves it will lose more water per unit area.

7. The leaves from the periphery of the tree were usually more
deeply lobed, more prominently toothed, and smaller than the leaves
from the center of the same tree.

8. The water content of the leaves from the center of the tree was
always higher than that of the leaves from the south periphery. The
amount of dry material per unit area in the exposed leaves bears a
relation to tolerance. The dry weight of the leaves of the most tol-
erant trees is less per unit area than the dry weight of the leaves of the
least tolerant trees, as, leaves from Acer saccharum contain 1.029 gr.
of dry matter per unit area, while leaves from Quercus macrocarpa
contain 1.272>er.

9. The differences in the total thickness between the south periphery
and the center leaves on the same tree are usually greater than the dif-
ferences heretofore reported from leaves of mesophytic and xerophytic
forms of the same species. The leaves from the south periphery have
more palisade tissue, gréater compactness of structure, thicker epi-
dermis and cuticle than the leaves from within the crown.

This subject, the structural response of leaves of the same plant to
measured environmental factors, is so large that this paper can only
be considered as an opening wedge into further investigation. De-
tailed studies are needed on specific aspects, as transpiration, water
content, etc.

LEAF-STRUCTURE AS RELATED TO ENVIRONMENT 559

In conclusion I wish to acknowledge my indebtedness to Dr. J. E.

Weaver, who suggested this investigation and assisted in securing a
pari of the data obtained during the summers of 1915 and 1916. I
also wish to express my appreciation of the encouragement offered by
Dr. Raymond J. Pool and of the facilities afforded by the depart-
ment of botany at the University of Nebraska.

THE UNIVERSITY OF NEBRASKA

BIBLIOGRAPHY

. Adamson, R.S. On the Comparative Anatomy of the Leaves of Certain Species

of Veronica. Journ. Linn. Soc. Bot. 40: 247-274. 1912.

. Areschoug, F. W. C. Ueber die Bedeutung des Palisaden-parenchyms fiir die

Transpiration der Blatter. Flora 96: 329-336. 1906.

. Baumert, K. Experimentelle Untersuchungen tiber Lichtschutzeinrichtungen

an griinen Blattern. Beitr. Biol. Pflanz. 9: 83-162. 1907.

. Bergen, J. Y. Transpiration of Sun Leaves and Shade Leaves of Olea europea

and Other Broad-leaved Evergreens. Bot. Gaz. 38: 285-296. 1904.
Relative Transpiration of Old and New Leaves of the Myrtus Type.
Bot. Gaz. 38: 446-451. 1904.

Concavity of Leaves and Hlumination. Bot. Gaz. 38: 459-461. 1909.

. Bonnier, Gaston. Cultures expérimentales dans la région mediterranéene pour

le modifications de la structure anatomique. Compt. Rend. 135: 1285-1289.
1902.

. Boodle, L.A. The Structure of the Leaves of the Bracken (Pteris aquilina Linn.).

Journ. Linn. Soc. Bot. 35: 659-669. 1904.

. Brenner, W. Untersuchungen an einigen Fettpflanzen. Flora 87: 387-439.

1900.

. Brenner, W. Klima und Blatt bei der Gattung Quercus. Flora 90: 114-160.

1902.

. Chrysler, M. A. Anatomical Notes on Certain Strand Plants. Bot. Gaz. 37:

461-464. 1904.

. Clements, E. S. The Relation of Leaf Structure to Physical Factors. Trans.

Amer. Micr. Soc. 26: 19-102. 1905.

. Copeland, E. B. The Variation of Some California Plants. Bot. Gaz. 38:

ZOL-A26.> 1904;

. Eberhardt, M. Action de l’air sec et de l’air humide sur les végétaux. Compt.

Rend. 131: 193-196, 513-515. 1900.

. Ewart, A. J. The Influence of Correlation upon the Size of Leaves. Annals of

Botany. 20: 79-82. 1906.

. Ganong, W. F. The Living Plant. P. 258. 1913.
. Haberlandt, C. Physiological Plant Anatomy. P. 388. I914.
. Hansgirg, A. Phyllobiologie nebst Uebersicht der biologischen Blatttypen von

ein-und-sechzig Siphonogamen-familien. Leipzig. Pp. 486. 1903.

. Harshberger, John W. The Comparative Leaf Structure of the Sand Dune

Plants of Bermuda. Proc. Amer. Phil. Soc. 47: 97-110. 1908.

560 HERBERT C. HANSON

20.

PA

Pala

23.

24.

The Comparative Leaf Structure of the Strand Plants of New Jersey.
Proc. Amer. Phil. Soc. 48: 72-89. 1909.

Hasselbring, H. Effect of Shading on the Transpiration and Assimilation of the
Tobacco Plant in Cuba. Bot. Gaz. 5'7: 257-286. I914.

Herriott, E. M. On the Leaf Structure of Some Plants from the Southern
Islands of New Zealand. Trans. N. Z. Inst. 38: 377-422. 1906.

Hesselman, H. Zur Kenntnis des Pflanzenlebens schwedischer Laubwiesen.
Eine physiologisch-biologische und _ pflanzengeographische Studie. Beih.
Bot. Centralbl. 17: 311-460. 1904.

Livingston, B. E., and Brown, W. H. Relation of the Daily March of Trans-
piration to Variations in the Water Content of Foliage Leaves. Bot. Gaz.
53: 309-330. I9I2.

. Lubimenko, M. W. Sur la sensibilité de l’appareil chlorophyllien des plantes

ombrophiles et ombrophobes. Rev. Gén. Bot. 17: 381-415. 1905.

. Oltmanns, F. Photometrische Bewegung der Pflanzen. Flora 79: 232, 233.

1892.

. Raunkiaer, C. Nogle Iagttagelser of Fors6g over Aarsagerne til Palissadece-

lernes Form og Stilling. (Sur les causes qui déterminent la forme et l’orienta-
tion des cellules palissades). Bot. Tidsskr. 27: 293-311. 1906.

. Renner, O. Beitrage zur Physik der Transpiration. Flora 100: 451-457. IgI0.
. Sampson, A. W., and Allen, Louise M. Influence of Physical Factors on Trans-

piration.. Mann: Bot: Stud.74': 33-59... 1900:

. Starr, A. M. Comparative Anatomy of Dune Plants. Bot. Gaz. 54: 263-305.

LOI.

. Wiegand, K. M. The Relation of Hairy and Cutinized Coverings to Trans-

piration. Bot. Gaz. 49: 430-444. I9I0.

. Wiesner, J. Anpassung des Laubblattes an die Lichtstarke. Biol. Centralbl.

Tir .15.4 S90:

. Zon, R., and Graves, H. S. Light in Relation to Tree Growth. Forest Service

Bulletin 92. Washington, I9gII.

tH) pee Key

"id

Nee

“BOTANICAL ‘SOCIETY OF

EDITORIAL COMMITTEE

ie c. ‘Newcomnr, ‘Aiton -in- Chie,
«University of clean

ney Stuart Ginae. Bidiacs Manager
- Brooklyn Botanic Garden lag

_Tnvine W. Barry,
ey, Institution

hin

oe ‘i. BARTLETT, - heart ne sae
Bice of Michigan eee

AL Gs JouNson, Sides ofl Waeonsin
"Representing 4 meniean Rt vicarage Society)

*

a ae scriptions.

nies The: pages ‘of. the foucuht are sonene to » members’ of :
iS America, | or candidates nes for eenierne Ae

dd be charged: at cost rates for illustrations and tabular platter ine

wore

r5. 15 cents fa

| date of dating ~The puviaters will supply missing zur
have been lost in the mails, | creas <

an

; yn Botanic Garden, Brocka, te or mae North i Queen Stet, Lancet

AMERICAN
JOURNAL OF BOTANY

———.
$+

Vor. 1V DECEMBER, I9I7 No. 10

iii PHY TOGEOGRAPHY OF MANOA. VALLEY,
HAWAITAN ISLANDS

VAUGHAN MACCAUGHEY

The present paper represents the first effort, in the long history
of botanical exploration in the Hawaiian Archipelago, to make a
detailed ecological survey of a representative Hawaiian phytogeo-
graphic area. Taxonomic lists and descriptions of new species com-
prise the bulk of the scientific literature dealing with the Hawaiian
flora, and in all of this material there is a conspicuous absence of
physiographic and ecologic data. The present paper is based upon
field observations extending over a residence of nine years on the
island of Oahu, of which time four years have been passed in Manoa
Valley itself. The writer has repeatedly visited all portions of this
beautiful and historic valley, and has conducted many collegiate field
excursions to the numerous points of prime botanic interest.

The writer has availed himself of all accessible records. The
nomenclature followed has been chiefly that of Hillebrand’s monu-
mental Flora of the Hawaiian Islands (1888). Although this nomen-
clature is somewhat obsolescent, it is in common usage in the island
literature, and it was deemed inadvisable to cumber too greatly these
pages with revised names of familiar plants. In numerous instances
however, the modern taxonomy has been introduced.

The College of Hawaii is situated in Manoa Valley, near Honolulu.
This valley is the immediate natural background of the College and
its botanical instruction. Manoa is a representative ecologic area of
the Hawaiian mountains. It presents a very clearly defined series of
life zones, both in vertical and horizontal planes. It is typical of
many valleys in the Hawaiian Islands, and in other parts of the
Polynesian Pacific. The phytogeography of Manoa Valley epitomizes
that of any similar physiographic region in the archipelago.

561

562 VAUGHAN MACCAUGHEY

RERRERER

'
Cress SECTION C~C |

3

Cross SECTION B~B a

1
CROSS SECTIONA~A 8

RERREERRERR

B

LonGa/TvainaL Secrion D~D

Scales PROF/LE MAP
Horizontal distonce 1° = 1000F% OF
MANOA VALLEY

Fic. 1. Vertical Cross and Longitudinal Sections of Manoa Valley / .

1. West lateral ridge; transition region; steep wall into hygrophytic valley-
head.

2. East lateral ridge; mountainward portion, covered with rain-forest; east-
ward slope is into the head of Palolo Valley.

3. West lateral ridge; Mount Tantalus; sloping into middle valley floor.

4. East lateral ridge; apex of foothill.

5. West lateral ridge; Roundtop; sloping into lower valley floor.

6. East lateral ridge; Manoa-Palolo Foothill, about midway between plain and

7. Summit Ridge, near Mount Olympus.

8. Puu Pueo, a median ridge lying along the central axis of the valley.

9. Middle portion of the valley floor, near the point of union of the main tribu-
taries of Manoa Stream.

10. The lower valley floor.

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 563

Manoa is situated on the island of Oahu, in the vicinity of the city
of Honolulu. Residential districts lie along portions of the mouth of
the valley and lower western slopes. Much of the floor is occupied
by agricultural lands—iaro, bananas, vegetable gardens, etc. Oahu
is third in size among the Hawaiian islands. It is 46 miles long and
25 miles wide, with an area of 598 square miles. It is topographically
distinguished from the other islands by being composed of two elongate
mountain ranges, Waianae and Koolau. These are of great antiquity,
deeply eroded, and give evidence of numerous and extensive elevations
and subsidences.

The Waianae Range, lying on a NW-SE. axis, is about 20 miles
long. Its highest peak, Ka-ala, is 4,030 feet high; this is the highest
point on the island. The highest point in the archipelago is Mauna
Kea, on Hawaii, 13,825 feet. The Koolau Range, in which Manoa
is carved, lies to the northeast of the Waianaes, parallel with the latter,
at a mean distance of eighteen miles. The Koolau Range is 37 miles
long, and is the longest range in the archipelago. It is low, its mean
elevation not exceeding 2,000 feet. The highest peak, Kona-hua-nui,
rises to 3,105 feet, and lies at the head of Manoa Valley. The range
is deeply sculptured by subaerial erosion. There are about fifty major
valleys, with numberless ravines and lateral gullies. Manoa is one
of the largest of the major valleys.

Manoa is a well-matured valley, with broad flat floor and slightly
expanded head. Measured from its mouth or portal (using the
100 feet contour as a base-line), an airline to the crest of the summit
ridge is 3.4 miles long. Its width, measured by airline from one
lateral ridge-crest across to the opposite ridge-crest, varies from 1.2
miles at the portal to 2.2 miles at the head. Like many other of the
larger Hawaiian valleys—Kalihi, Kahana, Iao, Pelekunu, Halawa,
Waipio—the head of Manoa is a constantly expanding amphitheater
of erosion. The valley widens progressively from portal to head,
at the rate of about 5—6 percent.

The Koolau Range lies along a NW.-SE. axis. All the valleys,
of which Manoa is one, that deeply furrow its leeward flanks have a
dominant southwesterly exposure. The trade winds, which blow
almost continuously through a major portion of the year, come from
the northeast. The leeward valleys are thus protected from the trade
winds by the mountain wall. The maximum of the torrential precipi-
tation that results from the rising of the moisture-laden trades over

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 565

the mountain rampart falls, not upon the summit crest, but just to
the leeward. The heads of the leeward valleys thus receive Oahu’s
maximum precipitation. The annual average for this is about 150
inches, whereas the precipitation along the crest itself is about 100
inches. Manoa is known locally as a very rainy valley.

Rainfall has been a dominant factor in forming the valley and
sculpturing its walls. At present it is the controlling factor in the
distribution of the plant life of the valley. The following data, sup-

fe
: Col beer oP ie slay Wg al egies ee roe Cer ae as
Locality | 2 S ie S 2 cI 2 s & iS 3 i
a A

Honolulu, on coastal plain, |
pret D A bee Scat Pi seed es EO Se leon ol he Seer a te ee oe At
Manoa, middle of yalley,el. 300 ft.’ 7| 7 |8-|'8 | 7| 6) 7| 8) 8 \-7 1.9) 9] 90
Manoa, wpper floor, el. 300 ft... ..| 22| 3 |. 11) 9. | 28| 12) 13} 19) 9 | 9 | 17| 24|176
Mount Tantalus, el. 1,360 ft...... Bi-8-|°9 29 1. 8) 78 ol 8) 9. |. Siabre | 06

Fic. 2./ Contour Map of Manoa Valley

1. Nuuanu Valley Gap; at the head of the valley, cutting completely through
the Koolau Range.

12. Ohu-ohi Amphitheater, on the windward side of the Koolau Range.

13. Mount Olympus. i

14. Mount Kona-hua-nui, with two summit peaks.

15. Mount Tantalus.

16. Round top Hill; Sugar-Loaf les directly mountainward, between the
former and Tantalus.

18. The Manoa-Palolo foothill.

19. Apex of the Manoa-Palolo foothill.

20. Transition region on the east lateral ridge.

21. Mountainward portion of the east lateral ridge.

22. Topographic transition region on the west lateral ridge.

23. Mountainward portion of the west lateral ridge.

24. Zone of precipices or palzs that bound the upper valley head.

25. Zone of the hanging valleys; in addition to the five or six large hanging
valleys there are numberless small ones.

26. Hanging valleys in the flanks of Mount Olympus.

27. The lower valley floor.

28. Manoa stream, crowded against the foot of the east ridge.

29. The middle portion of the valley floor.

30. Puu Pueo.

31. The upper valley floor.

32, 33, 34, 35. Talus zone and valley walls.

36. Windward wall or precipice of the Koolau range.

566 VAUGHAN MACCAUGHEY

plied by the Hawaii Section, U.S. Weather Bureau, shows the FLUE
in various parts of the valley, for 1916.

Translating these data into terms of ecologic zones, the approximate
annual rainfall is as follows:

Coastal:plain, seaward'of Manoa valley... 552. 41 inches
Middle ofvalley..ci chi acannon i OOMmme
Uppersvalley: floor 3.45). 2 3.8sg:ecupe eye cee Pee eee 17634

. Lower dorest/Zone. soya ees ee ee ee ee LOG ts

FIG. 3. 7 Map of the eastern end of Oahu showing the relation of Manoa valley to
the adjacent land areas. |

/

In general the valley becomes progressively hygrophytic as one
advances toward the head, and conversely, progressively xerophytic
as one approaches the sea.

In Luakaha, a region in upper Nuuanu Valley, and separated from
Manoa by only a narrow ridge, the annual rainfall is about 200 inches
(196.99).

The U. S. Hygrographic Survey maintained a rain-gauge on the
summit of Mt. Olympus (2,450 feet) for a period of 67 days, and

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 567

recorded a daily rainfall of .20 inch. This would be equivalent to
an annual precipitation of about 73 inches, a figure somewhat lower
than the general average for the rain-forest.

Shreve! gives the annual rainfall for three stations in the Jamaica
rain-forest as 105.70, 113.85, and 168.02 inches respectively. This
corresponds closely with records for the Hawaiian rain-forests, as
does his statement that ‘‘there is no other form of precipitation than
rain, hail and snow being unknown, although the former occurs at

Fig. 4. Niew of Manoa Valley, from a lateral ridge. Shows plainly the lower floor,
talus zone, wall and eastern foot hill. /

rare intervals in the lowlands. The frequency of showers too light
to register 0.01 inch is high, and they are not without influence on
vegetation. Although the number of rainy days is high and the
frequency of light showers is high, yet the bulk of the annual rainfall
is registered during the prolonged downpours. . . . Dew is formed
abundantly in open situations on clear nights at all seasons of the
year.” :

-1Shreve, Forrest, A Montane Rain-Forest. Carnegie Institution of Washing-
ton, I9I4.

568 VAUGHAN MACCAUGHEY

Although somewhat sheltered from the direct mechanical effects
of the trades, Manoa and the other leeward valleys of its class are
exposed to the periodical kona or southerly storms, which usually
occur during the late winter and early spring (January to April).
The kona storms are often characterized by heavy winds and excessive
rainfall.

The southwesterly exposure of Manoa shuts off its head from a
considerable portion of the morning sunlight, and gives prominence
to the afternoon heat and light. Manoa is much sunnier and warmer
than are the narrow, windy, northerly facing valleys of the windward
Koolau slopes. This climatic difference is sufficiently great to be
reflected in the respective floras of these two types of valleys.

1. THE REPRESENTATION IN MANOA OF THE HAWAIIAN E@OLOGIC
ZONES

In the Hawaiian Archipelago there are numerous well-defined
ecologic zones. The representation of these life-strata in the Manoa
region may be indicated as follows:

1. LITTORAL. a. Humid hitoral; windward.

b. Arid or semi-arid littoral; leeward. ‘The littoral of that
portion of the coastal plain which lies to the seaward of Manoa
Valley is of this type.

2. LowLanps. Up to 1,000-1,500 feet; with humid and arid sections,
depending upon relation of topography to trade winds, and
distance from interior mountains. In Manoa Valley the low-
land proper (valley floor) lies well below the 500-foot contour;
in early times the lower forest zone came down to this level.

3. THE Forest ZONE. a. The Lower Forest; 1,000—-2,000 feet; with
humid and arid sections. In Manoa this zone lies between 500
and 1,200 feet, and is almost wholly of hygrophytic or semi-
hygrophytic character, although some xerophytic forms do occur.

b. The Middle Forest; 1,800-5,000 feet; variable with humid
and arid sections on the various islands. In Manoa this zone
is typical Oahuan rain-forest; highly hygrophytic, and very
rich in endemic forms. Owing to the low elevation of the Oahu
mountains, this is the highest zone, and

c. The Upper Forest; 5,000—-9,000 feet, is restricted to the
high mountains of Maui and Hawaii.

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 569

4. THE SUMMIT REGIONS. a. Xerophytic Summits; 9,000—-14,000
feet; high mountains of Maui and Hawaii.
b. Hygrophytic Summits; 4,000—-6,000 feet; peaks rising into
the cloud zone, Kauai, Waianae, East Molokai, West Maui,
Kohala.

2. ECOLOGIC (ZONES OF -MANOA

The main ecologic zones in Manoa Valley are:
1. The Valley Floor:
a. The Lower Floor (near portal).
b. The Upper Floor (near head).
2. Manoa Stream and its Tributaries.
3. The Talus Zone.
4. The Valley Walls or Lateral Ridges:
a. The East or Manoa-Palolo Ridge,
1. Foothill.
2. Transition region.
3. Mountainward region.
b. The West or Mt. Tantalus Ridge,
1. Foothill (Roundtop).
2. Transition region.
3. Mountainward region. .
. The Kukui Zone, Ravines, and Precipices.
. The Zone of Koa and Lehua.
. The Hanging Valleys; Rain Forest.
. Summit Ridges and Peaks:
a. Olympus.
6b. Kona-hua-nul.
Topographic, edaphic, climatic, and biotic factors differentiate
more or less clearly these zones from one another. On the basis of
water, the grouping would be, as numbered above:

CON DU

fiydrophytic.:.2. Wiesophytice use. 2 oor Tae
Ebvecopmytic.. 210, 4023, 4023, 5, 6,7, 8. Xerophytic. 7. 2.) 4..%- 1a, 3, 4ai, 41.

On the basis of elevation:

ENDOVEI? OOONtIT Sy). sc od tees 8a and b.
Between I,000-3,000......... Te Between 1,200-1,700.......... 4a2, 4b2.
1,000—1,400 v.20. Ge 6. I,000—2,000.......... 4a3, 463.
200710008)... eso. 55: NOO= 8 SOO 8 es 2 nd ae

50-1. OOO seats oa ee: A4al, 4b1. 5 Os OO se ee erate. Bayes

570 VAUGHAN MACCAUGHEY

On the basis of mean temperatures:

Notably wat, ge... chee. 1, 400, 20 COOL gtr aan RO eh a 4a2, 402, 5.
Warm (2.2... Aes Phe An pa 2. Notably cool... b.-2 a2 403, 402,67, 8.

3.. THE VALERY BPEOOR

The floor of Manoa is conspicuously broad and flat, much more
so than are the floors of the valleys immediately adjacent to it, Makiki,
Pauoa, Palolo, and Waialae. This flatness may be considered as one
of the evidences of the maturity of this valley.

The designation ‘“‘floor’’ comprehends the region lying below the
300-foot contour; roughly an area 2.0 miles by .75 mile. It is com-—
posed chiefly of ‘mountain wash,” a heavy, dark reddish-brown, fine-
textured, adobe soil, that has been washed down from the surrounding
basaltic ridges and spread out as a deep blanket in the valley basin.
Along the lower western slopes are extensive deposits of volcanic
ash and cinders. The thickness of the soil bedding is not known;
along the center of the valley it must be very deep, perhaps hundreds
of feet. The red-brown adobe soil is fertile, stiff and intractable in
cultivation, and exceedingly retentive of moisture. When it becomes
dry to any considerable depth, as during the infrequent droughts, it
cracks conspicuously. The cracks are I to 4 inches wide and 12 to
40 inches deep. :

From an ecologic standpoint the valley floor may be transversely
divided into two regions, the floor of the lower valley, 1. e., near the
portal, and the floor of the upper valley, near the head. The lower
floor comprises the area from the portal up to the point at which
Manoa Stream diverts from the middle of the valley. The upper floor
continues from this latter point to the region beyond the bifurcation
of the floor at Puu Pueo. The lower floor is contrasted with the
upper floor by greater xerophytism; more brilliant illumination;
higher temperatures of air and soil; less surface water in the form of
streams, pools, and springs (although more in the form of irrigated
taro patches); more volcanic material such as surface lava, cinders,
ashes, etc.; and smoother contours. The upper floor has a higher
percentage of indigenous vegetation, and in former times was wholly
covered by the lower forest zone, as will be described in another
section. | |

The valley floor is principally occupied by introduced plants,

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 571

including both weeds and economics. The native vegetation is poor
in species and quantitatively insignificant. This condition is in
striking contrast with that of the rain-forest, only a few miles distant,
where the vegetation is almost wholly endemic or indigenous, and
where the introduced element is practically negligible.

Some of the larger and dominant plants of the valley floor (aside
from those actually under cultivation), are: Prosopis juliflora, Opuntia

Fic. sf View of Manoa stream and east valley wall. Looking toward head of
valley, which may be seen faintly through the rain.” Trees in mid-ground are Kiawe,
Prosopsis julifiora.

megacantha, Leucaena glauca, Lantana Camara, Psidium Guayava,
Xanthium strumarium, Ricinus communis, Indigofera Anil, Malvastrum
tricuspidatum, Cassia spp., Sida spp., Acacia Farnensiana, Ipomoea
spp., Commelina nudiflora, Crotalaria spp., Eugenia Jambolana,
Stachytarpheta dichotoma, Solanum Sodomeum, etc.

Manoa has been inhabited by the native Hawaiians since very
early times. Much of the lower floor was occupied by their tiny
plantations or kuleanas. The kalo or taro (Colocasia antiquorum

SV VAUGHAN MACCAUGHEY

Schott) was the principal crop, and was raised in small irrigated fields
or Joi. The water from these fields was skilfully diverted from Manoa
stream by a primitive but highly efficient system of ditches. An area
equivalent to several square miles was occupied by the kalo fields.
Much of this kalo land is in cultivation today, although the industry
has passed largely into the hands of Orientals.

Fic. 6. Opuntia megacantha, a dominant xerophyte of the Manoa lower valley
floor and foot hills.

Other crops raised by the primitive Hawaiians, and continuing
today in small patches here and there, are

SWeet. potato... ee GIG A Eey ere Ipomoea Batatas
Native banands 2 4.5. NIA aus tea Musa sapientum
DUSAL CAN sae ie eee KOs Sy ee geen Saccharum officinarum

Mention may be made of the eleven avian species that are dis-
tinctive of the valley floor and walls. Six of the introduced species
are common and of considerable phytogeographic significance, as.
they are abundant carriers of weed seeds and fruits.

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS

Ave Ss, OF THh VALLEY -FEOOR

ENDEMIC
Short-eared owl, Aszo accipitrinus Sandwichensts Bloxam. Pueo................ R
Ileamaidanecoot, i uiica aiat. Peale, Alaé keokeo. 0 cee ce a ew dds eee R
Hawaiian gallinule, Gallinula galeata Sandwicensis Streets. Alaeula............ R
INDIGENOUS
Black-crowned night heron, Nyctocorax nyctocorax griseus Bodd................. F
NATURALIZED
nice bird, Mania mesora punctaia Temm.. .... 0... 2. ce ee ae ee eens ec
tledichispa sho warleasser- COMeslicUus. Lic o.oo s.o% ag Ss ices si emi a A eee Vo ba oles Cc
by icllag eR MMA GTC OLREPOSUTISIIS Langa ce bok gee elk DAG Sha eb bed eda. wae te €
SANGRE CUO OTC IU STS Mle eth hse wot ao suc gan 4 vow athe dn he WS ade Se kao ated A on F
Ghmese reed warbler, Trochalopterum canorum Li... oe ele eee ees ‘S
Cinmesescurcie dove, Juriur Chimensts SCOP...) 6. ek ees oe ee ne dal nd none Cc
Australian: alac, Alae 22, Porphyrio melanotus Newton... . 0... .06. eee ee ew R
Explanation of Symbols
H—herbaceous; annual. V—liana habit.
S—shrubby. C—common.
A—arborescent. F—frequent.
P—perennial. R—rare.
Pier ReSeNLATIVE PLANTS OF THE VALLEY FLOOR
Group I. ENDEMIC Group II. INDIGENOUS
la, Hydrophytes Ila. Hydrophytes
Hibiscus Youngianus Gaud...... HSF Commelina nudifiora L. 6... ee. BPC
Kyllingia monocephala Rottb.....HPC
Ib. Mesophytes
Nama Sandwicensis Gray....... HF Ifd.’ Mesophyies
Sicyos cucumerina Gray........ HR Andropogon contortus Roem. &
‘‘ pachycarpa Hook. & Arn... HR Sehr tc. circ pal, oe eee eae HPF
Solanum aculeatissimum Jacq... .SF Caesalpinia Bonducella Fleming. .VSPF
Chrysopogon aciculatus Trin......HPC
Ic. Xerophytes Cyperus pennaiusstame., si... 3. HEC
Abutilon incanum G. Don.......HSF Ee bona-nox L............ HPVF
Chenopodium Sandwicheum Moq..HSF ; BONED hy Ua L...... 0... HVC
Erythraea sabaeoides Gray....... HF Nicandra phy sal oides Gaertn. . .. HF
Jacquemontia Sandwicensis Gray .HVF Panicum p Bens Trin.......... HPC
Plambaso Zeyianie@ Var. 202). ie SPE
Ge Paras Wrkstroemza foetida var. Oahuen-
Cuscuta Sandwichiana Chois.....HVF SC eggs fae re
IIc. Xerophytes
Ie. Pteridophy'es Boerhaana dvpusa Lies. ek HSF
Ophioglossum concinnum Brack...F Hleusine Indica Gaertn... ...... Hee

Ske)

574

Erythrina monosperma Gaud...
Ipomoea tuberculata Roem. &
Schult tit ce et ace ees

S100 jaan NVA. cee de ee oe:

" hombuolid Wins 4 ee
Tephrosia piscatoria Pers........
TVD UIESICUSEOTOCS Na Ga Nen ae

IId. Pteridophytes

Ceropteris calomelaena Link.. .
= ochracea Robins.......
Dryopteris propinqua Gilb.......
Microlepia strigosa Presl.........
Nephrolepts cordifolia Presl......
+ exaliata Schott... ..<.

Group III.
PRIMITIVE HAWAIIANS

Illa. Hydrophytes

Alocasia macrorrhiza Schott......
Colocasia antiquorum Schott...

IIIb. Hvgrophytes
Artocarpus incisa Lee... seit

Eusenta:. Malaccensis 1, (2... 2
VGISGRSO DICH UNG Vacate, Lua

IIIc. Mesophytes

Calophyllum inophyllum L.......
Cocos NuUcierals s5cs53, San he
Cucurbita maxima Duch... 45.

TPOMOLA BOGS Neti ee eee
Lagenaria vulgaris Ser.... 6.22.2.
Saccharum officinarum L........
Thespesia populnea Correa......

IlId. Xerophytes

Morinda citrifolia L.........
Pandanus odoratissimus L

© 0 e.Le 28 8

i

HPVC
ne
HSC

Halss®

SRE
SEE
SC

INTRODUCED BY THE

VAUGHAN ,MACCAUGHEY

Group IV. PLANtTs NATURALIZED
SINCE THE ADVENT OF |

EUROPEANS
IVa. Hydrophytes
Arundo: Donax i. 2) ee HPF
Coix lachryma Lio... HPC
Panicum barbinode Trin......... HC
sagiiaria sagititfolia L. 2) eee HC

TV Oe, yerophytes

Coffea arabica Llu ee AF
Hucenia Jambos \.i,1, 2 se ee Ce
Physalis, Peruviang Vee HPC

IVc. Mesophytes
1. Trees and Shrubs

Bixa Orellana Lic... ee F
Carica: Papaya Li int F
Cassia chamaccrista tu. ee F

“ ~ Jaevigata. Willde.32 alee F

| OCCIGEHIGLIS NG eek SL Cc
Cestrum diurnum L............. F
CLUS SDDao ese oe ae F
Eugenia Jambolana Lam........ c
Yatropha Curcas are) 2 eee R
Leucaena glauca Benth.......... AC
Melia Azedarach L............ AC
Mimosa pudical., 3 HPSR
Mammea Americana La)... sae AF
Psidium Catileyanum Sab....... AC

Guayava L. and vars... .ASC

RICMUS CONMUNISAL. SAC
SPOndies CulCTSMu i ee ee eee AF
Terminahia Catappa L.......... AC

2. Herbaceous Perennials
Bambusa vulgaris Schrad. &

Wendl so eee F
Canna indica 2 ee C
Cajanus Indicus Spreng... LG
Crotalaria fulva Roxb... ae C

rk soliang Andris. oes C
" spectabilis Roth... =. (S
Cynodon Daciyion Pers.........- C

Cyperus. rotundus Ves eee ee

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 575

Echinochloa Crus-galli (L.) IVd. Xerophytes
Beauv.. Deere eae Beret eee c 1. ‘Lrees and: Shrubs
Medicago ap tculata Willd Se ak a Acacia Farnensiana Willd....... (S
e se aee Willd... ... a Cassia occidentalis eo see ee C
£ ahd ee rane ge eh ats Casuarina equisetifola L........ F
a antertexta NII, 22 ee. FF Eucalyptus spp C
: 5 a. lupuling Le...s ss... S Indigofera ANU Misr us ee eee €
Mirabilis Jalap al... ee pwr en et C Nicotiana glauca Grah.... 2.3. F
Panicum barbinode RiGee. ae Oi ih Sela oe C
Paspalum conjugatum Berg... .. C Pehocnise Gdetyiifer ae. ames ee ea
Taraxacum officinale Weber......R Penis ee C
Wintcrod alata Bojvandwat) * 9 © =
auramiiaed Ite oe oe F 2. Herbaceous Perennials
Verbena Bonariensis L.......... C Abunlon-=sop C
Amaranthus spinosus L..........€
3. Annual Herbs Desmodium uncinatum DC,......C
Asclepias Curassavica L......... F ey ans hips ee ae anita at
Bothriospermum tenellum F. & M..F Ph ss al peace pes ee C
Brassica campesiris L....... 2. R p Te Poe ae ite os R
Cubhea hyssopifolia H. B. K.....R a BSBA d L. Pee bot ae F
Erigeron albidus Gray.......... F Pec L 3 aM ae C
e Cangdensts Vn. ce es ‘S lie) 3 aa en ae
Euphorbia geniculata Ort........ ‘Ss SP Aae pa ies yee Vahl. . -
‘4 Witntifera Lic s,s C anthium Strumarium L........
Fleurya interrupta Gaud......... R A eters
Franseria tenuifolia Gray....... F ss ate se : C
Mailva rotundifolia L............ 6. led BA ety i Bae eae .
Onairscornicuiata Vi... eke as F Bi ie nn aa aie hs C
SPIO SO i). Stuer Meeeeaet tee os.
ee ns a . Chaetochloa verticillata Scribn.....C
Peucedanum graveolens Benth. & Crenenodnnalun C
Fook. .:... an ee F tk hbrdun Leta E
PVOMEACOMIGION Lie ca eee R rr Wire hed ee F
Stegesbeckia orientalis L......... F DOL ee i
Eleusine Aegyptiaca Pers........ C
4. Vines ay Indica’ Gaettiis... 60 seas Cc
. Emilia flammea Cass............R
ee p Ss ee : Erodium cicutarium L’Her....... F
isla me eens Dele Pena ire te ee Euxolus viridis Mog..........-- Ee
Cardiospermum Halicacabum L...C Gynandropsis pentaphylla DC... .C
Ci erodendron fragrans Vent...... F Malvastrum tricuspidatum Gray. .C
Clitoria Ternatea L............. F Porinlaca ‘oleracea Wie te. Cc
Convoluulus 2 2) Sc C Raves A CClOCell Gain saree eoleee F
DolichostLavao le ee G Senebtera didyma Perse Cte
Ipomoea chryseides Ker-Gaul.....F Sonchus oleraceus L...........4% C
WEA CIC POCLIOG, lanes ois oes: CG SLOENYS OVDEH STS Neen een ree: F

EQSSUNOTG fOciida Mey eN eis ess F Vernonia conyzoides lo........... (:

576

VAUGHAN MACCAUGHEY

It will be noted that this list, which includes practically all of the
important species of this region, comprises the following groups:

Species. Hydrophytes... 7 9 a 9

Endémicws oie osha ee aes Ii’. Hygrophytes.....: ee 6
Indigenous.) 00.0... cee ee oe 29 ' Mesophytes:.... 070/240 82
Introduced by primitive Hawanans. 15. Xerophytes)...... 7)... 54 56
Introduced since the advent of Vines or lianas ... (22:5. ee 18
FE MPOPEATS ori. secure cis enter Dame 115. , Pteridophytes.’;. 0.3 eee PS ameeatt 8

6. MANOA STREAM

The surface drainage waters of Manoa escape as a single small and
fluctuating brook, known as Manoa Stream. A very considerable
percentage of the Manoa drainage makes its way to the sea through

-

Pic. 7. / Manoa stream near its mouth. The trees are Prosopsis julifiora. In
the distance is the west lateral ridge, with Round Top and Tantalus showing dis-
tinctly.

f
/

subterranean channels; this is a condition universal throughout the
islands. Manoa Stream is fed by numerous tributaries, which enter

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 577

the head of the valley over beautiful waterfalls. There is also a series
of springs along the foot of the cliffs at the head of the valley, which
contribute their waters to the stream. Other springs, notably those
on the grounds of the Kawaiahao Seminary and Punahou Academy,
occur on the valley floor.

The ten contributory branches of the stream are freshet in char-
acter and very intermittent in their flow. The stream proper is never
wholly dry. After the rare protracted droughts it becomes very low,
and its branches cease to flow. The upper course of the stream lies
in the middle of the valley. The lower course has been strongly
deflected against the east foothill, presumably by the outpourings of
lava and other volcanic material from the craters in the west foothill.
The stream leaves the valley at the extreme eastern side, hugging the
base of the east foothill, where it has carved a small narrow canyon
through the thick beds of ancient flow lava.

The stream is marked throughout its course by vegetation char-
acteristic of streamways and swampy places. Due to the general
utilization of the stream waters for irrigation, the swampy areas and
lot kalo (taro patches), adjacent to the stream itself are here considered
as a part of this ecologic zone. The vegetation of the streamway is
nowhere sharply. differentiated from that of the valley floor. In
numerous instances the species that grow most luxuriantly along the
stream are also forms most abundant on the valley floor.

From the ecologic standpoint the stream is at present a factor of
minor importance in determining the phytogeography of Manoa.
It undoubtedly had a more prominent réle in early times, before the
valley floor was overrun by introduced vegetation. One of the
influences of the stream is as an agent for the dissemination of seeds.
Frequently the seeds of montane species are carried to lower levels,
where they occasionally establish themselves. It is extremely sig-
nificant, however, that there has been no general seaward migration
of montane species via the stream; in general the forests have retreated
up stream. si

7- YEANTS ABUNDANT ALONG OR CHIEFLY CHARACTERISTIC OF
THE MANOA STREAM AND ITS TRIBUTARIES, INCLUDING
ADJACENT SWAMPS AND TARO-PATCHES

Group I, ALGAE A phanothece repens A. Br.

Anabaena confervoides Reinsch. Bulbochaeta spp.
“ variabilis Kuetz. Calothrix fusca Bornet.

578 VAUGHAN MACCAUGHEY

Chamaesiphon curvatus Nordst.
Chara coronata var. leptosperma.
“forma Oahunesis A. Br.
gymopus var. armata Nordst.
Cladophora Nordstedi De T.
Closteriopsis longissima Lemm. -
Coleochaete trregularis Pringsh.
i orbicularis Pringsh.
Conferva bombycina var. minor Wille.
ts Sandwicensis Ag.
Dactylococcus infustonum var. minor
Nordst.
Dictyosphaerium pulchellum Wood.
Draparnaldia macrocladia Nordst.
Gloeothece fuscolutea Naeg.
Gonium sociale Warm.
Hydrodictyon reiiculatum Lagerh.
Lyngbya aestuarit Liebman.
“ distincta Schm.
rivulartum Gomont.
Merismopedium glaucum Naeg.
‘Nitella Havatensis Nordst..
Nostoc commune Vaucher.
‘‘ paludosum Kuetz.
piscinale Kuetz.
punctiforme Hariot.
Raphidium polymorphum Fres.
Rivularia natans Welw.
Scenedesmus quadricauda Breb.
Scytonema crispum Bornet.
Spirogyra spp.
Stigeoclonium Falklandicum Kuetz.
Stigonema aerugineum Tilden.
. ocellatum Thuret.
Ulothrix minulata Kuetz.
ie subtilis Kuetz.
Xenococcus Kernert Hansg.
Zygnema spontaneum Nordst.

be

(a)

iz

bc

Group I]. PTERIDOPHYTES

Marsilea villosa Kaulf.......F, endemic.

iad 66

crenulata Desv.....R,

IAzZolla sp. noe F, recent introduction.

Group III. SpERMATOPHYTES
Illa. Indigenous
Aster divaricatus Torr. & Gray...HPF
Bidens chrysanthemoides Michx.. .HF

Cladium leptostachyum Nees &

Meyen. 3. 6.205 oe ee RPL:
Commelina nudiflora L.......... HPC
Cyperus auriculatus Nees........ HPL

“" jaevigatus Lee PE.
Eleocharis obtusa Schultes....... HPL
Ipomoea bona-nox L............. HPVF

x rebians Poir...) ee HPVF
Jussiaea villosa Lam: eee HPe
Kyllingia monocephala Rottb.....HPC
Naias major All HC
Polygonum glabrum Willd........ HF
Poe annua li. 2 ee HF
Poiamogeton fluitans Roth....... HPF

5D pauciflorus Pursh. ..HPF
SCL DUS TGCUSITIS 1a, ee ee HPF
<< mnariivmis Va eee HPR

Zingiber Zerumbet Ros..........

IIIb. Introduced by the Primitive

Hawaiians
Aleurites Moluccana Willd....... AF
Alocasia macrorrhiza Schott......HPF
Colocasia antiquorum Schott..... HPC
Eugenia Malaccensis L.......... AF
Hibiscus ulteceusi. eee AC
Musa sapientum L.............. HPF

IIIc. Introduced Since the Advent of

Europeans

Co: lachrymans... (2S ee eee HPC
Canna Indica, oy. ee HEE
Cyperus rotundus dae 1. be eee HPE
Echinochloa colonum Link....... HPG
Eugenia Jambos \2.. 2 3.4 eee AG
Hydrocotyle Asianea Wiss 20... ee HC

oe verticillata Thunb. ...HC
Leucaena glauca Benth.......... SC
Lemna minor Un. 3202 4. eee HC
Nasturtium officinale R. Br....... HF
Mazus rugosus Lout...:....22 4. eu
Panicum barbinode Trin......... HC
Pithecolobium Saman........... AF
Psidium Guayata 1... ee SC
Savitiaria sagitivjoliag a. ie He

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 579

See ie PARUS ZONE

Between the valley floor and the valley walls lies an intermediate
physiographic and floristic zone, which may be designated as the
talus zone. This zone comprises, as its name indicates, the talus
deposits at the foot of the valley wall and resting upon the floor. It
varies in width from approximately 100 to 1,000 feet. The slope
averages I0-I5°, as contrasted, on the one hand, with the 40° slope
of the wall, and, on the other, with the nearly flat floor. The soil of
the talus zone varies considerably in nature in various parts of the
valley—in some places it is fine-grained lava soil; in others, coarse
volcanic ash and cinders; and in others the ground. is littered with
massive boulders that have been dislodged from the heights above.
It is probable that the talus slope is not, in cross-section, wholly com-
posed of talus; the surface layers are of debris, and below them are
the ancient lava-sheets of the valley walls. This situation is revealed
by the little streamways that are cut through the talus zone.

The dominant plant of the Manoa talus zone is the guava, Psidium
Guayava. Secondary species are: Lantana Camara, Paspalum con-
gugatum, Andropogon contortus, Verbena Bonariensis, Psilotum nudum,
Morinda ciirtfolia, Nephrolepis exaltata, Solanum sodomeum, Con-
volvulus spp., Cassia occidentalis, Opuntia megacantha, Waltheria Amer-
acana,etc. ‘The talus zone, like the valley floor upon which it rests, is
covered almost exclusively with ruderal vegetation. Arborescent
forms are infrequent; vigorous and drought-resistant herbaceous-
woody shrubs are the prevailing types. .

In primitive times the talus zone of the upper valley was com-
pletely clothed with native trees, the species being those of the lower
forest zone. The forest retreated before the incursions of man and
wild live-stock, and exposed the talus zone to the invasions of foreign
vegetation. The hilo grass (Paspalum conjugatum) has been notably
pernicious, as it forms a dense sod and effectually prevents the native
species from reseeding themselves. |

The talus zone of the lower valley probably has been always more
or less xerophytic in character. Many of the indigenous or endemic
xerophytes of Hawaii have become extinct or are now upon the verge
of extinction. This condition is pronounced in several leeward
localities on the various islands—Hawaii, Maui, Molokai, Kauai—
and undoubtedly obtained in Manoa.

580 VAUGHAN MACCAUGHEY

9. THE VALLEY WALLS OR LATERAE RIDGES

The Manoa portal opens to the southwest and is bounded on east
and west by the plainward terminations of its two irregularly sculp-
tured lateral ridges. These terminal areas of the ridges may be
designated as foothills; that on the east is the Manoa-Palolo foothill;
that on the west is the Roundtop foothill.

The ridges extend from the coastal plain up to the main summit-
ridge of the range, which here has an average elevation of 2,300 feet.
Like all the ridges which define the Hawaiian valleys, these are the
remnants of an original volcanic dome. The lower or foothill ends of
the ridges are sufficiently bare of vegetation to reveal the laminated
series of basaltic lava flows, of which they are mainly composed.

Each lateral ridge may be divided by vertical lines into three
sections or areas:

1. The terminal or outlier foothill, which fronts upon and rests
upon the coastal plain.

2. The transition or intermediate “‘knife-edge”’ region.

3. The mountainward region, wherein the ridge connects with
or springs from the main summit ridge.

The Manoa-Palolo or East Foothill.—This, viewed from above, is a
fan-shaped mass, with the expanded portion abutting upon the coastal
plain. The upper slope narrows to a high (1,200 feet) apical region.
The seaward slope of the foothill has an angle of about 8°; the valley
wall is abrupt, rising at about 40°. The origin and physiography of
the foothill is due to the remarkably localized distribution of the
rainfall, as has been referred to in a previous section of the paper.
The rainfall on the foothill itself is comparatively slight. Therefore
erosion has advanced much more in the mountainward districts, and
has left the foothill as a more or less isolated and xerophytic outlier.
In the Waianae district, on Oahu, are found the culminating stages
in the isolation of the foothill from the main range.

The Roundtop Foothill.—The lateral ridge which constitutes the
western wall or boundary of Manoa Valley terminates in Roundtop
(Uala-kaa). This whole ridge is distinguished by a series of ancient
explosive volcanoes, of which Roundtop is the most seaward and
Mount Tantalus (Puu Ohia) is the highest and most conspicuous.
The highest points are Roundtop, 1,000 feet, Sugarloaf 1,400 feet,
and Tantalus 2,013 feet. Tantalus has a well-defined crater; the
craters of the other cones are either eroded away, or hidden under

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 581

volcanic ejecta. There are a number of lesser, unnamed craters on
the west ridge, and on the lower valley floor, which are very obscure,
and not of phytogeographic significance.

The three named craters in prehistoric times discharged vast
quantities of volcanic ashes and cinders. This material was deposited
in thick blankets over the local topography, obliterating the original
land features, and masked the irregularities which had been produced

7;
Fic. 8./ View on the slopes of Round Top, showing garden patches, papaia trees
and general physiography.

by erosion. Thus the surface of the Roundtop Foothill is much
smoother than that of the Manoa-Palolo foothill. On the latter the
surface material is the ponderous basaltic sheet lava of which the
original volcanic dome was composed;. on the Roundtop foothill the
surface material is a secondary volcanic product—lapilli—light in
weight, very porous, and produced within relatively recent geologic
times.

This difference in the nature of the surface material has resulted
in a striking difference in the agricultural utilization of the two foot-

582 VAUGHAN MACCAUGHEY
hills. The eastern one is so rocky and rugged that it is untillable,
and is used only as cattle land. The steeper slopes are covered with
various introduced weeds, which have been enumerated. The Round-
top foothill, on the contrary, is rich volcanic ash, thoroughly drained,
easily cultivated, and giving high yields of such crops as sweet potatoes,
papaias, onions, carrots, and various other garden vegetables. There
are many little garden patches on the upper slopes of Roundtop,
cultivated by Portuguese, Hawaiians, Orientals, and others. The
lower slopes are occupied by residences.
The outstanding .ecological characteristics of the foothill region are:
1. A strong tendency toward xerophytism, indicated by the presence
of many xerophytes and semi-xerophytes. |
2. Brilliant insolation, due to the fact that the foothills lie seaward of
the mountain cloud-cap, and under a sky which is largely cloudless.
3. Exposure to the winds, both trade and kona, owing to the smooth
topography.
4. No surface water, except during and immediately after rains.
5. Topography has permitted wild live stock to overrun the foothills
and to exterminate most of the native vegetation.
6. Invasion by a great variety of foreign: weeds, the woody or her-
baceous woody type being dominant.

10.. PLANTS OF THE .FOOTHILL AND ITS WALES

Most of the plants which occur upon the foothill and its walls
also occur on the valley floor; they are chiefly naturalized xerophytic
ruderals.

Group I. ENDEMIC Group II. INDIGENOUS
Cassia Gaudichaudit Hook. & Andropogon contortus Roem. &

AIM, eee ee eer eens SF Schult cate Ae ene HPF
Chenopodium Sandwicheum Mogq. .HSF Boerhaavia diffusa L...........- HSF
Lepidium Oahuense Cham. & Chrysopogon aciculatus Trin...... HPC

Soh sa Att Re ees cree: HPE

; Cyperus polystachys Rottb....... HPR
Lipochaeta connata DC. var. ; :

: Daucus pustilus Michx.......... HR
decurrens illebr eee PSR:

Te . : Dracaena aurea Mann........... AR
Nama Sandwicensis Gray....... HF
Neraudia melastomaefolia Gray ..SF Lup horbia cordata Meyen....... HPF
Phyllanthus Sandwicensis Gnaphalium luteo-album L....... HF

Mueller... Se eee SR a purpureum L.......HR

> Reynoldsia Sandwicensis Gray. ..AR Ipomoea pentaphylla Jacq....... HVE
Scaevola Gaudichaudii Hook. & ii tuberculata Roem. &
PAG Mele leper A cee Se ate eee SR Schult sk. ek Oe eee HPVF

Solanum aculeatissimum Jacq... .SF

Osteomeles anthyllidifolta Lindl...SVF

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 583

Panicum pruriens Trin.......... laWAe: Crotalaria fulva, KOxDs. 0.2 fs oe: HPC
a: torredum Gaud......... HC Gs SOLON HANG ered oe ee HPC
Phaseolus semterectus L......... Hee i spectabilis Roth....... HEC
Plectranthus parviflorus Willd... .HC Cynodon Dactylon Pers......... HPF
Plectronia odorata Benth & Hook. .SF CY PTUs TOlUNGUS Wi ctie te oto. lel
Pilea peploides Hook. & Arn.....HF Desmodium uncinatum DC......-HC
soem cordtjolid la, 202k eee: He Eleusine Indica Gaertn.......... HC
BLONDE OIG Linc. osteo ate 63 HPC Erigeron albidus Gray.......... HC
memiQuiane \Nalpsec io. weak ek HPC . COnGdensts len. wee toe lak
Solanum nodiflorum Jacq........ HR Eugenia Jambolana L........... AR
Stenotaphrum Americanum diunolus airidis. NOG). 3. sue. es HF
Bean Ke ot. haat hie to ss HPF Indigofera Anne lance. a eae led ©,
Waltheria Americana L.......... HPC LEONG COMO TES ia ce =16
Wtkstroemia uva-urst Gray...... PPok = levcecena clauca Benth... 9.7.) SF
Mirabilis Jalapa Lo... 2.2... HPF
Group III. INTRODUCED SINCE THE Momordica charantia L.......... HPVR
ADVENT OF EUROPEANS Nicotiana glauca Grah.......... AR
Acacia Farnensiana Willd....... S16 ae eae megacanih % as Sh See eee
Ageratum conyzoides L.......... HC pag ee CON EGE, meen ee” PC
T-OSSULONG fOCLIGR ts. ratte HPVF
Argemone Mexicana L....:....-HC : :
TAGeUCnvapOcltad: Warn hannah rte: oe HPC
Asclepias Curassavica L......... HF Wee
i cenvtlantitus Nordea Ne ae HF
isydens puos@ lie. ee ee ss HC
; Pithecolobium Saman Benth.....AF
Bryophyllum calycinum L........HPF :
; : PIC OSOLMOsOF Neste | ans ae ok HR
Cardamine hirsuta Le... 26... ss. HF a
e f PSidviinGuayaie Nae, eee. SC
Cassia occidenialis Li... os. wee HEC Prosopis juliflora L....-..-2.... AC
contra ee. Cav........ ae R Ricinus communis L............- SC
CO SEN Gis 3 as Salvia occidentalis Swartz........ HR
Centaurea melitensis L...........HF SONCHUS OICMACCHSL ne HC
Cestrum diurnum L............. AR SUOCHNS OPUCNSIS AA al a eos ox HF
Chenopodium album L........... HC Stachytarpheta dichotoma Vahl....HPC
Clerodendron fragrans Vent...... HPVR = Solanum Sodomeum L........... SF
Crepis Japonica Benth.......... HC Verbena Bonariensis L.......... Hee

11. THE TRANSITION REGION

This term is used to designate the ‘‘knife-edged”’ portion of the
lateral ridge, which lies between the foothill and the mountainward
termination of the ridge in the main range. The mountainward limit
of the foothill area is clearly defined by an eminence or little peak; -
beyond this the ridge abruptly descends and narrows. The con-
spicuous vertical erosion which has produced the “knife-edged”’ crest
so characteristic of this portion of the ridge, indicates clearly the
heavy rainfall to which it is subjected. The crest of the foothill is a
broad, sloping, triangular plane; the crest of the transition or inter-

584 VAUGHAN MACCAUGHEY

mediate region is very narrow, in many places being only 2 or 3 feet
in width. The valley walls of the foothill are relatively smooth and
unfurrowed; the walls of the transition ridge are deeply fluted, with
numerous alcoves.

The Transition Region marks the area intermediate, in ecologic
features, between the high, humid ridges of the rain-forest proper,
and the low, arid foothills with their covering of xerophytic and
semi-xerophytic vegetation. It marks with considerable accuracy
the usual seaward limit of the summit-ridge cloud-cap.

On the west ridge there is a marked discrepancy between the
situations of the topographic transition region and the vegetational
transition region. These two do not coincide; the topographic
transition region lies two miles mountainward of the vegetational
transition region. This difference is due to the presence of the
Tantalus series of volcanic craters along the west ridge; these have
pushed the topographic region much further mountainward than it
otherwise would have occurred.

On the east ridge practically none of the normal vegetation of the
lower or middle forest zones occurs seaward of the Transition Region.
On the west ridge Mount Tantalus rises to a height of 2,000 feet on
the seaward side of the topographic Transition Region, and supports
a luxuriant lower- and middle-forest flora.

The east transition ridge is but 1,200 feet high, at its lowest point,
whereas the west transition ridge is about 1,700 feet high. The rain-
forest, which on the east ridge does not extend beyond the Transition
Region, on the west ridge covers, not only the ‘‘transition”’ region,
but also the mountainward half of the Tantalus mass. This condi-
tion clearly illustrates that rainfall and not topography determines
the lower limits of the montane forest.

12. CHE. VALLEY HEAD

The head of Manoa Valley is an expanded amphitheater of erosion,
rimmed by abrupt and deeply dissected walls.. From the standpoint
of plant life it is an ecologic complex, comprising the following elements:
1. The upper valley floor, already described.

2. A zone of broad, gentle, grassy slopes, lying above the valley floor
and below the kukui zone. Many of these ridges are knife-edged
and precipitous in their upper courses, and separate deep, narrow
ravines (700—-1,400-ft. contours).

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 585

3. A-series of cliffs or palis, which lie between the ridges, and are more
or less covered with vegetation. These cliffs are 200-300 feet
high, and are cut at fairly regular intervals by V-shaped gorges
and hanging valleys, from the mouths of which waterfalls issue
and drop down the face of the cliffs.

4. Above the cliffs is a series of hanging valleys, separated from one
another by steep ridges. These ravines have an average eleva-
tion of 1,400-2,000 feet and open above the face of the precipice.
They extend abruptly back and up to the main summit ridge, a
distance of .50-.75 mile.

5. The summit ridge.

Fic. 9. /Typical physiography of valley head, summit ridge and peaks. Note
j ravines and hanging valley formation.

The general structure of Manoa Valley, with reference to plant
geography, is fundamentally the same as that of the other valleys
along the leeward flanks of the Koolaus. However, variations of
marked phytogeographic significance may be noted. Nuuanu Valley,
for example, has cut completely through the range, and so its head is
much more windswept than that of Manoa. The difference in the
windiness of the heads of these two valleys has produced an observable
difference in their respective vegetations, that of Nuuanu being

586 VAUGHAN MACCAUGHEY.

conspicuously wind beaten. ‘The heads of the valleys in the Punaluu
region support a much finer type of forest than that of Manoa, for the
former region has been practically free from the ravages of wild goats
and other herbivores, and the forest is in its primitive condition.
The Manoa Valley head occupies an ecologic position somewhat
intermediate between the extremely arid and depleted valleys toward
Makapuu Point, and the hygrophytic valleys of the central part of
the range.

13.- PULU PUBOr(EURITA)

The upper floor is bifurcated by a ridge which emanates from the
main summit ridge and which terminates in a green grassy hill known
as Puu Pueo, the Owl Hill. This median ridge is about 2 miles long,
its lowest point is 300 feet above sea-level, and Puu Pueo rises 500 feet
above the valley floor. Due to the extensive erosion in the region
mountainward of the hill, the ridge is conspicuously saddle-shaped,
when viewed from the side.

Puu Pueo was at one time, like the region immediately adjacent
to it, densely covered with the mantle of the lower forest; the ravages
of wild goats and cattle, wood-cutters, and in recent times, dairy
cattle, have stripped from the hill practically all of its forest growth.
The principal plant now is the ubiquitous Paspalum conjugatum; other
plants occurring here and there upon the hill are Scaevola Chamis-
soniana, Acacia Koa, Microlepia strigosa, Cordyline terminalis, Cler-
montia macrocarpa, Pipturus albidus, Sadleria Hillebrandu, Osteomeles
anthyllidifolia, etc.

This ridge originally extended down the valley much further than
it does at present. It is not unlikely that there were other ridges
lying parallel with it, and that the physiography was considerably
more complex than that of Manoa today. The present broad floor
may be the result of the almost total elimination of several of these
ancient ridges. Under this hypothesis the plant life of the valley
under these early conditions was probably more diversified and
precinctive than it is at present. Erosion has caused an infinitely
gradual shifting of plant groups and zones. Projecting this vision
into the future, the head of the valley will become increasingly larger,
all contours more regular, and the life conditions more mesophytic.
Puu Pueo will have vanished and the foothills will have been com-
pletely isolated as outliers, with low open gaps into Nuuanu and

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 587

Palolo. The foreign lowland vegetation will dominate the entire floor
and its adjacent slopes.

14. THE MANOA LOWER FOREST OR KUKUIL ZONE

As one views the upper portion of the valley, from the floor or
mouth, the most conspicuous plant zone is the kukut or lower forest.
This is due to the fact that the kukuz foliage is pale silvery green,
quite distinct from the yellow green of the grass lands or the heavy
somber green of the rain-forest. The kukui groves form a broad,
more or less broken band across the head of the valley.

Fic. 10. Kukuz tree in lower forest zone, covered with epiphytic plants, chiefly
pteridophytes, bryophytes and lichens.

The kukut or Candle-nut Tree, Aleurites Moluccana Willd., is an
euphorbiaceous tree. It was probably introduced by the primitive
Hawaiians from the South Pacific, where it is abundant. It is now
thoroughly established in the lower forest zone throughout the Ha-
waiian Islands, and is the dominant tree in many regions. It attains
a height of 40-60 feet, but is usually about 30 feet high, with a broad,
dome-shaped crown.

In Manoa the kukut occupies an irregular horizontal zone, lying
across the head and around the sides of the valley, mainly between

588 VAUGHAN MACCAUGHEY

the 300- and 1,000-ft. contour lines. Along its lower fringe or level
the kukui gives way to various species of woody or shrubby plants,
conspicuous among which are: Psidium Guayava, Lantana Camara,
Osteomeles anthyllidifolia, Eugenia Malaccensis, Cordyline terminalis,
Verbena Bonariensis, Hibiscus tiliaceus, Pandanus odoratissimus, Melia
Azedarach, Cassia spp., Leucaena glauca, Bambusa, etc. Along its
upper border or level it is more or less abruptly replaced by such forms
as Acacia Koa, Metrosideros polymorpha, Ilex Sandwicensis, Pelea spp.,
Pittosporum spp., Cheirodendron Gaudichaudu, and other rain-forest
forms.

Fic. 11. Ina Manoa hau (Hibiscus tiliaceus) jungle. The foliage canopy is thirty
feet above the men.

Along the lateral walls of Manoa the kukui extends seaward until
it reaches a point whereat the increasing xerophytism, and the devasta-
tions of wild goats and other pests, have inhibited its growth. There
is ample evidence that in early times the kukui forests of Manoa
extended much further seaward along the walls and floor of the valley
than they do at present. There has been extensive encroachment by
man and his live-stock upon all the native forests.

The kukui is a moisture-loving tree and in Manoa reaches its
finest development in the little vales or alcoves which furrow the

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 589

walls and head of the valley. It does not grow upon the exposed
ridges which separate these alcoves from one another, nor upon the
crests of the lateral ridges, but nestles in the alcoves. On the lateral
walls it ascends to within 100 feet of the crest of the ridge. At the
valley head the rain-forest rises 2,000 feet above the upper limits of
the kukui zone. |

The chief botanical features of the Manoa lower forest are as
follows:

1. The lower forest presents a series of life conditions much less
rigorous than those of the rain-forest. The slope, soil, protection
from wind, and mean temperature, are all more favorable for plant
development than are those same factors in the rain-forest region.

2. The lower forest, in the days of the primitive Hawaiians, was
an important zone for the raising of crop plants. Twelve species
were cultivated in little clearings here and there along the skirts of
the lower forest. This zone was subjected to the direct and indirect
effects of human utilization to a much greater degree than was the
rain-forest. In ancient Hawaii the rain-forest was not much fre-
quented by the natives—they made occasional visits for birds, canoe
timbers, etc.

3. The physiognomy of the lower forest zone has been strikingly
modified by introduced Hawaiian crop plants, particularly Aleurites
Moluccana, Cordyline terminalis, and Eugenia Malaccensis. ‘The
former has become the dominant tree, and in parts of Manoa and
elsewhere in the islands forms pure stands of considerable magnitude.

4. The lower boundary of this zone is undoubtedly at present at a
much higher level than ever before in the history of the islands. In
other words, the forested montane area is continuously diminishing;
the forest margin is slowly creeping up the mountains. In geological
time this movement was due to slow subsidence (according to the
subsidence theory); in recent times it has been tremendously accel-
erated by herbivorous animals.

5. The undergrowth of herbaceous and herbaceous-perennial vege-
tation is much richer in the lower forest than in the rain-forest. The
pteridophyte representation is much greater, however, in the latter;
the Manoa rain-forest possesses 93 species, the lower forest has 4o
species.

6. The lower forest tends to be more or less open, whereas the rain-
forest is a completely closed association. Epiphytic vegetation is
much more abundant in the rain- than in the lower forest.

590 VAUGHAN MACCAUGHEY

15 REPRESENTATIVE PLANTS WHICH EXCLUSIVELY OR IN MOST
PART INHABIT THE MANOA LOWER FOREST ZONE;
KUKUI ZONE

Group I. ENDEMIC Solanum Sandwicense Hook. & Arn..F
Trees Vaccinium penduliflorum Gaud. var.
Charpentiera ovata Gaud............ C calycinum Hillebr. PE gett it? -R
Clermontia macrocarpa Gaud........ C Wikstroemia foetida var. Oahuensis
Dracaena aurea Mann... ss R Grays... eee eee eee F
Elaeocarpus bifidus Hook. & Arn. ...C ere
Eugenia Sandwicensis Gray........ R oS igi ae
Gardenia Brighamii Manny......... R Anoectochilus Sandwicensts Lindl....R
« Remyt Manns ....~.. 0... R Canavalia galeata Gaud:.. ... ......VR
Maba Hillebrandéi Seem............ R _-Ltparis Hawattensis Mann......... R
Ochrosia Sandwicensis Gray........ R
Pteridophyt
Osmanthus Sandwicensis Knobl......F na ee
Perrottetia Sandwicensis Gray....... F Athy salts Potretianum Presl........F
Pipturus albidus Gray............. € Asplenium Macraei Hook, & Grev...F
Rauwolfia Sandwicensis A. DC...... C Cibotium Chamissot Kault. oanceenae Cc
Reynoldsia Sandwicensis Gray...... F Doryop HAS decipiens J. Sm......... C
Santalum Freycinetianum Gaud...... F Dry on ng nuda Underw. = SoS Ra F
Urera Sandwicensis Wedd.......... F rubsformis Robins Bie ees F
a stegnogrammoides C. Chr..R
Shrubs and Herbaceous Perennials Polyp odium Hillebrandu Hook Bea R
itvin oleae Goud VC Sadleria Hillebrandit Robins........ F
aa tier ye e lystichoides Heller.......R
Eragrostis variabilis Gaud.......... ( Panes ji a Candee
Euphorbia Hookeri Steud........... F Bt
i multiformis Hook. & Arn..C
Group II. INDIGENOUS
Freyewmeiia Arnon Gad. > oacsan C as ae
Gahnia Beecheys Mant... 2... .. 22. F Trees
MC”. plobosa Wianny 2s ae F Dodonaea wmscosa 1..." ee c
Gouldta-comacea Tillebis, . ose. ane F Maba Sandwicensts ADC ae ae C
Joinvillea adscendens Gaud.........R Pisonta umbellifera Seem........... C

Kadua acuminata Cham. & Schlecht FE Plectronia odorata Benth. & Hook. ..F
‘* cordaia Cham. & Schlecht... .F

Lipochaeta connata DC. var. de- Shrubs and Herbaceous Perennials
Currens Wale Drie. aie oes ee R Adenosiemma viscosum Forst........ Cc
Lysimachia Hullebrandi Hook. f..... F Caesalpinia Bonducella Flem........ R
i rotundtfolia Hillebr...... R Dianella odoraia Blume. .>.....+-.. F
Osteomeles anthylidifolia Lindl....... © Ipomoea mnsulans Steud... 5 ee €
Rhynchospora thyrsoidea Nees & Kyllingia monocephala Rottb........F
Méyen.. cis. code eee ee F Lythrum meniiimum TB. Woe ae F
Rollandia grandifolia Hillebr........ F Oplismenus compositus R. & Schult.. .C
i lanceolata Gaud. ... 0... gh Pontoum pruriens Vrin. 2.) ee Cc
Scaevola Chamissoniana Gaud....... G Phytolacca brachystachys Moq....... F
Sstda Meventana Walp... .. 2. 2e.2F Styphel.a tametameia F. Muell..... AG

Smilax Sandwicensis Kunth. ....... VC ‘Zingiber Zerumbet-Roscoe.... see iG

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 591

Herbs Group III. INTRODUCED BY THE
Commelina nudiflora L.......... tye PRIMITIVE HAWAIIANS
Daucus pusillus Michx,............ R Trees
Solanum nodiflorum Jacq........... C Aleurites Moluccana Willd.......... (@
Broussonettia papyrifera Vent....... F
Pteridophytes Eugenia, Malaccensis ta... 3... bacte a C
Adiantum capillus-veneris L......... C :
Asplenium caudatum Forst......... 18 Shrubs and Herbaceous Perennials
és horridum Kaulf..........€ Alocasia macrorrhiza Schott........ S
“ De AR Sar on ee Cc Colocasia antiquorum Schott........ F
‘6 unilaterale Lam......... @ Cordyline terminalis Kunth......... €
Dicranopteris linearis Underw....... Cc Curcumea longa Vee i. ae a R
Dryobteris cyatheoides Kuntze.......C Dioscorea pentaphylla L............F
‘6 Sandwicensis C. Chr...... EF " SOUUG MS. le! Soe nee eee e
Ceropteris calomelaena Link........ C Musa sapientum L................. C
‘ ochracea Robins.......... @ Piper methysticum Forst.....:...... R
Coniogramme fraxinea Diels.........C Tacca pinnatifida Forst............R
Dryopteris truncata Kuntze......... F Touchardia latifolia Gaud..........F
Lycopodium cernuum L............. C
Microlepia speluncae Moore........ R Group IV. INTRODUCED SINCE THE
ch SIRIQOSM ISAT aa. 8a C ADVENT OF EUROPEANS
Weotopieris Nidus J. Sm........... Cc Trees, Shrubs, Herbaceous Perennials
Odontosoria Chinensis J.Sm........ C Bambusa vulgaris Schrad. & Wendl. .C
Pellaea ternifolsa Link........ ae eee F Jantang COMar ONE Rs ete eee C
Phymatodes elongata Presl..........C Psidium Cattleyanum Sabine........ F
e spectrum Presl......... R a GHOVGUG NG as et Sere ees oe C
Psiloium nudum Griseb............ CG POSSULOTG CAUITS SIMS. ee) ee: VF
Pteridium aquilinum Kuhn......... F * lonrtfolia ia ee oe VF
CMSA POLICU Ma rae eS 2k des oe 8 C
Sadlerta cyatheoides Kaulf.......... C Herbs
Selaginella Menziesii Spring........ S Bidens puosa. Whee eee ee R
Tectarta cicutarta Robins........... Cc Grepis Japonica. Benth. t+). cian... CG
Trichomanes Bauerianum Endl...... F Physalis Peruvrang Wee, 0. 2 oe C
oe jiwle: Worst? ... &. wk F WENCCLOSUULCANIS Wet etn eas eae F
Votarnarreida Kaull. 700.5... 2. c SOUCIUS CLCTACCUS Amat see: aia Fea F

16. ALGAE OF THE STREAMS AND WATERFALLS OF THE LOWER
AND MIDDLE FOREST ZONES

Anabaena catenula Bornet. Cylindrospermum catenatum Ralfs,
A phanothece Naegelt Wartmann. > stagnale Bornet.
Cladophora fracta Ag. Draparnaldia macrocladia Nordst.
A nitida Kuetz. Fischerella ambigua Gomont.
Coleochaeie irregularts Pringsh. Gloeocapsa magma Kuetz.

66 66

orbicularis Pringsh.
Conferva bombycina var. minor Wille.

polydermatica Kuetz.
quarternata Kuetz.

66

592 VAUGHAN

Lyngbya cladophorae Tilden.
t Martensiana Menegh.
Mougetia capucina Ag.
Nostoc foliaceum Mougeot.
‘“ verrucosum Vaucher.
Oedogonium crispum var. Haviense
Nordst.
Oedogonium spp.
Oscillarotia sancta Kuetz.
a formosa Bory.
Phormidium favosum Gomont.

MACCAUGHEY

Phormidium papyraceum Gomont.
Schroederia setigera Lemm.
Scytonema guyanense Bornet.
mE ocellatum Lyngb.
rivulare Borzi.
varium Kuetz.
Spirogyra spp.
Spirulina major Kuetz.
Stigeoclonium tenue Kuetz.
Tolypothrix distorta Kuetz.
Ulothrix minulata Kuetz.

bc

6c

17. SRAVINES

Between the grassy ridges specified as ‘‘zone two”’ of the valley
head are deep, narrow, steep-walled ravines, lying between the 700-
and 1,400-ft. contours. ‘These ravines are not to be confused with the
hanging valleys, which occupy a higher level—1,400 to 2,000 feet—
and are mantled with the true rain-forest vegetation. The ravines
are occupied by plants of the lower forest zone. These narrow, humid
gorges are the regions of minimum illumination in the valley. Their
floors receive no direct sunlight until an advanced hour of the morning.
The eastern arc of the sky is shut out by the mountain wall. These
ravines are so narrow—their streamways are but 8 to 15 feet wide—
that sunlight can enter only directly from above, and from the front,
1. e., facing the main valley. The subdued illumination is augmented
by the cloud-cap that lies across the summit ridge. The gloominess
contrasts strikingly with the glare of the main valley floor.

The larger arborescent species that are most prevalent in the
ravines are: Aleurites moluccana, Eugenia Malaccensis, Charpentiera
ovata, Pipturus albidus, Urera Sandwicensis, Elaeocarpus bifidus,
Clermontia macrocarpa.

Under the shade of these trees occur a number of smaller species
that are characteristically shade tolerant, for example: Lysimachta
Hillebrandiu, Rollandia grandifolia, R. lanceolata, Cordyline terminalis,
Smilax Sandwicensis, Oplismenus compositus, Zingiber Zerumbet,
Alocasia, Colocasia, Dioscorea spp., Curcuma, Musa, Touchardta lati-
folia, Crepis japonica, and many pteridophytes.

The plants that grow in these cool, humid, shady, protected ravines
are sharply contrasted, from the ecological standpoint, with those
that inhabit the hot, arid, glaring, windswept foothill slopes. These
two habitats represent two environmental extremes.

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 593

isa) DHE ZONE OR KOA AND, ERHUA

Directly above the kukui zone and commingling with it along
its upper limits is the zone dominated by the koa, Acacia Koa, and the
lehua, Metrosideros polymorpha. Originally the koa was much more
abundant than it is at present; at this time practically all of the large
koa has been cut or killed and the trees which remain are only of
medium stature. The lehua is the most abundant tree in the Manoa
forests, and in the forests of the archipelago as a whole. Both it and
the koa attain their optimum development on the island of Hawaii,
particularly in the region of Puna and Olaa. In these districts trees
of 75-90 feet are not uncommon; in Manoa the average height is
35 feet.

The zone of koa and Jehua does not have as sharp horizontal boun-
daries as do some of the other plant zones. The koa thrives in Manoa
at elevations as low as 50 feet and was at one time fairly plentiful in
the valley floor, in districts from which it has been absent for the
last fifty years. The upper limit of the koa is also somewhat indefinite,
averaging 1,200 feet, but sometimes rising to nearly 1,800 feet. On
the island of Hawaii the finest stands of koa occur at elevations of
4,000 to 5,000 feet. The /ehua occurs scattered throughout the Manoa
rain-forest, particularly along the ridges, and ascends the highest peaks.
On the island of Hawaii it rises to a height of 9,000 feet.

Pao) DISTINCTIVE FEATURES OF THE MANOA’ RAIN-FOREST

1. The forest flora is composed almost wholly of arborescent,
shrubby, or woody species. Most of them are endemic and many are
confined to the island of Oahu. There are no gymnosperms.

2. The average stature of the trees is about 25 feet; many do not
exceed 20 feet. The more stunted forms occur on the steep slopes
and ridge crests; along the floors of the ravines the trees may rise to
heights of 30 to 40 feet.

3. Most of the shrubs are tall and semi-arborescent in character;
it is difficult to discriminate between the two habits.

4. The substratum is a thin layer of stiff, red soil, derived from the
basaltic lavas which directly underlie it. This soil is continuously
wet, and is exceedingly tenacious of its water. It contains very little
organic matter, owing to the steepness of the slope and the rapidity
of the erosion.

594 VAUGHAN MACCAUGHEY

5. The forest forms an almost unbroken mantle, covering the peaks,
slopes and ravines. The only gaps are those upon the very steep
cliffs, and the rents caused by landslides. The landslides vary in
width from 10-40 feet and in length from 20—400:feet. At any given
time there are approximately 125 landslide scars visible in the Manoa
rain-forest.

6. The foliage of the rain-forest vegetation is, in general, small,
simple, oval, thick, coriaceous, and with a glossy upper surface. The
prevailing color is a dark, dull, heavy green, approaching olive.

7. The vegetation is very slow-growing, and relatively small
shrubs and trees show that they have attained considerable age (30
to 50 years).

8. The undergrowth is scanty, and consists mainly of bryophytes
and the lesser species of pteridophytes. There is practically no grass
or annual vegetation.

9g. The flowers of the rain-forest are small and inconspicuous.
There is no well-defined flowering season, and very few showy species.

10. Tree-ferns and palms comprise a very minor element in the
rain-forest. Orchids are rare. Lianas are of a relatively few species,
and are not as abundant as in the lower forest zone. Plants along the
summit ridges, exposed to the wind, tend to assume krumholz forms.

11. Despite the heavy precipitation, the streams of the hanging
valleys and ravines of the rain-forest are exceedingly inconstant in
character, filling with great rapidity after a storm, and soon running’
almost dry.

12. In the absence of definite records for the Manoa rain-forest,
the data given by Shreve? for the Jamaican rain-forest may be pre-
sented as suggestive and probably very nearly the same as for Manoa:

Temperature -
of the Soil Of the Air
Annualmeane- sieht see alae: 61.67, 60.8° F.
Annual: mean range =. 4.0. Gae. 2.9° Bes

The humidity of the Jamaican forest (annual summary of monthly
means for I5 years), is 84.1 percent; Manoa conditions are closely |
comparable to this.

20. »HANGING VALLEYS
Above the abrupt slopes and precipices that frame the valley
head is a series of little hanging valleys. They are separated from

2, Loc; cit.

PHYTOGEOGRAPHY OF MANCA VALLEY, HAWAIIAN ISLANDS 595

Fic, 12./ Typical Kodlan summit ridge and peak. Elevation of camera, about ~
2200 ft. Note precipices and forest mantle.

596 VAUGHAN MACCAUGHEY

one another by steep-walled, knife-edged ridges. The ravines open
upon the precipices, with vertical walls of 200 to 300 feet directly
below their mouths, so they are true hanging valleys. They lie
chiefly between the 1,400-2,000-ft. contours, although some reach up
the slopes of Kona-hua-nui to 3,000 feet. The hanging valleys,
like the summit ridges and peaks, are mantled with the somber
greenery of the rain-forest.

The sides of these ravines are steep, and very difficult to climb.
They are 45°-65°; the steeper declivities are constantly marked by
landslides. These wounds cut through the soil to the underlying rock
and remain bare for a long time.

21. SUMMIT RIDGES AND PEAKS

That portion of the main summit ridge of the Koolau Mountains
which lies directly above the head of Manoa Valley, 7. e., between
Kona-hua-nui and Olympus, is 1.7 miles long, measured along the
crest. The ridge, viewed from above, is strongly curved, with its
concave side facing northeast (windward), into the Ohu-ohi amphi-
theater. The windward wall is a great precipice, about 1,000 feet
sheer, covered for the most part with scrubby vegetation, but im-
passable. The summit ridge forms an arc of 90°. The eastern half
of this arc definitely bounds Manoa; the western half is part of the
Kona-hua-nui mass. Erosion is rapidly bevelling the summit ridge,
which has a strongly serrate silhuette. In the process of time a gap
will be formed through the mountains, similar to the gaps at the
heads of Nuuanu and Kahili Valleys. The summit ridge and peaks
are covered with the dense drapery of the rain-forest.

The climate of the Manoa rain-forest is similar to that of all
tropical montane forests. The temperatures are very constant and
low as compared with those of the lowlands. Frost is unknown, and
in the absence of accurate records, 45°-50° may be taken as a mini-
mum. The rain-forest is far enough removed from the warm low-
lands to be little influenced by them. The Oahu altitudes are not
sufficient for alpine influences to be felt; this contrasts with the
great mountains of the island of Hawaii, which rise to nearly 14,000
feet. .

22. MOUNT KONA-HUA-NUI

Mount Kona-hua-nui is the highest peak—3,105 feet—in the
Koolau Range. It lies as a mighty rampart directly northeast of

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 597

the head of Manoa. Although not physiographically an integral
part of the Manoa region, it is of such ecologic importance that it is
considered herewith.

The airline distance from the valley-head precipices to the extreme
summit of Kona-hua-nui is about one mile. The most northern
branch of Manoa Stream originates at an elevation of about 2,600 feet,
very near the mountain summit. There is no other point along the
Manoa summit-ridge that rises above 2,400 feet, and the average is
about 2,200 feet. Thus all of the Kona-hua-nui region above 2,400

Fic. 13. Trail and camp in the Manoa rain forest. In the upper portion of a
hanging valley on the side of Kona-hua-nui.

feet comprises a life area which is without counterpart in any other
adjacent portion of the Manoa mountains. Certain plants are very
distinctive of these upper levels, and are rarely or never met below
the 2,400-ft. contour. Some of these are: Hesperomannia arborescens,
Chetrodendron platyphyllum, Exocarpus brachystachys, Vaccinium
pbenduliflorum var. gemmaceum, Lobelia Gaudichaudu, Lobelia macro-
stachys, etc.

Owing to its elevation Kona-hua-nui is a great rain-maker. The
trade winds become chilled in rising over it, the copious moisture
condenses, and a characteristic cloud-cap covers the mountain summit

598 | VAUGHAN MACCAUGHEY i
during most of the year. Torrential precipitation occurs on both the
windward and leeward slopes, and averages about 100 inches annually.
This heavy rainfall has cut the east and west faces of the mountain
into very steep precipices. The north and east faces are part of the
famous Koolau palit. The south and west faces are fretted with
hanging valleys, which debouch into Nuuanu and Manoa Valleys.

Fic. 14. View in the rain forest, showing lianas. Note man’s‘head in center fore-
ground, indicating height of undergrowth.

23; - MOUNA. OLYMPUS

Mount Olympus (Awawaloa) forms on the summit ridge the eastern
boundary of the Manoa region. It rises to an elevation of 2,447 feet
and closely resembles Kona-hua-nui in physiography and vegetation.
It is covered with the typical rain-forest vegetation; the peak itself
is wind-swept and the vegetation, like that of all the summit peaks
and ridges, gives every evidence of very unfavorable life-conditions.

24. REPRESENTATIVE PLANTS OF THE: MANOA RAIN FOREST

Group I. ENDEMIC ; Broussaissia arguia Gaud........... C

la. Trees Charpentiera ovata Gaud............ C

Wicacia 0d Gray . a7 ee ee ee F Cheirodendron Gaudichaudu Seem....C
Antidesma platyphyllum Mann...... F zs pblatyphyllum Seem... .R

BovcavelaivonmGaud; \.. 2 eae F Claoxylon Sandwicense Mueller..... F

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 599

Dubautia plantaginea Gaud.........

«€ Campylotheca Sandwicensis Hillebr...F
Elaeocarpus bifidus Hook. & Arn... .C ‘e (Coreopsis) macrocarpa
Eugenia Sandwicensis Gray........ R Gray, and Vats..c2:. C
Pua sandwicensis Gray... ....... R Clermontia macrocarpa Gaud........ C
Exocarpus brachystachys Hillebr.....R os oblongifolia Gaud........ F
Hesperomannia arborescens Gray....R Coprosma foliosa: Gtay >.) kb eae G
Ties Sandwicensis Loes.... 2.6 6.0... € * longrolid, Gray. foe F
Labordia membranacea Mann....... Cc Cyanea angustifolia Hillebr. and
i. SOSSUISMGTAY <M oh a. e VS? ket. n Meee alia a ee tanee te
bs POMVOMONGEAY 5.08 ee eae. a F US, ROCUMIMNGIG leDie: 28 #05 ee.
Masa Sandwiensis DC........5....C “. wGromestane Gaudie Sane. «ih:
Metrosideros macropus Hook. & Arn..R Cyriandra cordifolia Grav..........

* polymorpha Gaud......C ‘ gracilis Hillebr. and vars. .

Zi TWCOSUMGTAY. S037 tad c grandiflora. Gaudes «su.

_ tremuloides Rock...... F Hiullebrands Oliver.......
Ochrosia Sandwicensis Gray........R Kaithi Wawra ke.
Osmanthus Sandwicensis Knobl......F a latebrosa aie bie aay...
Pelea-elustacfolia-Gtay.... 0.05.0... F i LessonianaGtayan. faust

Peusanawicensis Gray... :.2 5... .. @ % Wiactaer Gta, «tree a
ne enormmagoiid Gray oc. ke es F i paludosa Gaud. and vars. .
Perrottetia Sandwicensis Gray....... S s
Pittosporum glabrum Hook. & Arn...F is triflora Gaud. and vars....

be

C
F
(®.
S
R
F
F
F
F
Cc
R
CG
PUChCHin gt tay ee Cc
F
glomeratum var. Delissea subcordata Gaud........... C
F
Cc
F
C
Cc
C
R
R
F
c

acutisepalum Hillebr..F Dubauna lond Took, &-Atn.e.. 3
i spathulatum Mann.....F Euphorbia clustaefolia Hook. & Arn..
Platydesma campanulata Mann..... F fi Hooker, Steud nthe get
os CONN Ube ILO: Mei ot R ey multiformis Hook. & Arn..
Pritchardia Marttt Wendl.......... F Gouldia coraces, Gtay ea ce.
Psychoiria hexandra Mann... ... . C Hibiscus Arnoinanus Gray . 62 3
Piteralyxia macrocarpa Schult....... R o TCORTO SELIM GD Ease, ne es serosa
Pterotropia gymnocarpa Hillebr...... R Joinvillea adscendens Gaud.........
Santalum Freycinetianum Gaud...... F Kadua cordata Cham. & Schlecht... .
Sapindus Oahuensis Hillebr......... R ‘acuminata Cham. & Schlecht..
Sideroxylon Sandwicense Benth. & | Labordia lophocarpa Hillebr. and
OO Kae ee eyes a nantes F Vial Neate hes. ene ee Nurs ay id ee F
sivaussia Fauriet Levl....: 1.2 oe: R Lipochaeta connata DC. var.
i kaduana Gray and vars... .C decurrens Tile Dr. ae rb. Ages oe Cc
a VONMGUSSUITOMNOCK Pi'5. gs a ok C Lobelia Gaudichaudi DC... . E
>, WWMarinvnaGiay.... 2... F ““ macrostachys Hook. & Arn...F
Suttonia Lessertiana Mez........... C Nothocestrum longifolium Gray..... BGs
Tetraplasandra meindra Harms..... @: PelcaLvdvatembinilebras.-. 2. Ages oR
* Oahuensis Harms....F _ “ oblongifolia Gray............R
Xanthoxylum Oahuensis Hillebr...... F Phyllostesta glabra Benth........... c
si dipetalum Mann...... R at® agrandijloraspenth.. ..... Cc
: : % larsuta benth) 0, 6... 2 €
Ib. Shrubs “* baruyjiora Benth... <2... .. C
Sometimes more or less arborescent Plantago princeps Cham. & Schlecht..R
Artemisia australis Less............ R Rollandia calycina G. Don......... R

600 VAUGHAN MACCAUGHEY

Rollandia grandiflora Hillebr....... F
e Humboldtiana Gaud.......F
2 lanceolata Gaud. and vars..F

longiflora Wawra var.

4c

angustifolia Hillebr....... R
‘ racemosa Hillebr.......... R
Scaevola Chamissoniana Gaud....... (C
5 glabra Hook; Ge Arn. 1. os. F
i molits Hook. & Arn. 233 3. & F
Schiedea Nutigiit Hooks elo. F
Smilax Sandwicensis Kunth........ VG
Solanum Sandwicense Hook. & Arn..F
Stenosvnersppsiie aut dias cmien wees F
Suttonia Sandwicensis Mez......... F
Tetramolopium Chamissonis Gray...R
Urera Sandwicensis Wedd........... F
Vaccinium penduliflorum Gaud......C
Viola Chamissoniana Gingins....... F
Viscum articulatum Burm. and vars..C
Wekstroemia Oahuensis Rock... ...F

Ic. Herbaceous Perennials and Herbs

Astelia veratroides-Gaud.. i. ek F
Alyxia olivaeformis Gaud. 2.22. . C
Baumea Meveni Kunth 232050. <4 F
Carex. Oahuensts Neyer = veins F
Gahnia Beecheyt Mann............. F
Gunnera petaloidea Gaud........... R
Isachne distichophylla Munro....... R

oy. pailens Ilillelra! air ese oe R
Iaparis Hawauiensis Mann......... R

PDT blanda Kunth.

Eekana-G. DC,

hypoleuca Miq.

insularum Miq.

s Koolauana C, DC.

: latifolia Miq.

membranacea Hook. &
Arn.

m pachyphylla Miq.

2 parvula Hillebr.

: reflexa Dietr.

Sandwicensis Miq.

Rhynchospora thrysoidea Nees &
eV eniy Auras. cd hes ee eee Bi

Id. Pteridophytes
Asplenium acuminatum Hook. &

Arn ce R
ty contiguum Kaulf......... FP
glabratum Robins........ R.
i Hillebrandn Cy Chr =. R
A Kaulfussai Schlecht...... R

u lobulatum Niettso*s. -24 R
iy nitidulum Hillebr........ R
" patens Kaulfn 7a R
pavonicum Brack........ R
pseudo-falcatum Hillebr...C
schizophyllum C. Chr.....F
vexans Teller o> > see ae R
Athyrium deparioides C. Chr........ R

a proiiferum: CiChr. 2 ae
Botrychium subbtfolzatum Brack.....R
Cibotium Chamissot Kaulf.......... C

“ Menzies Hook... 3. F
2 glaucum Hook. & Arn...... R
Cyrtomium Boydiae Robins......... R
Dicranopteris emarginata Robins.....F
Dielia pumila Brack: ). 3 R
“2. faleata, Bracks 1. ayer ee R
Diplazium arboreum Robins.........R
. Henzlionum ©, Cine R
Doodsa Kuntinane Gaud:. eve
Doryopteris decora Brack. °-. #2 RS
POs acuudens © Chie yee R
crinalis GC. Chr Pers eS: F

Hi Keraudreniana C. Chr....F

latifrons Kuntze. soo F
i. nuda Underwe. ....4 00 9G
as rubiginosa Kuntze....... F

squamigera Kuntze......F

‘ unideniata ©, Chr... 2 F
Elaphoglossum micradenium Moore. .C
Ms reticulatum Gaud.....C

Filwx Dougiassit Robins. (2.3 a R
EL CURIE Baldwini Eaton....R
recurvum Gaud....C

: lanceolatum Gaud..R
Lycopodium nutans Brack. (as
venustulum Gaud....... R

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS

Lycopodium polytrichoides Kaulf.. .
Maratiia Douglassit Baker.......
Odontoloma macraeanum Brack....
Polypodium abietinum Eaton......

M adenophorus Hook. &

is Saffordit Maxon

GrouPp II. INDIGENOUS

Ila. Trees

WOdGwAed VISCOSG. La. ook ee
Metrosideros polymorpha Gaud.....
Pisonia umbellifera Blume........
Trema Amboiensis Blume.........

Haahiolanum Brack...
re hymenophylloides Kaulf..
- pellucidum Kaulf......
pseudogrammitis Gaud. .
pumilum Robins.......
sarmentosum Brack... .
tamariscinum Kaulf.. .
tripinnatifidum Presl.. .
iyaris irresularts Walt... cet oc.
Sadleria Souleytiana Moore........
Seneaca robusta. Baker...........
Schizostege Lydgate: Hillebr........
Selaginella arbuscula Spring........

‘a Spring Gaus. obits...
Trichomanes cyrtotheca Hillebr.......

Ild. Shrubs and Herbaceous Perennials

Dianella odorata Blume...........
Lythrum maritimum H. B. K.......
Strongylodon lucidum Seem........

IIc. Herbs

Mancusepusiius Mich. 20.6... 8a
meriera depressa Banks ©. . < o.0 6 tiie <:-

IId. Pteridophytes

Adiantum capillus-veneris L.........

Asplenium horridum Kaulf.....

66
6b

66

unilaterale Lam....

insiticium Brack........
monanthes Werks OS

i

Coniogramme fraxinea Diels.......
Cyrtomium caryotideum Presl.......
Dicranopteris glauca Under.......
linearis Underw.....
Diplazium Sandwichianum Diels.... .
Dryopierts truncata Kuntze.......
Elaphoglossum aemulum Brack... .
gorgonium Brack.....

bc

66

- hirtum C, Chr
Hymenophyllum obtusum Hook. &

Hypolepis punctata Mett.........
Lycopodium cernuum L..... 2.2...
Serralim /Phunb: 33:2

66

ie pbhyllanthum Hook. &

Ophioglossum pendulum Is..........
Polypodium Hookert Brack........
Pteridium aquilinum Kuhn.......
Pieris excelsa Gaudin sane.
Psilotum complanatum Sw.......:
NUGUIN GtISCWen wey ee Cee:

66

éé

Group III.
PRIMITIVE ‘HAWAIIANS

Touchardia tatifolia Gaud.........
MSG SG DICINIVN mt eens ees:

Group IV.
ADVENT OF EUROPEANS

Buddleia Astatica Lour..........
TOHONnE: COMGrA ee eS
PSUCLUIME GUANOUGs ik. een es eae:

SUMMARY OF THE RAIN-FOREST

VEGETATION
SDB cret es = team NICE Reon t eI UR Jy Re et te
Sie See ee ame eh cease
Herbaceous perennials... ........
PETS eee ere een ete cee ata ee

ee ee @

Pareulum Poin fe.
Viliaria ricida. Kaulionn vere ee,

INTRODUCED BY THE

INTRODUCED SINCE THE

Species

ee)
A OS

602 VAUGHAN MACCAUGHEY

Endemic sic. icon iis OSes 198 Introduced since advent of Euro-

Indigenotis <492:0. ea ra ae nau he a7 peans.. 6. anal 2 ee 3
Introduced by primitive Hawatans...2° .Pteridophytes. 2... 1.4 23 93

ENDEMIC VEGETATION OF THE RAIN-FOREST

Common Frequent Rare
PCOS chet tSR ANE ae eae ets. ea age ee 18 19 13
STU stk, Ves Pe ees See eae econ eee 26 26 12
Herbaceous-perennials and herbs............. 2 5 4
(11 spp. Peperomia, abundance uncertain)
Pteridophytes 3) oats, re eee eg ree I2 21 22

25. BIRDS OF THE MANOA RAIN-FOREST
ALL ENDEMIC

Group I. Species that Have Become Extinct within Historic Times

Oahu Thrush, Phaeornis Oahuensis Wilson.
Oahu Akialoa, Hemignathus Ellisianus Gray.
Oahu Akiapolaau, Heterorhynchus lucidus Lichenst.
Oahu Akepeuie, Loxops rufa Bloxam; on verge of extinction.
Oahu Ou, Psititrostra olivacea Rothsch.
Oahu O-O, Moho apicaulis Gould.
Group II. Species that are Present, in Small Numbers, at the Present Time
Oahu Elepaio, Chasiempis Gayt Wilson.
Oahu Amakihi, Chlorodrepants chloris Cabanis.
Oahu Creeper, Oreomyza maculata Cabanis.
liwi, Vestarta coccinea Forster.
Akakani, Himatione sanguinea Gmelin.

26. ORIGIN OF THE ENDEMIC -FLORA

One of the most interesting problems connected with a study of
the Manoa phytogeography is that of the origin of the large endemic
flora, particularly that of the rain-forest. Shreve’s excellent state-
ment? is worthy of quotation at length: 3

“There is no type of vegetation in which may be found a wider
diversity of life forms than exist side by side or one above the other
in a tropical montane forest. Together with the structural diversities,
discoverable in the field or at the microscope, are diversities of physi-
ological behavior, discoverable by observation or experiment, and
sometimes correlated with the structural features. There are quite
as high degrees of specialization to be found in the rain-forest as may
be sought in the desert. The prolonged occurrence of rain, fog, and

3 Loc. cit., pp. 109-10.

PHYTOGEOGRAPHY OF MANOA VALLEY, HAWAIIAN ISLANDS 603

high humidity at relatively low temperatures places the vegetation
of a montane rain-forest under conditions which are so unfavorable
as to be comparable with the conditions of many extremely arid regions.
The collective physiological activities of the rain-forest are continuous
but slow; those of arid regions are rapid, but confined to very brief
periods. In the regions of the earth which present intermediate con-
ditions between those of the desert and the reeking montane rain-
forest may be sought the optimum conditions for the operation of all
essential plant processes. It is indeed, in such intermediate regions—
tropical lowlands and moist temperate regions—that the most luxuri-
ant vegetation of the earth may be found, and it is also in such regions
that the maximum origination of new plant structures and new species
has taken place.”

From the standpoint of conditions in the Hawaiian Islands, the
closing words of the above quotation are of particular significance.
Evidence is accumulating which indicates the former elevation of these
islands far above their present levels. There undoubtedly has been
a period of prolonged subsidence, amounting perhaps to several
thousands of feet. The very rich endemic flora that today occupies
the Manoa rain-forest very likely did not originate there, but rather
upon warm lowlands that are now submersed beneath the ocean.
In other words, Hawaii’s remarkable endemic flora evolved upon
prehistoric lowlands, and through slow subsidence of the land has been
slowly crowded up the mountain slopes, into zones distinctly unfavor-
able for plant evolution. This hypothesis is also applicable to the
various groups of animals—birds, snails, and insects, that today
occupy the upper levels.

CoLLEGE OF Hawa, HONOLULU

REVISION OF THE HAWAIIAN SPECIES OF THE -GEN@s
CYRTANDRA, SECTION CYLINDROCALYCES HIEEEBR:

JosEPH F. Rock

INTRODUCTION

The genus Cyrtandra is represented in the Hawaiian Islands by a
considerable number of species. To the 32 enumerated by C. B.
Clarke in his monograph, several new ones have already been added
and there still remain to be described at least seven species and as
many varieties.

It is to be regretted that much confusion exists in the taxonomy
of the Hawaiian species. This was mainly caused through the works
of C. B. Clarke and Hillebrand, both of whom described the same
_ species of Cyrtandrae contemporaneously, the one not being aware
of the other’s labors.

C. B. Clarke’s monograph, as far as Hawaiian species are con-
cerned, is based mainly on the collections of Gaudichaud, Barclay,
Wawra, Mann and Brigham, Nuttall, Asa Gray, and partly on speci-
mens forwarded by Hillebrand with manuscript names.

Hillebrand had evidently not forwarded a complete set of his
duplicates to Berlin and Kew, for practically none of the Hillebrand
material in the Berlin Herbarium bears C. B. Clarke’s determinations,
whereas they are present on all specimens collected by Gaudichaud,
Wawra and other earlier botanists, whose material is deposited in the
various herbaria of Europe and America.

In the Berlin Herbarium, where the writer was privileged to work
on the Hawaiian collection, he found C. B. Clarke and Hillebrand’s
species still in separate covers, notwithstanding the fact that Hille-
brand’s species, or at least some of them, are identical with Clarke’s
species.

For example, Hillebrand’s Cyriandra latebrosa (Fl. Haw. Isl. 337.
1888) is Cyrtandra longifolia Hillebr. var. degenerans C. B. Clarke,
and published as such in the latter’s monograph on the tribe Cyrtan-
dreae. Hiullebrand distributed material of this species under Cyrtandra
longifolia, which name was adopted by Clarke, giving Hillebrand credit

604.

REVISION OF THE HAWAIIAN SPECIES OF CYRTANDRA 605

for it as author, while Hillebrand himself published the same species
under the name Cyrtandra latebrosa without mentioning his former
manuscript name. On still another sheet we find for the same species
still another name—Cyrtandra paradoxa. Again, a specimen in the
herbarium at Vienna, No. 1991, marked Cyrtandra paludosa Gaud.
var. a longifolia Wawra, which is a synonym of Cyrtandra longifolia
Hillebr. (in Clarke Monogr. 276. 1883), was selected by C. B. Clarke
as the type for Hillebrand’s manuscript name C. longifolia.

It is the writer’s desire to clear up all this confusion. He was
in a position to examine the material which formed the basis for C. B.
Clarke’s monograph, and he also compared the same with Hillebrand’s
collection.

The larger portion of Hillebrand’s material was not known to C.
B. Clarke, as it came into the possession of the Berlin Herbarium
after Hillebrand’s death.

This paper is the first of a series on the genus Cyrtandra and con-
tains a critical revision of the section Cylindrocalyces Hillebr.

SECTION ONE: CYLINDROCALYCES Hillebr. Fl. Haw. Isl. 326. 1888

Calyx campanulate, cylindrical or fusiform, five-cleft into unequal
lobes, splitting laterally at last; corolla usually large, curved, bilabiate;
flowers single or few, subumbellate to cymose; leaves mostly glabrous,
or, when tomentose, usually so along the midrib and nerves under-
neath, with blackish or dark brown to reddish hair, often thickly
matted (C. longifolia, var. degenerans).

This section possesses now seven species, ten varieties, and four
forms, of which one species, two varieties, and four forms, are here
described for the first time.

The species are closely related, for example: C. paludosa and C.
longifolia; C. cyaneoides and C. waianuensis; C. grandiflora and
C. filapes; C. oenobarba var. petiolaris appears to come more or less
close to C. longifolia var. calpidicarpa, in the long cylindrical fruit.

CYRTANDRA PALUDOSA Gaud. Bot. Voy. Uranie 447. 1830. Var. a@
ayeicA C.7B. Clarke. Monogr; Cyrt. 5> 275.. (1883-1887

A low shrub, the young parts silky to rusty-tomentose; branches
glabrous, somewhat fleshy; leaves opposite, elliptical-oblong, thick,
chartaceous, acuminate at both ends, with crenate to serrate margins,
almost glabrous when mature, dark green above, pale underneath,
the veins straight and prominent, 10-22 cm. long, 4-6 cm. wide, on

60G aa! JOSEPH F. ROCK

petioles of 2-6 cm.; peduncle short, 5 mm. long; cyme few-(3-7)
flowered; bracts 6-8 mm., long-acuminate, covered with reddish
brown hair; pedicels 6-10 mm. long; calyx cylindrical to campanulate,
thin, IO-I15 mm., unevenly 5-fid to the middle or less into lanceolate,
acuminate lobes, splitting laterally, caducous when with fruit, partly
hirtellous or glabrous; corolla 15-20 mm., suberect, exserted, glabrous,
white; fruit 12-20 mm. long, 5-8 mm. broad, glabrous. 3

OauHu: ex. Coll. Gaudichaud no. 154, Iles Sandwich, visit 1841,
in herb. Berlin, and part of type in College of Hawaii herbarium;
Ins. Sandwic. Oahu, Meyen 5/31, labeled C. Garnottiana det. C. B. Cl.
C. paludosa, and Meyen C. triflora Gaud. det. C. B. Cl. C. paludosa
Woahoo, Ins. Sandw. Macrae, Maio 1825, in herb. Soc. Hort. Lond.
and in herb. Berlin; Lindley visit 1832 in herb. Berlin; Hawaiische
Inseln, Wawra no. 1665, Oahu, fruiting and flowering (three sheets)
in herb. Vienna and herb. College of Hawaii, and no. 2375 leg. Hbd.
comm. Dr. Wawra, in herb. Vienna; Niu Valley, Oahu, leg. Lydgate,
Willi, 1870, herb. Hillebr. Berlin; Kalihi, Oahu, Jan. 1870, leg. Hbd.
fruiting specimen in herb. Berlin; Palolo Valley, main ridge, flowering,
Nov. 7, 1908, Rock no. 96 in herb. College of Hawaii; Punaluu Mts.,
Koolau, flowering Nov. 14-21, 1908, Rock no. 291 in herb. College of
Hawaii; Waikane Mts., flowering, Jan. 23, 1909, Rock no. 1251 in
herb. College of Hawaii.

The Oahu specimens are the typical C. paludosa a typica C. B.
Clarke. The species occurs on Hawaii also, but is much smaller in
every way.

Hawalt: Kilauea, leg. Hillebr. April 1868, flowering, in herb.
Berlin; Hilo, leg. Lydgate in herb. Berlin (with small narrow leaves) ;
Kalanilehua, Kilauea, flowering, May 1912, Rock no. 10343 in herb.
College of Hawaii; Alakahi Kawainui along ditch trail, flowering and
fruiting, July 13, 1909, Rock no. 4473 (two sheets) in herb. College of
Hawaii; Alakahi ditch in swampy forest, flowering, June 1910, Rock
no. 8513 in herb. College of Hawaii.

The specimens from Alakahi and Kawainui gorges, near the
summit of the Kohala mountains, at an elevation of 4,200—4,500 feet,
differ considerably from the typical specimens occurring on Oahu;
on Hawaii where they grow in dense swampy forests in thick Sphagnum
moss they are only 2-3 feet in height, the leaves are smaller, ovate-
elliptical, much more coarsely serrate, of thicker texture, and on
shorter petioles; the peduncles are shorter than in the Oahu specimen, —
or are almost wanting; the calyx is glabrous and not thin. It would

REVISION OF THE HAWAIIAN SPECIES OF CYRTANDRA 607

perhaps be better to class these plants as a distinct variety, but owing
to the polymorphism of most of the Hawaiian plants it may cause
confusion if raised to varietal rank.

The plants from Kilauea (no. 10343) have still smaller leaves (ellip-
tical-lanceolate) and the branches are very slender, otherwise the same.

CYRTANDRA PALUDOSA Gaud. var. BREVICALYX Hillebr. Fl. Haw. Isl.
226. 1888

Leaves as in the species a typica C. B. Cl. though broader, on
stout petioles of 2.5—5 cm.; peduncle very short or almost wanting,
often only one-flowered, the filiform pedicels 18-36 mm.; calyx only
one third the length of the corolla, 6-8 mm., cyatiform with broad
acuminate lobes or teeth; berry slender, fusiform, 20-24 mm. long.

OaHu: Kaala range, Hillebrand, July 1870, in Herb. Berlin
(doubtfully referred here a specimen from the Waikane Mts. flowering
and fruiting Jan. 23, 1909, Rock no. 1135 in herb. College of Hawaii).

This variety differs from the species in the slender, long pedicels,
short peduncle and small calyx, perhaps only a form of the typical
species.

The writer’s specimen from the Waikane Mts. have long, very
slender, pedicels (a little over 18 mm.) but a distinct peduncle, the
calycine lobes being very short as is the calyx tube.

CYRTANDRA PALUDOSA Gaud. var. ALNIFOLIA Hillebr. Fl. Haw. Isl.
236. 1888
Young shoots and inflorescence hirsute with dark ferruginous
hairs; leaves broadly ovate, 10-12.5 cm. long, 6—7.5 cm. wide, some-
what obtuse, rounded at the base, the strong ribs and veins pubescent;
peduncle and pedicels 12-14 mm. each; calyx and corolla faintly
pubescent.

Oauv: Hillebrand in herb. Berlin.

This variety can be retained; in its general appearance it is a
typical C. paludosa but differs from a typica only in the longer pe-
duncles and pedicels, and in the young leaves, which are covered
with a fulvous tomentum.

CYRTANDRA PALUDOSA Gaud. var. MICROCARPA Wawra
Flora 55: 560. 1872 (not Hillebr.)

Plant low, I m., rarely more, scarcely branching; leaves as in C.
paludosa a typica though somewhat larger, light green above, fawn

608 JOSEPH F. ROCK

colored underneath, the midrib and veins prominent, covered with a
silky, brown pubescence; petioles about 4 cm.; inflorescence densely
clustered in the leaf-axiles covered with a brown, coarse pubescence;
peduncles short, o-4 mm., many-flowered; pedicels 5-20 mm., um-
bellate; calyx 8 mm. long, subglabrate, caducous; corolla 12-14 mm.
long, tube narrow, curved, glabrous; fruit 1 cm. long, 5-6 mm. broad,
numerous, subconglomerate.

Kauat: In the forests above Waimea, Wawra no. 2056 in herb.
Vienna, and herb. College of Hawaii; Kealiaand Waimea leg. Knudsen,
in herb. Berlin; at the head of Olokele canyon along rockwalls,
flowering Oct. 1909, Rock no. 5414 in herb. College of Hawaii.

Hillebrand’s var. 6 muicrocarpa is identical with Wawra’s var.
microcarpa and is therefore preoccupied by Wawra; there are three
sheets in the Vienna herbarium ex Coll. Wawra.

Hillebrand says: “including probably Wawra’s var. confertiflora
and herbacea.’”’ His variety confertiflora does not belong to C. paludosa
Gaud.; it was described by C. B. Clarke in his monograph as C. con-
fertiflora (Wawra) Clarke and that justly, for the plant has no re-
semblance to C. paludosa.

Wawra’s var. herbacea does not belong to C. paludosa Gaud. but
to C. oenobarba and Heller’s combination (C. oenobarba herbacea
Heller) is therefore correct.

CyYRTANDRA PALUDOSA Gaud. var. SUBHERBACEA Wawra, Flora 55:

559. 1872
Plant glabrate; leaves broadly ovate or suborbicular, rounded at
the base, shortly acuminate at the apex, on long, stiff petioles; in-
florescence almost as in C. paludosa typica; peduncle glabrous, but
with a reddish tomentum at the base, 3-7-flowered; calyx sub-
campanulate, membranaceous, glabrous, caducous; lobes little smaller
than in C. paludosa typica, and subdeltoid.

Kaval: plateau of Waialeale, Wawra no. 2155 in Herb. Vienna
(two sheets), det. by C. B. Clarke; part of type in herb. College of
Hawaii.

CYRTANDRA PALUDOSA Gaud. var. Gayana (Heller) Rock.

Cyrtandra Gayana Heller, Minn. Bot. Stud. 9g: 887, pl. 59. 1897.

A small tree, 3 m. high, the trunk usually Io cm. in diameter;
leaves opposite, lanceolate, tapering at both ends, 5—7.5 cm. long,
about 2 cm. wide, entire, bright green above, with impressed midrib
and veins, covered with a brown pubescence underneath, the petioles

REVISION OF THE HAWAIIAN SPECIES OF CYRTANDRA 609

1.5 cm.; flowers solitary or two, in the axils of the leaves, on a peduncle
of 3 mm.; pedicels 6-15 mm.; calyx thin, slightly pubescent, as in
C. paludosa typica; fruit ovate-oblong, 10 mm., tipped with the per-
. sistent style.

KAUAI: on the ridge west of the Hanapepe river, elevation 3,000
feet, and on the plateau above Waimea 4,000 ft. elevation, Heller,
no. 2495; forests of Kaholuamano, above Waimea, flowering, March
3-10, 1909, Rock no. 2280 in herb. College of Hawaii; flowering and
fruiting Sept. 1909, Rock, no. 5600 in herb. College of Hawaii.

Heller says: “‘It belongs to the group of which C. paludosa is the
type.”

In fact it is merely a variety of C. paludosa, differing from it in its
smaller, entire leaves and arborescent habit, otherwise the same. .
Heller states that it may be identical with Wawra’s Cyrtandra paludosa
var. arborescens. ‘This is however not the case. The writer was able
to examine Wawra’s plant, through the courtesy of Dr. Alexander
Zahlbruckner of Vienna. Wawra’s plant is now C. longifolia Hillebr.
var. arborescens C. B. Clarke, and is identical with Hillebrand’s C.
paludosa var. integrifolia.

CYRTANDRA PALUDOSA Gaud. var. haupuensis Rock n. var.

A small bush with thick angular branches; young shoots pubes-
cent; leaves opposite, elliptical-olbong, subcoriaceous, dark green
above, light brown underneath, glabrous on both sides, subentire,
with a slightly undulate margin and faint crenation, 15-22 cm. long,
3.5-5 cm. wide, acute at the apex, gradually narrowing at the base
into broadly auriculate margins; petiole 1 cm. long; inflorescence
axillary; peduncle. I-2 mm., 3—4-flowered; pedicels 10-12. mm.;
calyx thin, glabrous, caducous, nearly as long as the tube of the
corolla; fruits (immature) cylindrical-oblong, acuminate at the apex.

Kauai: Lihue, near the summit of the Haupu range, flowering and
fruiting March 18, 1909, Rock, type no. 2473 in herb. College of Hawaii.

A very distinct variety, nearly worthy of specific rank; it differs
from the species in its robust habit, subcoriaceous, subentire, auricu-
late leaves, and very short-peduncled inflorescence.

SPECIMINA EXCLUDENDA

C. PALUDOSA Gaud. var. HERBACEA Wawra no. 2070 in herb. C. B.
Clarke = Cyriandra oenobarba Wawra var. herbacea Heller.

C. PALUDOSA Gaud. var. CONFERTIFLORA Wawra no. 2057 in herb.
Vienna = Cyrtandra conferitflora C. B. Clarke.

610 JOSEPH F. ROCK

be

HANCAILARN TSLANDS

%, Litt
PT EP HAW AL

DOK

Fic. 1. Cyrtandra paludosa Gaud. var. haupuensis Rock. Type in the College of
Hawaii Herbarium.

REVISION OF THE HAWAIIAN SPECIES OF CYRTANDRA 611

C. PALUDOSA Gaud. var. INTEGRIFOLIA Hillebr. Knudsen no. 137,
Kauai, in herb. Berlin = Cyrtandra longifolia Hillebr. var. arbor-
escens C. B. Clarke.

Hillebrand’s specimen from the Kohala Mts. Hawaii and referred
by him to the above variety with Knudsen’s no. 137, is an entirely
different plant and has absolutely nothing in common with C paludosa
or with C. longifolia; the leaves remind one very much of a species
of Shorea. It represents an undescribed species.

CYRTANDRA LONGIFOLIA Hillebr.; C. B. Clarke, Monogr. Cyrt. 5: 276.
1883-87

Cyrtandra paludosa Gaud. var. longifolia Wawra, Flora 55: 558. 1872.

Branches scarcely quadrangular, glabrate, the young parts hirsute
with ferruginous hair; leaves opposite, shortly petiolate, o-1 cm.,
elongate-oblong, acute, subentire, 22 cm. long, 5 cm. wide, nearly
glabrous when mature, with a yellowish wool along the median nerve
on the lower surface; peduncles 0-5 mm., often one-flowered; bracts
narrow; pedicels 3 cm. long, or longer, with a reddish-brown tomen-
tum; calyx 12 mm., the tube campanulate,; with a reddish-brown
tomentum outside, the lobes deltoid-acuminate; tube of corolla 14
mim., cylindrical, curved upwards, glabrous outside, the lobes 7 mm.
long, 4. mm. wide, minutely pubescent inside; fruit 22 mm. long,
I cm. broad, broadly-oblong; calyx at first ampliate, persistent,
later caducous.

' Kauval: Hanalei forests, collected by Wawra flowering and fruit-
ing, no. 1991 in herb. Vienna.

This species is only known to the writer from the type which is
Wawra’s No. 1901 in the herb. Vienna.

There are two sheets in the Vienna herbarium both bearing the
number 1991.

The flowers are on long pubescent pedicels.

CYRTANDRA LONGIFOLIA Hillebr. var. ARBORESCENS C. B. Clarke,
; Monogr. Cyrt. 5: 276. 1883-87

Cyrtandra paludosa Gaud. var. arborescens Wawra, Flora 55: 558.
£372.

Cyrtandra paludosa Gaud. var. tntegrifolia Hillebr. Fl. Haw. Isl. 337.
1888. e

Branches thick, woody; leaves subentire, lanceolate, broader
towards the apex, as in C. longifolia typica, but attenuate at the base

612 JOSEPH F. ROCK

Ky

Fic. 2. Cyrtandra longifokia Hillebr. Sketched from the type in herb. Vienna,
ex coll. Wawra, no. 1991. Hanalei, Kauai.

REVISION OF THE HAWAIIAN SPECIES OF CYRTANDRA 613

and merging into a winged petiole of 2-7 cm., or subsessile; peduncels
nearly wanting, one-flowered, the rigid pedicels 1-2 cm., with reddish
tomentum; calyx 16 mm. long, glabrous, caducous; fruit ovoid-
elliptical.

Kavuat: Wawra no. I991 0 (not ¢, as given by C. B. Clarke) in
herb. Berlin, and portion of type in herb. College of Hawaii.

There are two sheets of this variety in the Vienna Herbarium;
not represented in the Berlin Herbarium. ‘This variety is only known
to the writer from Wawra’s collection.

Hillebrand’s C. paludosa var. integrifolia from Kauai belongs here,
rather than to the species on account of the much shorter pedicels.

CYRTANDRA LONGIFOLIA Hillebr. var. DEGENERANS C. B. Clarke
Monogr. Cyrt. 5: 277. 1883-87

Cyrtandra paludosa Gaud. var. degenerans Wawra, 55: Flora 558. 1872.

Cyrtandra paradoxa Hillebr. ms.

Cyrtandra latebrosa Hillebr. Fl. Haw. Isl. 337. 1888.

Cyrtandra Hawauensis Drake del Cast. Ill. Fl. Ins. Mar. Pacif. 7:
253, 1802, not C..B. Clarke.

Cyrtandra degenerans (Wawra) Heller, Minn. Bot. Stud. 9: 887. 1897.

Stem straight, 2-4 m. high, with a thick glutinous sap, the young
shoots and inflorescence dark ferruginous, with a thick squamaceous.
tomentum; leaves verticillate, 3-5 in a whorl, narrow-oblanceolate,
12-25 cm. long, 2—5.5 cm. wide, acute to acuminate, entire or shortly
dentate, chartaceous, dark green above, brownish underneath with a
short and soft tomentum, prominently penninerved; peduncle very
short, 2-4 mm., I-5-flowered, the flowers drooping on pedicels scarcely
longer than the peduncle; bracts linear-lanceolate, 10-14 mm. long;
calyx caducous, fleshy, shaggy outside and inside with dark squama-
ceous tomentum, 20-30 mm. long, fusiform in the bud, with a lateral
slit through which the corolla protrudes, the peaked top remaining
entire or splitting into five short teeth; corolla slightly exserted,
glabrate, curved, with large spreading limb, bilabiate, the upper lip
deeply emarginate, the lower three-lobed, 8-10 mm. long; style
twice as long as the glabrous ovary; berry olive-shaped, 26 mm. long.

OaHu: Wawra no. 1781 (two sheets) in herb. Vienna (type), and
part of type in herb. College of Hawaii; in deep and dark ravines of
Kalihi and Manoa, Hbd. without date or number, in herb. Berlin and
herb. College of Hawaii (Kalihi spec.); Mts. of Punaluu, Koolau
range along stream bed, flowering Aug. 1908, Rock, no. 9 in herb.
College of Hawaii; Punaluu Mts., flowering Nov. 14-21, 1908, Rock

614 -

PEG,

Cyrtandra longifolia

JOSEPH Fs ROCK

FLORA OF

a

Hillebr. var. degenerans C. B. Clarke. {Typical
specimen.

REVISION OF THE HAWAIIAN SPECIES OF CYRTANDRA 615

no. 933, 934; flowering Dec. 3-14, 24-29, 1908, Rock no. 157 & 407
in herb. College of Hawaii; Kaukonahua gulch, Wahiawa, flowering,
May 15, 1909, Rock, no. 3029 in herb. College of Hawaii.

Mo.oxat: Mapulehu Valley, Hbd. in herb. Berlin without date
or number.

The Oahu specimens are the typical var. degenerans; the plant
occurs in the very wet forests usually in deep ravines and along stream
beds in dense shade. It usually does not branch, but bears a crown
of leaves at the end of a stem 3-4 m. high, with the inflorescence
clustered in the leaf-axils. It is a rather handsome plant and re-
sembles somewhat certain species of the lobelioideous genus Cyanea
in habit.

Hillebrand’s specimens from Kalihi are identical with the writer’s
plants from the Punaluu Mts. of the same range.

CyYRTANDRA LONGIFOLIA Hillebr. var. DEGENERANS C. B. Clarke,
forma subglabra Rock
Cyrtandra latebrosa Hillebr. var. 8 subglabra Hillebr. Fl. Haw. Isl. 338.

1888.

Leaves quaternate, almost glabrate underneath, only the midrib
and veins reddish-tomentose, pale on both sides, shortly dentate,
thin, chartaceous, obovate-oblong, acute to acuminate, 16-30 cm.
long, 4-6.5 cm. wide, gradually contracting toward the base, sub-
sessile or running out into a broadly winged petiole; calyx mem-
branous, hirsute with dark brown hair, occasionally glabrate outside,
but hirsute inside, fusiform; corolla as long as the calyx, only the
lobes exserted, slightly hairy or glabrous.

MoLokatl: Kalae, Hillebrand in herb. Berlin without date or
number; Mapulehu Valley, flowering March 1910, Rock no. 12518 in
herb. College of Hawaii.

Mavi: Honomanu Valley, along stream bed, northern slope of
Mt. Haleakala, flowering, May 1911, Rock no. 12519.

Peawaim: «Valley of Holopalaw vin, Kohala, Hibd.; «biamakua,
Paauhau no. 3, forest, flowering July 5, 1909, Rock nos. 4061, and 4062;
Holokaiea gulch, back of Waimea, flowering and fruiting July 1o,
1909, Rock no. 4081 in herb. College of Hawaii.

The Maui and Hawaii specimens have green, thin, glabrous calyces,
and only the pedicel and nervature of the calyx is slightly hirsute,
while the Molokai specimens have the whole calyx densely hirsute.

The leaves in the Maui and Hawaii specimens are also thinner and
broader towards the apex, than those of Molokai.

616 JOSEPH F. ROCK

The specimen which Hillebrand records from Waiehu, Maui,
belongs to the writer’s forma cymosa.

CYRTANDRA LONGIFOLIA Hillebr. var. DEGENERANS Wawra,
forma cymosa Rock n. f.

Branches angular; leaves quaternate, broadly obovate-oblong,
thin chartaceous, sparingly pubescent on both sides, dark green above,
paler underneath, irregularly dentate, acute at the apex, 14-24 cm.
long, 4.5-9 cm. wide, contracting at the base into a broadly margined
petiole, subsessile, or on petioles of 2-3 cm.; inflorescence a cyme,
hirsute with brownish hair throughout; bracts linear-lanceolate, acute,
to subfoliaceous; peduncle 1.5-3 cm., 3—-8-flowered; pedicels 1.5-2.5
cm.; calyx not fusiform, split into subdeltoid or linear-lanceolate,
acute lobes, nearly glabrate or hirsute with brownish hair; corolla
exserted; berry unknown.

MoLoKal: Mapulehu Valley, flowering March 1910, Rock no.
10334 in herb. College of Hawaii.

Maur: Valley of Waiehu, Hillebrand, without date or number.

Hawatt: Holokaiea gulch back of Waimea, elevation 4,000 ft.,
flowering July 10, 1909, Rock nos. 4075 and 4479 in herb. College of
Hawaii.

The Hawaii specimens differ slightly from the Molokai specimens
in the larger and denser flowered cyme and the linear-lanceolate, acute
calycine lobes.

CYRTANDRA LONGIFOLIA Hillebr. var. DEGENERANS C. B. Clarke,
forma oppositifolia Rock n. f.

Branches quasi quadrangular; leaves opposite, obovate-oblong,
acute at the apex, hirtellous on both surfaces, especially along the
prominent midrib and nerves, with brownish hair, 15-22 cm. long,
4—6 cm. wide, gradually contracting into a petiole of 2-3 cm.; flowers
single, or three on a common peduncle of I-1.5 cm.; bracts linear- .
oblong, acute, 12 mm.; pedicels 8-10 mm.; calyx not fusiform, split
to near the base into 5, linear-oblong, acuminate lobes; peduncle,
pedicels, and calyx hirsute with reddish-brown hair; corolla slightly
exserted, lobes large, spreading, of unequal size, hirtellous or nearly
glabrate; berry unknown.

Maur: Western division, Honokawaii gulch, flowering Aug. 1910,
Rock no. 8206 in herb. College of Hawaii.

Differs from the other forms in the opposite leaves, deeply divided
calyx, and large spreading corolla lobes. .

REVISION OF THE HAWAIIAN SPECIES OF CYRTANDRA 617

fewpilebea ASE fF,

ha ihe fi Hemet ay,
é ,
Shr: ee

No. 7¢ Go
FLOR: A OF THE H., oh AHAN ISLANDS

LILES oF ia i HERBAR MI
41d

FIG. 4. Crytandra longifolia Hillebr. var. calpidicarpa Rock. Type in the College
of Hawaii Herbarium.

618 JOSEPH F. ROCK

CYRTANDRA LONGIFOLIA Hillebr. var. DEGENERANS C. B. Clarke,
forma auriculaefolia Rock n. f.

Stem fleshy; leaves quaternate, broadly obovate-oblong, obtuse at
the apex, dark on both surfaces, as well as slightly pubescent, 18 cm.
long, 5.5-6.5 cm. wide, subsessile, and broadly auriculate, with a basal
diameter of about 3 cm. in the older leaves; peduncle 3 mm., usually
three-flowered; pedicels 5 mm.; calyx fusiform, thin, 20 mm. long,
subglabrous; corolla slightly exserted, the lobes small, unequal.

Maur: Western division, Honokawai gulch, deep, shaded places,
along the stream, flowering Aug. 25, 1910, Rock no. 8159 in herb.
College of Hawaii.

Differs from the other forms in the dark, broadly obovate-oblong,
auriculate leaves.

CyRTANDRA LONGIFOLIA Hillebr. var. calpidicarpa Rock n. var.

Shrub I m. high, branching; leaves quaternate, elliptical-oblong,
membranous, pale underneath, light green above, glabrous on both
sides, excepting a reddish brown pubescence along the midrib, shortly
and unevenly dentate, subentire in the lower portion, acuminate at the
apex and base, 10-16 cm. long, 3-4 cm. wide, on a petiole of 2.5—3 cm.;
peduncle 2 mm., 2—3-flowered, reddish-tomentose; pedicels 2 mm.;
bracts foliaceous nearly as long as the calyx, the latter caducous, thin,
fusiform, glabrate, excepting the acuminate lobes; corolla curved,
exserted, 20 mm. long, including the spreading, subequal lobes; berry
long-cylindrical, 3.5-4 cm. long, 4 mm. wide.

OaHu: Windward side, Waiahole Valley, on rocky wall, near
waterfall at the head of the valley; flowering and fruiting Jan. 17,
1909, Rock, type no. 1093 in herb. College of Hawaii.

Remarkable for its long cylindrical fruit which, in shape, reminds
one of those of Calpidia. It is so far the only Cyrtandra found in
these Islands, with a 4 cm. long, cylindrical fruit.

Cyrtandra Waianuensis Rock n. sp.

Plant 1.5-2 m. high, erect, single stemmed, not branching; stem
somewhat fleshy towards the apex, thick, woody and brittle towards
the base, with a large crown of sessile leaves at the apex; leaves
broadly oblong, subentire, or faintly dentate, dark green above, pale
underneath, glabrous above, pubescent below, with fine yellowish-
brown hair, obtuse or subacute at the apex, 30-45 cm. or more long,
15-20 cm. wide, thin, membranous to chartaceous, suddenly contract-
ing at the base, sessile to subsessile; inflorescence densely clustered
in the axils of the leaves on a common peduncle of 2-3 mm., with

REVISION OF THE HAWAIIAN SPECIES OF CYRTANDRA

FIG. 5.

4 f

tf¢ Pahccilia WA $8 BLAS EO MP
J

* (aye!

no fA Ef

FLORA OF THE HAWAIIAN ISLANDS
COLLEGE OF HAWAR HERBARIUM
MOHLECTER FY FF BOCK

i
a. i cf 5 4
teak ter in WALA wield (oe

i

{

HE Bes

619

Cyrtandra Waianuensis Rock. Type in the College of Hawaii

Herbarium.

620 JOSEPH F. ROCK

numerous bracts at the apex, bracts oblong, acute, 12x 4 mm.;
pedicels 1 cm.; calyx caducous, thin, glabrous, green, fusiform, 15
mm. long, the lobes acute, short; corolla slightly exserted, puberulous,
the lobes very small and subequal; berry (immature) ovate-oblong,
I4 x5 mm.

OaHu: Waianu Valley, windward side of the Koolau range, near
the head of the valley, along stream-bed, flowering, Jan. 22, 1909,
Rock, type no. 1167 in herb. Gollege of Hawaii.

A very interesting and striking species, remarkable for its large
leaves, which are sessile. The species is single stemmed and at first
glance resembles some of the Hawaiian Lobeliads, especially species
of the genus Rollandia, in whose company the plant grows.

It is related to Cyrtandra longifolia var. degenerans and its forms.

CYRTANDRA CYANEOIDES Rock, Bull. Coll. Haw. 2: 39. 1913

Plant subherbaceous, somewhat woody at the base, the stem
erect, not branching, 11-12 dm. high, 4 cm. in diam., bearing a crown
of leaves at the apex, not unlike a species of Cyanea; leaves 45-55
cm. long, including the thick, winged petiole, 22.5—35 cm. broad; mid-
rib fleshy, obovate, rounded at the apex, suddenly narrowing below
into a margined petiole, the latter 1.5 cm. thick, texture of leaves
thick, but coriaceous to fleshy, upper surface deeply rugose, the
veins impressed, dark green, lighter underneath, glabrous, dull;
young leaves and petioles covered with a light silky brown tomentum,
with erose margins, the young leaves almost fringed; flowers numerous
in subsessile clusters surrounding the stem, and hidden; calyx with
prominent veins, curved, yellowish brown, five-cleft, bi-labiate, the lower
lip consisting of two sepals, 12 mm. long, suddenly narrowing into
filiform apices, the upper lip of three sepals half as long as lower lip,
the two outer ones only beaked, all parts covered with a silky brown
tomentum; corolla enclosed in the calyx, white, 36 mm. long including
the 25 mm. long tube, slightly curved, two upper petals rounded and
smaller than the three lower which are acute, pubescent; stamens
adhering in the lower half of the tube, the filament Io mm. long;
style 14 mm. long, green, thickening towards the base; stigma flat-
tened, two-lobed, the lobes obtuse, 2 mm. long; fruit ovoid (immature),
the calyx deciduous from fruit, on pedicels of 8-10 mm., and covered
in its young state with a brown tomentum.

Kauat: Forests of Kaholuamano, elevation 4,000 ft., on cliffs,
near streams or waterfalls, along the trail of the Waialae Valley,
flowering March 3, 1909, Rock, no. 2282 in herb. College of Hawaii.

One of the most striking species of Cyrtandra. It resembles a
species of Cyanea of the section Palmaeformes, hence the specific
name. The native name of this species is Mapele.

REVISION OF THE HAWAIIAN SPECIES OF CYRTANDRA 621

CYRTANDRA FILIPES Hillebr. Fl. Haw. Isl. 336. 1888

“Habit of C. grandiflora; leaves three or four in a whorl, flaccid,
pale underneath and glabrate, oblanceolate, 10-15 cm. long, 2.5—3.75
cm. wide, on petioles of 8-12 mm., acuminate at both ends, remotely
dentate or serrulate; flower solitary on a short peduncle of 1-6 mm.,
the slender pedicels many times longer, 18-36 mm.; bracts filiform,
4—6 mm., soon caducous; calyx herbaceous, glabrous, cylindrical or
campanulate, 12-24 mm., bilabiately five-fid to less than the middle,
into sharply pointed, lanceolate lobes, deciduous from the fruit;
corolla glabrous, as long as the calyx or longer, 20-28 mm., shaped
asin C. grandiflora; ovary glabrous; berry slender fusiform 18 mm.’’!

WEstT Maut: Gulches of Honokawaiand Kaanapali, Hillebr., without
date or number in herb. Berlin, part of type in herb. College of Hawaii.

The writer is only acquainted with this species from Hillebrand-
material in the Berlin herbarium. It is related to the writer’s C.
longifolia var. calpidicarpa.

CYRTANDRA GRANDIFLORA Gaud. Bot. Voy. Uranie 447. pl. 55.
1826

Cyrtandra Endlichertana Walp. Nov. Act. Nat. Cur. 19, Suppl. 1.

25021. 10.. 1843.
Cyrtandra Ruckiana Meyen Reise 2, 125. 1834.

A small shrub 1-2 m. high; branches quadrangular, the new parts
ferruginous-tomentose; leaves opposite, elliptical, acuminate at both
ends, subentire, 10-14 cm. long, 4-8 cm. broad, obscurely crenulate,
thin, chartaceous, glabrate above, pilose underneath especially along
the midrib and nerves; petioles 2 cm.; cymes few-flowered; peduncle
2-6 cm., 1—7-flowered; bracts foliaceous, 2-3 cm., ovate-lanceolate,
subpetiolate or clasping at the base, greenish, deciduous; pedicels
O-I cm.; calyx herbaceous, campanulate, 18-24 mm. long, unequally
five-fid, the lobes broadly triangular, acute; corolla large, glabrous,
exserted, 30-32 mm., the tube curved, the limb bilabiate, large, spread-
ing; ovary glabrous, style articulate at the base; fruit 16 mm. long,
8 mm. broad, ellipsoidal, glabrous, white, the calyx deciduous.

INSULIS SANDWICENSIBUS, GAUDICHAUD.

OanHu: Beechey in herb. Kew; Meyen, flowering specimen in
herb. Berlin, two sheets, one labeled Cyrtandra Rucriana, only leaf-
specimen, det. by C. B. Clarke as C. grandiflora; Mann et Brigham
no. 40 in herb. Kew; Nuttall in herb. British Museum; Nuuanu

1 Hillebrand’s description is here quoted, only the measurements have been
changed from inches to the metric system.

622 JOSEPH F. ROCK

Valley, flowering 1868, Hillebrand, without number, in herb. Berlin;
Wawra no. 1746 (flowering), in herb. Vienna (four sheets); Pauoa
Valley, flowering, Nov. 4, 1908, Rock no. 704; same bes flowering
Oct. 29,1909, Rock no: 10246:

In the Berlin herbarium with the Hillebrand material, is a sheet
labeled C. grandiflora, collected in the Malay peninsula, State of
Pahang in 1909, no. 13673; this plant does not belong to our C. grandi-
flora Gaud.

The calyx and flowers of that specimen are silky tomentose, and
in other respects it does not agree with our plant. The Meyen speci-
men is labeled C. Rucriana, while C. B. Clarke cites it in his mono-
graph as C. Ruckiana.

A very distinct species common in the valleys back of Honolulu, a
branching shrub with large white attractive flowers; occurs only at
lower elevations of 500—1,000 feet, usually in dense shade along water
courses.

CYRTANDRA OENOBARBA Mann, Proc. Amer. Acad. 7: 189. 1866

Low, decumbent, 3-6 dm. high, fleshy, the stem and petioles
shaggy with stiff, dark-brown, reflected hairs; leaves ovate or sub-
cordate, acute at the apex, denticulate, hirsute with reddish hairs
along the veins underneath, glabrate above, fleshy, 6.25-8.75 cm.
long, about 6 cm. wide, on stout petioles of 2.5—-5 cm.; peduncle one-
to two-flowered, as long as the petioles; calyx shaggy, oblong, 5-fid,
the lobes ovate lanceolate, sharply acuminate, foliaceous; corolla
slightly exserted, fully 2.5 cm. long, glabrous, the limb large, spreading.

Kauai: Wahiawa falls and Waioli, foot of Waialeale, Mann and
Brigham no. 616.

It seems that this species has only been collected by Mann, for
the writer found no material of it in any of the collections, neither in
those of Wawra nor SIL Serial C. B. Clarke in his monograph says
also “‘species non visa.’

The writer is not acquainted with the species but with the variety
petiolaris Wawra.

CyYRTANDRA OENOBARBA Mann var. PETIOLARIS Wawra, Flora 55: 563.
1872
Cyrtandra oenobarba Mann var. rotundifolia Wawra, |. c.

Cyrtandra oenobarba Mann var. obovata Hillebr. Fl. Haw. Isl. 338.
1888.

REVISION OF THE HAWAIIAN SPECIES OF CYRTANDRA 623

Plant low, procumbent, 12-36 cm. high; leaves opposite, ellipti-
cal, acute or rounded at both ends, 10-14 cm. long, 6 cm. wide, re-
motely serrate, glabrate above, with blackish-brown tomentum along
the midrib and veins, otherwise pale and glabrate; petioles 5-7 cm.,
with blackish hair; peduncle very short 5 mm., densely villous with
blackish hair; bracts 8 mm., oblong, deciduous; pedicels 2-4, 0-8
mm. long; calyx 2 cm. long, narrow, tubular, divided to the middle
into 5, linear-lanceolate lobes, covered with blackish to yellowish hair;
corolla 3 cm., glabrate; fruit 16 mm. long, 3 mm. broad, narrow cylin-
drical, the calyx persistent.

KAvAI: Wawra no. 2012, 2157 in herb. Vienna, and portion of type
of no. 2012 in herb. College of Hawaii; Hanapepe fall, Abbe Faurie,
flowering Dec. 1909, no. 625 (distributed as C. oenobarba Mann), in
the herb. College of Hawaii, as no. 12520.

A distinct variety, differing from the species in the long petioles,
very short peduncle, and pedicels; it is identical with Wawra’s var.
rotundifolia which seems to differ from it only in the glabrous leaves.

CYRTANDRA OENOBARBA Wawra var. HERBACEA (Wawra) Heller, Minn.
Bot. Stud. 9: 890. 1897

Cyrtandra paludosa Gaud. var. herbacea Wawra, Flora 55:559. 1872.

Herbaceous, procumbent; branches fleshy, villous with reddish to
grayish hair; leaves fleshy, elliptical or subovate, 17 cm. long, 6-10
cm. wide, coarsely serrate, on petioles of 2-6 cm.; peduncles very
short, 0-9 mm., many-flowered; pedicels short, often 0-7 mm.; calyx
glabrous; corolla large, curved, glabrous; fruit unknown.

Kauai: Hanapepe falls, Wawra no. 2070 in herb. Vienna; same
‘locality, July, Heller no. 2490 in part, distributed as C. oenobarba.

Wawra’s specimen no. 2070 is a distinct variety but comes close
to C. oenobarba var. petiolaris. It has nothing in common with
C. paludosa.

COLLEGE OF Hawall, HONOLULU

ON THE DISTRIBUTION OF ABNORMALITIES” IN: fae
INFLORESCENCE OF SPIRAEA. VANHOUT#Etr

J. ARTHUR HaArRIS
I. INTRODUCTORY REMARKS

Experimental breeders, primarily de Vries, have shown that in
certain races individuals of more or less mutually exclusive charac-
teristics may regularly occur in fairly constant proportions. Students.
of hybridization have devoted their chief effort for the past fifteen
years to a study of the laws of segregation of parental characters in
filial generations. Those interested in experimental morphology
recognize the fact that a pure bred individual may in the course of
its ontogeny display characteristics which might belong to distinct
varieties or species.

In its relation to both genetic and morphogenetic problems the
investigation of the distribution of abnormalities among the syn-
chronously developed organs of the same individual seems of im-
portance.

The purpose of the present note is to call attention to peculiarities
of the frequency disrtibutions of certain abnormalities of the pedicels
in one of the most splendid garden spiraeas, S. Vanhouttet.

II. History oF SPIRAEA VANHOUTTEI AND DISCUSSION OF
MorPHOLOGY OF INFLORESCENCE

1. History of S. Vanhoutier (Briot) Zbl.

Briot writes of the origin of the form which he refers to as Spiraea
aquilegifolia vanhouttei:' ‘Cette variété, obtenue par M. Billard...
de graines du Spiraea aquilegifolia.”’ He also states: ‘Le Spiraea
aquilegifolia est, dit-on, une forme du Spiraea trilobaia.”” In his
original description Zabel? gives no statement as to where or when
this “‘hybrid’”’ was formed. In a later paper? he merely refers to it

iBriot, Rev. Hort. 37: 269... 1866:

a Zobel, H., Gart. Zeit: 3% 496.5 1604.

3 Zabel, H., Mitteil. Deutsch. Dend. Ges. 1904: 59.

624

ABNORMALITIES IN INFLORESCENCE OF SPIRAEA VANHOUTTEI 625

as a hybrid between Spiraea cantoniensis and S. trilobata, without
giving any details or actual proof of its hybrid origin.

Schneider‘ follows Zabel in regarding the form as a hybrid between
S. cantoniensis and S. trilobata. So far as I can make out there is no
really valid ground for this conclusion. |

The early writers on S. vanhouttet noted abnormalities of the
inflorescence. Briot® describes some in the original material. Zabel®
has even described a new form, which he calls S. Vanhouttet var.
phyllothyrsa, in part distinguished by abnormalities of the inflorescence.

Those who desire to compare the anomalies described in this
paper with those hitherto recorded may consult these papers. My
purpose has not been to describe in detail all the types of aberration
which may occur, but rather to throw them into categories usable for
statistical analysis.

2. Descriptive Morphology of Inflorescence

The normal inflorescence of S. Vanhouttei isa many-flowered umbel-
like raceme. In general, the pedicels originate fairly close together,
but occasionally the lowermost flowers are considerably scattered.

Normally each ray, as I shall sometimes call the pedicels, is simple,
terminated by a single flower, but occasionally more or less com-
pounded. The normal inflorescence, composed exclusively of simple
rays, is too familiar to require illustration. Figs. 17 and 18 give a
good idea of the abnormal inflorescence, the latter figure representing
a rather advanced though by no means extreme stage of compounding.

In general it is the lowermost rays of the flower cluster which
become compound, but there are inflorescences, and perhaps indi-
vidual plants, in which this is not true.

The range of variation in the abnormal pedicels is, as shown in
Figs. 1-16,’ of the two plates, very great. In the earlier work with
the form J devoted much attention to the attempt to classify the
various anomalies into logical groups: for example, to distinguish
between synanthies and the compounding of the flower stalk, and
between synanthies and the production of an accessory pedicel im-
mediately below the terminal receptacle.

4 Schneider, C. K., Illust. Hand. Laubholzk. 465. 1905.

5 Briot, Rev. Hort. 37: 269. . 1866.

6 Zabel, H., Mitteil. Deutsch. Dend. Ges. 1904; 59-60.
7 Figs. I-16 are natural size, Figs. 17-18 twice natural size.

626 J. ARTHUR HARRIS

Typical cases of synanthy are shown from the side in Figs. 5, 6,
8 and 9 and from below in Figs. 2 and 3, to which Fig. 1 of a normal
flower is joined for comparison. Examples of the production of an
accessory pedicel below the normal receptacle are:shown from the
side in Fig. 10 and from below in Fig. 4. But all possible gradations
may be found between these two types of anomalies: hence it is idle
to recount the criteria which have been applied in an attempt to
distinguish between them. For example, it is difficult to decide just
how the cases illustrated in Figs. 7, 11, 12 and 14 shall be classified.
They combine in some degree the characteristics of perfectly consti-
tuted secondary inflorescences, of synanthous flowers and those in
which there is a production of an adventitious pedicel from below the
receptacle.

The numbers of flowers involved in synanthy varies considerably.
Generally it is 2, but 3, 4 or 5 may be found. Figs. 2-3, 5-11, and 14
serve as illustrations. The number of secondary rays originating
below the inflorescence is also variable. It is interesting to note that
very frequently, and I believe in the great majority of the cases, the
secondary pedicel extends considerably above the flower below which
it originates, as shown in Figs. 10 and 12.

Between synanthy, or pedicels showing secondary rays inserted
below the receptacle, to the most perfectly formed secondary ‘‘um-
bels,’’ as shown in Fig. 16, all possible transitions, both in number of
flowers and perfection of formation, are found.

Ordinarily the rays of the secondary umbels are inserted at about
the same position, but occasionally examples are found in which one
ray is considerably lower than the rest, or in which the lowermost
rays are rather scattered. The number of secondary pedicels varies
greatly.

Some abnormalities of inflorescence structures are almost invariably
formed in any large series of plants. Without going into details con-
cerning the general observations of the past several years, I think it
may be safely stated that variation in the inflorescence is to some
extent dependent upon the peculiarities of the individual plants and
to some extent determined by environmental conditions.

III. Discussion AND ANALYSIS OF DATA

The first problem to require consideration is that of the frequency
and the nature of the distribution of abnormal pedicels.

ABNORMALITIES IN INFLORESCENCE OF SPIRAEA VANHOUTTEI 627

Confining our attention to records from plants in which
abnormality occurs in considerable abundance, we may ex-
amine the actual and the percentage frequencies given in
Table I. Here the frequencies for 1909 represent the results
of countings on 18 individual plants. Those for 1913 are
based upon determinations on three large individuals grow-
ing at Cold Spring Harbor. |

The figures show clearly that the number of inflorescence
with no abnormal rays is far in excess of these with any
other number. Thus in 1909 61.5 percent and in 1913 from
34 percent to 40 percent of the inflorescences were without
abnormality, and this notwithstanding the fact that all
these series of material were selected for abnormality.
Furthermore the frequency of the inflorescences decreases
as the number of rays which are abnormal increases. This
is evident from Table I, the results of which are represented
graphically in diagram I.

In the foregoing table and diagram the percentage fre-
quencies have been computed by using the total number of
inflorescences as a base. It is instructive to determine the
relative frequen-
cies of different
o—0 =/. 09 18 Plant
Se: number of ab-
sm Mee normalities in
the inflores-
Abnormal Pedicels per liplorescence cences_ which
0 2 4 6 8 70... /2 ey have at least one

60

50

40

JO

20

/0

Ferc erage Frequency

DIAGRAM I. Percentage frequencies of number of abnormal abnormal pedi-
pedicels per inflorescence in all inflorescences. cel. The results

are shown in
Table II, and represented graphically in diagram 2.

Both methods give skew distribution, the highest frequency falling
on 0 or I abnormal ray, and the frequency decreasing from this class
to those with higher numbers of abnormal pedicels.

That the skewness of distribution of the number of abnormal
pedicels per inflorescence is not due to skewness in the distribution
of number of pedicels in the inflorescence as a whole is shown clearly
by diagram 4, which gives the percentage frequencies of number of
pedicels in three of the series. All of these distributions are fairly
symmetrical.

628 J. ARTHUR HARRIS

TABLE I

Actual Frequencies, f, and Percentage Frequencies of Number of Abnormal Pedicels per
Inflorescence in All the Inflorescences

Namoer 1909, 18 Plants 1913, Plant I 1913, Plant II 1913, Plant IIT
of Abnormal
Pedicels ye % Vi % Je) % ve %

fo) 1,255 | 61.51 388 | 34.18 | 334 34.25 | 364 39-91

I 203 9.95 1203) 2hl.4'5 136 13.95 iT 12,17

2 [22 6.47 93 8.19 105 10.77 82 8.99

3 120 5.88 98 8.63 94 9.64 68 7.46

4 95 4.66 83 Vey! 89 9.13 66 7.24

5 71 | 3:48 O54. 98.37 65 6.67 46 5.04.

6 55 2.69 64 5.63 65 6.67 48 5.26

iA 54 | 2.65 68 | 5.99 37 3-79 42 4.61

8 19 985% 38 3.35 22 2.26 30 3.29

9 II 53 29 2.56 10 1.02 21 2.30

10 12 .59 23 2.03 8 .82 16 1.75
Il 7 +34 9 79 3 -31 13 1.43
I2 50 gn eee 3 26 2 Aca 3 33
i3 aa = 9 79 2 ot I II
14 | I .04 4 35 2 21 I II
15 — — — — — — — ae
16 == —= I .08 = = = ae
Totals tat ses 25040°>| 99.96) |! 1,135) 4:99.96 975 100.01 gI2 100.00

TABLE I]

Percentage Frequencies of Number of Abnormal Pedicels in Abnormal
Inflorescences Only

Number Percentage Frequency
of Abnormal Z
Pedicels 1909, 18 Plants 1913, Plant I 1913, Plant II z913, Plant III
ie 25.86 17.40 21:22 20.26
2 16.82 12.44 16.38 14.96
2 15.28 P2312 14.66 12.41
4 Pre fo 1 OD a 13.88 12.04
5 9.04 12.72 10.14 8.40
6 7.08 8.56 10.14 8.76
i 6.87 9.10 5-77 7.66
8 2.42 5.08 BuA2 5.47
9 1.40 3.88 1.56 3.83
10 1.52 3.08 1.25 2.92
II .89 1.20 47 2:37,
12 -63 -40 47 oF)
13 ea 1.20 Bgl .18
14 | a2 54 21 .18
T5 — — — _
16 Te 13 a ee

ABNORMALITIES IN INFLORESCENCE OF SPIRAEA VANHOUTTEI 629

The fact that the three collections made from large individuals
in the spring of 1913 show the same type of frequency distribution as

30
20
YN
1S)
‘Ss
g
Jos o—e = 1909, 18 Plants
R Sai o---0 = /9/3, Plant]
ee o—o = /9/3, Plan7 LT
a : o---0= /9/3, Plarit Hl
5
iS)
a
0 R - —-3- ----e

Abrormal Fedicels per Inphorése eee
2 ¢ 6 8 /O0 /2 /4 16

D1AGRAM 2. Percentage frequencies of number of abnormal pedicels per inflor-
escence in abnormal inflorescences.

TABLE III

Number of Abnormal Pedicels per Inflorescence in Inflorescences of Various Total
Numbers of Pedicels

Total :

eee | Abnormal Pedicels ee
in Inflo-

rescence I 2 3 4 5 6 7 8 9 Io 13 § r2 | 14
Beer Se fee Leer gle fr ee) I
6— 8..., —}—);—}—}) — | — | — |) — |} — |) — | —- er er
Onli. 2 I te SS 2
io Aees |) ad I BAe eee All Stee ee aoe ai ee eee eo ee TiS
Males 23). 13'| 12 ol 2O 2 OP ee ee ee eee a
Pee 2Ong el 54.1043) |, 33 (2 3t"|) EGG) TI 02" he Te ong I [, | ==!) 206)
OMe ec 7ZOol| 3815-37. |. Bh" \) 240 | 17 419 8 5 Sale| 2. 56
22a er 27h | 20.) 23 14T8 9) D2. |, TON GhO 9 5 Ant 2 Ig Fk Sts 7
27-29...) I4 i eile ae 7 9) 3 2 | ea 2h Sau ee Uh oe Oe
30-22"... B 5 Aa i3 2 1 ee | Veet Se |e IN a een Ne [oe it,
22R-6 5.1 Pie | eae ge pee ease fame Leen So pce eo) a 2
26.308 | Mes a eae terran cee ) eras Pa A acl cnet nla Noe I
Motale 203) 132' |k120)|95" 12710 55), 54% LO 19° | 12 9) 7 Boone Os

the massed materials of 1909 is sufficient proof that the skew distri-
bution of number of abnormal rays per inflorescence is not due to any

630 J. ARTHUR HARRIS

process of combination of materials from individuals differentiated

with respect of the number of abnormal rays which they produce.
Table III shows the distribution of number of abnormal rays per

inflorescence in groups of inflorescences of similar numbers of total

1909 /8 PlarTs

/0 oe
0 oe
a 19/8, Plant
10 o
0
20 1913. Plant 11
10
0
2 2 113, Plarit Fil
/0
0 ee
3 5 7 9 Vey: TE AT

Number of Fhowers per Pedicel

DIAGRAM 3. Percentage frequencies of number of abnormal flowers per pedicel
in abnormal pedicels. Note the bimodal nature of the distribution.

rays in the’ data for 1909.. In this table are included wholly normal
inflorescences, all of which are entered in the zero class, as well as the
inflorescences which contain abnormalities. Whenever the countings
are sufficiently numerous to justify conclusions, the frequency distri-

ABNORMALITIES IN INFLORESCENCE OF SPIRAEA VANHOUTTEI 631

bution of numbers of abnormal pedicels per inflorescence in groups of
inflorescence with similar number of pedicels is skew, just as it is in
the series as a whole.

‘The second peculiarity of the distribution of abnormality in the
inflorescences of this species is to be seen in the frequency of numbers
of flowers per pedicel. This is excellently shown for a combined
series of countings made from 18 shrubs in 1909 and from three large
individual plants examined in 1913. ‘These frequencies, reduced to a
percentage basis, are represented graphically in diagram 3.8

TABLE IY

Bimodal Distribution of Number of Flowers per Pedicel in the Abnormal Inflorescences
of Individual Plants

Number of Flowers Produced by Abnormal Pedicels
f 2 2 4 5 6 7 8 Oye ato |iaera 12 | 13 |] 14 /15/16/17

I 27 5 7 Oils Lo sialest >On aon Ly | Ou) Talal al ler 26m
2 I <i jae ka I Be eye me Balle 2: al sae as 20
2 II Gaile ie OF FeO en erg Oe en | ee 69
4 264|-15, | 21 Oa) 807 10) OterOe| 5G) | 86: 2 he rae
5 Chipeta Sapa) (a4 LOT: | 24Ne DB 7 Sal aa i=l O08
6 22 TeleLOwee A | [Ou eid POo a PGo.TOr et? 1 (Oe ie F225
7, £2) 229 Oo, 12 Fon Shales orl ah Gael) S| = 96
8 ATE 25s 15 Ld FAA E20! oat 1620.) 52 O62 155427 O ri 280
9 2OWeROL TAs -7 | QO 1eLO |. -8. 15) beto Bee oat| enO
10 So SS 2 I TI} 4.) 6)..6)°4)0 —) ={=)=/44l- 35
II 2ialsht Sal Sa Bale Asie 2 e221 eh 7) | Aen eee | al oT}
12 SEEM POs eee ese Osh Stl aol Kr ecw oll 37
13 Ae aie TOU tA TO Le eSt SO eA 280) 2a eA aE hl 2S
14 FOU LTO" Or “7, -3 Sale hOntel 2a elon e2 Ort tle oulte |||] Te
15 AAa TOs! 7a le LOn es Or le Or le22) p85 AD. | AT. | 26; 260) 2)|—|—|=|" 253
16 22 lene. 9 Sine ae. 3e lise ot 2k 7 Ns Py a eh Nor ein fs Fho
17 22 Falls Ori. Oy mo eel We TAn eS a Onl (Oa) ok) sien
18 Bonk On OA =O. \y 8 OF 2Ou Aye COM nOde a2 Oa) 56 \r—r lef oo
‘Rotal. .|364 |173 |149 |121 |124 |188 |274.|373-|399 |307 |192 185 |32 |9|1/1| 2,792

In all four series there is a major mode on 8, 9 or Io flowers per
pedicel and a secondary mode on two flowers per pedicel.

This species adds, therefore, one other to the series of dimorphic
characters, which have been reviewed elsewhere.?®

That the bimodal condition for the 1909 series is not due to the
mixing of inflorescence from a series of shrubs, some of which have a

8 A few synanthous flowers have been included among the two-flowered pedicels.

9 Harris, J. Arthur, A Bimodal Variation Polygon in Syndesmon thalectroides
and its Morphological Significance, Amer. Nat. 44: 19-30. I9I0.

632 J. ARTHUR HARRIS

modal frequency of two flowers per pedicel in the abnormal pedicels
and other shrubs which have a modal frequency on a higher number
of flowers in the secondary umbels, as we may for convenience call
them, is shown at once by an examination of the frequency distri-
butions for number of flowers per pedicel in the abnormal pedicels
of the individual shrubs. This is given in Table IV. From these
data it appears that 16 of the 18 individual plants which contributed

-

o—e - 109 /8 Plarils
eo --0 =/H3,Plgnl LT
—o=/WF a7 lL

é 8 /0 /2 /4. /6 /8 20 22 24 26 28 30 32 34

DIAGRAM 4. Frequencies of number of pedicels in Spzraea.

to the total series of 2,792 abnormal rays upon which the polygon for
1909 in diagram 4 is based, show a distinct secondary mode on 2
flowers per pedicel.

The two plants which are exceptions to this rule are represented
by only 20 and 37 pedicels each. These numbers are entirely too
small to justify any conclusion concerning the nature of the distribu-
tions. Had larger. series of countings from these individuals been
available, they might have shown the same bimodal condition as the
other sixteen. |

ABNORMALITIES IN INFLORESCENCE OF SPIRAEA VANHOUTTEI 633

The whole series of available data may be taken to indicate with
remarkable consistency that the bimodal nature of the distribution
of number of flowers per pedicel is not due to heterogeneity of material,
so far as this may originate from the combination of inflorescences
derived from differentiated individuals. It represents, therefore, the
resultant of some group of factors innate in the individual.

Having eliminated the possibility of an influence of the individual
plant as a determining factor of the peculiar frequency distribution,
it seems worth while to enquire whether any characteristic of the
inflorescence itself may have an influence upon the distribution of

TABLE V

Bimodal Distribution of Number of Flowers per Pedicel in Inflorescences with Various
Numbers of Abnormal Flowers per Pedicel

Number of Flowers Produced by Abnormal Pedicels

Total Abnormal

Pedicels Total

2 3 4 5 6 Wi 8 9 ike) II I2
I 320:).- 8 5 ayia 245 Phd olin Pee oy edge 62 Gi) ant 130
2 Dehetre 9) Oot LO Sor hyo E ligated | 4a 186
3 27ers sie TO 5231 ND 63 150: 28) Ql a= 2904
4 21 nO 1S3)f 42)" 58) (62.62 |s—32 (oso B32
5 B44 85. 13 EO) 20+), 661135) 102)|) Ai ln i2 I 475
6 24. LO 4 05) | 30h 56.1680) |. OT 47 | LO I 384
7 20g BOR) O23 494) 794132) 96.) 44 WIZ 476
8 15 4 Ae tile P2051) 450") vO" O41 Az.) he I 304.
9 SO en ey A R29u" SARE SON 598) 20/5 Be ls 261
10 8 2 I EE A ZOU 54914677 2-46) |) £6 = 230
at 2° as I 2 ete | TOE, 20.428 6; -|- 99
12 I = I 3 I 4/| 13| Io 2 ale 36
13 2 = = = To R20! ler 2ar «20.1.8 ON a 17
14 P| = I A TSA 205). °h2 I ey | tees 56
16 tly = I I 5 5 4 —-}| -|] - 16
Dotal...... 225: (104 87.130 12310580. (.651 O08 (b322r 670) |*.2 3,396

the number of flowers per pedicel. It is quite conceivable, for example,
that the inflorescences which are the most highly abnormal as measured
by the total number of abnormal pedicels, should have a larger number
of flowers in their secondary umbels than those which are only very
slightly abnormal. The combination of a series of inflorescences,
some only slightly abnormal and others highly abnormal, might, under
such conditions, result in a bimodal distribution of flower number in
the series of pedicels from the combined inflorescences.

Table V shows the number of flowers per pedicel in the 3,396
abnormal pedicels produced by the 747 inflorescences examined from

634 J. ARTHUR HARRIS

Plant I in 1913, arranged according to the number of abnormal pedicels
per inflorescence. ‘This table is quite typical of the others which
have been made. The bimodal distribution characterizes all the
arrays in which the number of observations is large enough to give the
distributions critical value.

Differentiation between individual plants and differentiation due
to correlation between the characteristics of the pedicels and those
of the inflorescence would seem to be the most probable source of a
spurious bimodality in the frequency distributions of number of
flowers per pedicel. That neither of these factors underlies the
observed form of the frequency distribution seems quite clear from
the foregoing tables. Bimodiality seems rather to be due to innate
factors operative in the morphogenesis of the individual pedicels.

Now in examining the graphs on diagram 3 the reader will note
that the mode on 2 flowers per pedicel is but a transition stage to the
higher mode—not represented in the diagram—on a single flower per
pedicel—that is to the normal condition.

TABLE VI
Distribution of Number of Flowers per Pedicel for All Pedicels

Number of Flowers mares
Perec ece 1909, 18 Shrubs 1913, Plant I 1913, Plant II 1913, Plant III

I | 14,250 8,436 5,196 5,714

2 364 225 177 193

3 173 94 83 103

4 149 87 G2 IOI

5 121 139 87 119

6 124 319 159 245

7. 188 580 280 367

8 274. 851 506 616

9 373 698 601 458

10 399 321 305 114

II 307 79 112 22

I2 192 3 35 8

ig) 85. * os 9 3

14 32 — I I

15 9 ra Tate oe

16 I ar a ere

17 I — a et
Motal\. 2.650. 17,042 11,822 72622 8,064

Tabulating the actual number of flowers per pedicel for all pedicels
examined, the frequencies in Table VI are obtained.

ABNORMALITIES IN INFLORESCENCE OF SPIRAEA VANHOUTTEI 635

These seriations bring out clearly that there are minimum fre-
quencies on four flowers per pedicel, and that the number of pedicels
increases rapidly as the number of flowers becomes smaller, reaching a
maximum on the single flowered (normal) pedicel, and that on the
other hand it increases more gradually as the number of flowers
becomes larger to a secondary maximum on 8—10 flowers per pedicel.

CONCLUDING REMARKS

This paper provides illustrations of the chief types of variation
which occur in the inflorescence of Spiraea Vanhouttet, and records
the results of statistical studies of the distribution of these abnormal-
ities.

The inflorescence of S. Vanhouttei is described by taxonomists as a
simple raceme-like umbel. A number of the pedicels are, however,
often compound. The compounding may range from a simple
synanthous condition to the production of a perfect, many-flowered,
secondary umbel in the place of the solitary flower which normally
terminates the pedicel.

The distribution of abnormal pedicels among the inflorescences in
plants in which abnormality is of frequent occurrence is not to be
represented by a normal or Quetelet’s curve but forms a one-sided
or skew frequency distribution, in which the frequency of occurrence
decreases as the number of abnormal pedicels per inflorescence becomes
larger. This is the first law of variation in the abnormalities of the
inflorescence.

If the normal pedicels be included with the abnormal to form a
single frequency distribution of number of flowers per pedicel in
inflorescences in which some of the pedicels are abnormal, it will be
seen that the two flowered pedicels form a transition frequency from
classes of larger numbers of flowers to those with a single flower per
inflorescence.

On examining these bimodal frequency distributions the reader
may be tempted to formulate some hypotheses concerning the co-
existence of determiners of mutually exclusive characters or con-
cerning the segregation of alternative, but variable, characters in
the morphogenetic processes of the individual plants. S. Vanhouttei
is known to be of garden origin and is supposedly a hybrid between
S. cantoniensis and S. trilobata. The evidences are not, however,

636 J. ARTHUR HARRIS

such as would be readily accepted by a careful geneticist. Further-
more, both of these assumed parent species have simple inflorescences.
S. Vanhoutiet is horticulturally such a splendid form that it would

be well worth while for some unoccupied geneticist to attempt further
crosses in this group of Spiraeas.

STATION FOR EXPERIMENTAL EVOLUTION,
CoLp SPRING Harpor, N. Y.

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE XXX.

HARRIS ;: ABNORMALITIES IN SPIRAEA.

AMERICAN JOURNAL OF BOTANY. VOLUME IV, PLATE XXXI.

HARRIS: ABNORMALITIES IN SPIRAEA.

ke , eos ee
rf ; Pe
/ :
. i"
1
( a
it
/ *
if
, » .
sf
i r
{ ss
“
4 2
‘
1 ’
a
‘
: D “
.
7
« f *-
|
i , €
* |

A.

=>

@

INDEX TO VOLUME IV

(New names and final members of new combinations are in heavy-face type.)

Abies grandis, duration of leaves in, 149;
host of Uredinopsis, 332; /astocarpa,
host of Uredinopsis, 332

Abnormalities in the inflorescence of
Spiraea Vanhoutiet, 624

Achlya, a form lacking sexual reproduc-
tion, 354

Aecidium abscedens, 46; Borreriae, 46;
circumscriptum (syn.), 49; Crsst (syn.),
49; Chbadii (syn.), 49; decoloratum
(syn.), 49; expansum (syn.), 51; passt-
floritcola, 46; Tournejortiae, 46; tubulo-
sum, 46

Agaricaceae, taxonomy of, 315

Agarics, development of, 221

Alternaria, taxonomic characters of, 439

Antagonistic action of other sugars
toward galactose and mannose, 430

Arbutus menziesu, 153

Arctostaphylos tomentosa, uva-ursa, 154

Ascospore production, in Endothia, 18

Asthma, caused by wheat protein, 417

BarRTLETT, H. H. (see La Rue, Carl D.,
119)

Berberts aqutfolia, nervosa, 154

BuizzaARD, A. W. The development of
some species of Agarics, 221

Botryorhiza Hippocrateae, 47; 52

Carbohydrates, effect of, on development
of rusts, 197 .

Carbon dioxide, effect of lack of, on
development of rusts, 195

Castanea dentata, host of Endothia, 3,
112; pumila, host of Endothia, 3

Ceanothus velutinus, 153

Chestnut blight, 1

Chimaphila menziesii, umbellata, 154

Chlorophyll formation, influence of, 407

Cichorium tntybus, fertility in, 375

637

Cissus sicyoides, host of Endophyllum,
49

Clibadium arboreum, Donnell-Smithi,
erosum, surinamense, hosts of Endo-
phyllum decoloratum, 50

Climatic factors, influence of, on the
development of Endothia parasitica, I

Clitocybe adirondackensis, cerussata, de-
velopment of, 226, 228

Clitopilus noveboracensis,
of, 230

Coleosporium solidagints, 334

Conductivity of electrolytes, 411

Cortinarius pholideus, development of, 520

Crown-rot of fruit trees: histological
studies, 477

Cyrtandra, revision of the Hawaiian
species of, 604

Cyriandra cyaneotdes, 620; filipes, 621;
grandiflora, 621; longifolia, 611, var.
arborescens, 611, var. calpidicarpa, 618,
var. degenerans, 613, formae auriculae-
folia, 618, cymosa, 616, oppositifolia,
616, subglabra, 615; oenobarba, 622,
var. peitolaris, 622, var. herbacea, 623;
paludesa var. alnifolia, 607, var. brevi-
calyx, 607, var. Gayana, 608, var.
haupuensis, 609, var. microcarpa, 607,
var. subherbacea, 608, var. typica, 605;
Waianuensis, 618

development

Development of Cortinarius pholideus,
520

Development of some species of Agarics,
221

Drabas of North America, 253

Draba alpina, 257; argyrea, 266; astero-
phora, 263; Belliz, 261; cruciata, 265;
cyclomorpha, 263; Douglasii, 259;
fladnizensis, 257; globosa, 257; incerta,
261; laevicapsula, 262; Lemmont, 263;

638

lonchocarpa, 262; Mulfordae, . 264;
Nelsonii, 259; nivalis, 262; oligosperma,
259; oreibata, 257; pterosperma, 266;
sphaerocarpa, 266; sphaeroides, 265;
sphaerula, 258; subsessilis, 267; un-
_ cinalis, 265; ventosa, 264; vestita, 261
Duration of leaves in evergreens, 145

Effect of soil temperature on the growth
of bean plants and upon their suscep-
tibility to a root parasite, 513

Electrolytes, measurement of the con-
ductivity of, 411

ELLIOTT, JOHN A. Taxonomic char-
acters of the genera Alternaria and
Macrosporium, 439

Endemic flora, origin of, in Hawaii, 602

Endophyllum circumscriptum, 49, 52;

-decoloratum, 49, 52; Stachytarphetae,
50, 52; Wedeliae, 49, 52

Endophyllum-like rusts of Porto Rico, 44

Endosperm color, inheritance of, in
maize, 396

Endothia parasitica, influence of climatic
factors on the development of, 7;
gyrosa, 7; radicalis, 7; Influence of
temperature on growth of, 112

Endothia pigments, 336

Environment, as related to leaf-structure,
533

Evaporation, relation
plant succession, I61

Evergreens, duration of leaves in, 145

between, and

Fertility in Cichorium intybus: the
sporadic occurrence of self-fertile
plants among the progeny of self-
sterile plants, 375

Fritillaria, rusts on, 368

Fusarium Martit phaseoli, root parasite
of beans, 513

Galactose, toxicity of, for green plants,
430

Galls, produced by Gymnosporangia, on
cedar, 241

Galvanometer, use of, with a 60-cycle

INDEX TO VOLUME IV

alternating current in the measure-
ment of the conductivity of electro-
lytes, 411

GATES, FRANK C. The relation between
evaporation and plant succession in
a given area, I61

Gauliheria shallon, 154

GREEN, Newton B. The use of the
vibration galvanometer with a 60-
cycle alternating current in the meas-
urement of the conductivity of elec-
trolytes, 411

GROSSENBACHER, J. G. Crown-rot of
fruit trees: histological studies, 477

Gunnera petaloidea Gaud., a remarkable
plant of the Hawaiian Islands, 33;
manicata, 33

Gymnosporangium globosum, producing
galls on cedar, 246; Juniperi-virgin-
ianae, producing galls on cedar, 241

Halorrhagaceae, 35

Hanson, HERBERT C. Leaf-structure
as related to environment, 533

Harris, J. ARTHUR. On the distribu-
tion of abnormalities in the inflores-
cence of Spiraea Vanhouttet, 624; and
LAWRENCE, JOHN V. The osmotic
concentration of the tissue fluids of
Jamaican montane rain-forest vegeta-
tion, 268

Harvey, R. B. and True, R. Hoy Phe
influence of light and _ chlorophyll
formation on the minimum toxic con-
centration of magnesium nitrate for
the squash, 407

Hawaii, Gunnera petaloidea occurring in,
33; phytogeography of Manoa Valley
in, 561

Hawaiian species of the genus Cyrtandra,
section Cylindrocalyces Hillebr., 604

HAWKINS, Lon A. and STEVENS, NEw E.
Endothia pigments I, 336

Hippocratea volubilis, host of Botry-
orhiza, 48

HvuBERT, ERNEsT E.

Ree7)

(See Weir, James

INDEX TO VOLUME IV

Humidity, effect of, on development of
rusts, 188

Immunochemical studies of the plant
proteins: proteins of the wheat seed
and other cereals, 417

Influence of certain climatic factors on
the development of Endothia parasitica
(Murr.) And., 7

Influence of temperature on the growth
of Endothia parasitica, 112

Inheritance of endosperm color in maize,

396

Jamaican montane rain-forest vegeta-
tion, 268

Juniperus scopulorum, duration of leaves
in, 149

Kalmia poltfolia, 154

KNupDSON, Lewis. The toxicity of
galactose and mannose for green plants
and the antagonistic action of other
sugars toward these, 430

La Rue, Cart D. and BaRTLetT, H. H.
Matroclinic inheritance in mutation
crosses of Oenothera Reynoldsu, 119

LAWRENCE, JOHN V. (See Harris, J.
Arthur, 268)

Leaf-structure as related to environment,
533

Ledum groenlandicum, 154

Light and chlorophyll formation, in-
fluence of, on the minimum toxic con-
centration of magnesium nitrate for
the squash, 407

Light, effect of, on development of rusts,
I9I

Linnaea americana, 153

Lutz, ANNE M. Fifteen- and sixteen-
chromosome Oenothera mutants, 53

MacCauGHEY, VAUGHAN. Gunnera
petaloidea Gaud., a remarkable plant
of the Hawaiian Islands, 33; The
phytogeography of Manoa Valley,
Hawaiian Islands, 561

639

Macrosporium, taxonomic characters of,

439

Magnesium nitrate, toxic concentration

of, 407

Mains, E.B. The reiation of some rusts
to the physiology of their hosts, 179

Maize, inheritance of endosperm color in,
396

Mannose, toxicity of, for green plants,
4390

Manoa Valley, Hawaii, phytogeography
of, 561

Matroclinic inheritance in mutation
crosses of Oenothera Reynoldsti, 119

Micromeria douglasit, 153

Mikanta cordifolia, odoratissima, hosts of
Endophylloides, 51

Mineral salts, effect of, on development

of rusts, 190

Murritt, Witt1Am A. The taxonomy

of the Agaricaceae, 315

Mutation, from Xanthium canadense, 43

Observations on forest tree rusts, 327;
on an Achyla lacking sexual reproduc-
tion, 354

Oenothera albida, 72; bipartita, 63, 64;
gigas X Lamarckiana, 68; Lamarck-
1ana, 65; Lamarckiana X gigas, 67;
lata, 55, 59; /ata X gigas, 66

Oenothera Reynoldsu, matroclinic inheri-
tance in mutation crosses of, 119;
pedigree of, 122, 123

Oenothera mutants, with fifteen and
sixteen chromosomes, 53

OLIVE, E> Wy and:tWHETZEL;« HE.
Endophyllum-like rusts of Porto Rico,
44

Omphaha chrysophylla, development of,
222

Opuntia megacantha, 572

Origin and development of the galls pro-
duced by two cedar rust fungi, 241

Osmotic concentration of the tissue fluids
of Jamaican montane rain-forest vege-
tation, 268

Overwintering of rusts, 332

640
Oxycoccus oxycoccus intermedius, 155

Pachistima myrsinites, 155

Payson, EDWIN BLAKE. The perennial
scapose Drabas of North America, 253

PEASE, VINNIE A. Duration of leaves
in evergreens, 145

Perennial scapose Drabas
America, 253

Phytogeography of Manoa Valley, Ha-
waiian Islands, 561

Picea sitchensis, duration of leaves in, 149

Pigments of Endothia, 336

Pinus contorta, monticola, duration of
leaves in, 149

of North

Proteins, immunochemical studies of
wheat, 417

Pseudotsuga taxifolia, duration of leaves
in, 150

Pteridium aquilinum pubescens, host of
Uredinopsis, 331

Puccinia Scleriae, 46; substriata, 46;
coronata and Sorght, used in experi-
ments on relation to physiology of
their hosts, 185

Pucciniastrum pustulatum, 334

Raphanus sativus, viability of seeds of,
as affected by high temperatures and
water content, 229

REDDICK, DONALD. Effect of soil tem-

perature on the growth of bean plants >

and upon their susceptibility to a root
parasite, 513

REEs, CHARLES C. The rusts occurring
on the genus Fritillaria, 368

Relation between evaporation and plant
succession in a given area, I6I

Revision of the Hawaiian species of the
genus Cyrtandra, 604

Rhamnus purshiana, duration of leaves
in, 152

Rhododendron californicum, 155

Rock, JosepH F. Revision of the
Hawaiian species of the genus
Cyrtandra, section Cylindrocalyces

Hillebr., 604

INDEX TO VOLUME IV

Rubus laciniatus, pedatus, ursinus, 153,
154

Rusts, of Porto Rico, 44; relation of, to
the physiology of their hosts, 179; of
forest trees, 327; on cedar, producing
galls, 241; occurring on the genus
Fritullaria, 368

SAWYER, W. H., Jr. The development
of Cortinarius pholideus, 520

Self-fertile plants of Cichorium, 375°.

SHULL, CHARLES A. An _ interesting
modification in Xanthium, 40

Spiraea Vanhoutter, distribution of ab-
normalities in, 624

Squash, used in light and chlorophyll
formation experiments, 407

Stachytarpheta cayennensis, dichotoma,
hosts of Endophyllum Stachytarphetae,
50

Sterility of seeds of Oenothera Reynolds,
125

STEVENS, Net E. The influence of
certain climatic factors on the develop-
ment of Endothia parasitica (Murr.)
And., 1; The influence of temperature
on the growth of Endothia parasitica,
112; (see Hawkins, Lon A., 336)

Stout, A. B. Fertility in Czchorium
intybus: the sporadic occurrence of
self-fertile plants among the progeny
of self-sterile plants, 375

Succession, relation between, and evap-
oration, 161

Taxonomy of the Agaricaceae, 315

Taxus brevifolia, duration of leaves in,
150

Temperature, effect of, on development
of rusts, 187; affecting viability of
seeds of radish, 229; effect of, on
susceptibility of bean plants to a root
parasite, 513; effect of, on growth of
bean plants, 513

Thuja plicata, duration of leaves in, 150

Toxicity of galactose and mannose for
green plants and the antagonistic

INDEX TO VOLUME IV

action of other sugars toward these,
430

Transpiration of Fraxinus, 533

Treviso, a red-leaved variety of chicory,
376

Weun, R. H.: (See Harvey, R. B., 327)

Tsuga heterophylla, duration of leaves in,

I5I

Uredinopsis pteridis, on Abies grandis,
329; on Pleridium aquilinum pubes-
cens, 331

Uredo Trichiliae, 46

Uromyces aecidiiformis, 369; Fritzlariae,
272). Holway1,~ 360; Miurae, 371;
proéminens, 46

Vaccinium ovatum, 155; parvifolium, 152

Valerianodes cayennensts (syn.), 50

Viability of radish seeds as affected by
high temperatures and water content,
229

WaaGconer, H. D. The viability of
radish seeds (Raphanus sativus L.) as
affected by high temperatures and
water content, 299

641

Wedelia trilobata, host of Endophyllum
Wedeliae, 49

WEIMER, J. L. The origin and develop-
ment of the galls produced by two
cedar fungi, 241

WEIR, JAMES R. and HUBERT, ERNEsT E.
Observations on forest tree rusts, 327

WEsToN, Wm. H. Observations on an
Achlya lacking sexual reproduction,
354

Wheat protein, immunochemical studies
of, 417

WHETzEL, H.H. (See Olive, E. W., 44)

WHITE, ORLAND E. Inheritance of
endosperm color in maize, 396

WopEHousE, R. P. Immunochemical
studies of the plant proteins: proteins
of the wheat seed and other cereals.
Study IX, 417

Xanthium, an interesting modification
in, 40
Xanthium canadense var. globuliforme, 41

Zea Caragua X California Golden Pop,
inheritance of endosperm color in, 396

Zoological and Botanical Supplies—AII Kinds
Biological Specimens—Every Description
Instruments for Plant Physiology

Field and Herbarium Equipment
of Standard Stock

Driers, Mounting Papers, Genus Covers, Plant Presses
(Samples of Paper Matled on Request)

3 Microscopical Accessories of Highest Grade
COMPLETE LABORATORY EQUIPMENT—Physics, Chemistry, Botany

CAMBRIDGE BOTANICAL SUPPLY COMPANY

3 Lexington Street WAVERLEY, MASS.

HUGO DE VRIES

Species and Varieties, Their Origin by Mutation
Edited by D. T. MacDougal. $5.00.
Plant- Breeding

Comments on the Experiments of Nilsson and Burbank. Illustrated.
Cloth, gilt, $1.50.

The Mutation Theory

Experiments and Observations on the Origin of Species in the Vegetable
Kingdom. (2vols.). Numerous illustrations, colored plates. Translated
by Prof. A. B. Farmer and A. D. Darbishire. Cloth, per volume $4.00.

Intracellular Pangenesis

Including a paper on Fertilization and Hybridization. Translated from
the German by C. Stuart Gager. Cloth, $3.00.

Special offer. The above five volumes sent prepaid for $10.00.

OPEN COURT PUBLISHING CO. Chicago, Ill.

Always mention the JOURNAL OF BOTANY when answering our advertisements.

BOTANICAL SUPPLIES

During the past year the Marine Biological Laboratory Supply Department
has sent botanical specimens to Scotland, England, Holland, China and Canada,
as well as to all parts of the United States. Let us supply you also.

At the present time the following material can be supplied:

Marine or fresh water algae, preserved or mounted, or fresh during their
season; fungi, representing all the important groups, for study or demonstra-
tion; complete life cycles of four genera of mosses, and sporophytes of several
more; several genera of ferns; seed plants such as Pine, Zamia, Capsella and
sometimes Ginkgo and others.

Sixty or more species of Pure Cultures of Fungi are available.

During the past year we have worked out and put up in Riker Mounts life
histories of Pinus, Marchantia, Sphagnum, Zamia and others; also groups of
Mosses, Liverworts, Slime Moulds, and Lichens. These mounts are of great
value in the class room as they can be passed around a class without being
damaged, and with proper care will last for years.

From October to May the Department makes shipments of live Fucus for
demonstration of fertilization. The past fall we have made a number of
successful shipments as far as the middle west.

Zoological and Embryological supplies also furnished. Catalogs of all material’
and other information sent on request. Address all communications to

GEORGE M. GRAY

Curator of the Supply Department Woods Hole, Mass., U.S. A. -

No Botanical Library No Thorough Botanist

Can afford not to have

The Vegetation of The New Jersey
Pine Barrens

By JOHN W. HARSHBERGER, Ph.D.

Professor of Botany, University of Pennsylvania

This book is the result of twenty-five years’ careful study of the region.
It covers the subject from every point of view and is an important and
permanent contribution upon one of the most interesting botanical fields
of the world.

There are 340 pages, large 12mo, 284 illustrations and a large colored
Map. Price by mail, $5.00

Christopher Sower Company, Publishers
124 North Eighteenth St. PHILADELPHIA

Always mention the JOURNAL OF BOTANY when answering our advertisements.

ELECTRIC SLIDE WARMER

FOR MOUNTING PARAFFINE SECTIONS

21070. Slide Warmer

Slide Warmer for Mounting Paraffine Sections, designed by Prof. S. Walter Ranson, of
the Northwestern University Medical School, as a convenient method for the heating of slides
in the mounting of paraffine sections slowly and evenly over their entire extent to a tempera-
ture just below the melting point of paraffine, giving a maximum flattening without damage
to the section.

The apparatus consists of anelectrically heated block of transite 24 inches long by 6
inches wide by + inch thick on a brass frame 3 inches high and fitted with an electric thermostat
so that the temperature can be held constant at any desired temperature. The heating is accom-
plished by a series of nichrome wires embedded in the block of transite, and the apparatus may
be attached to an ordinary lamp socket. The surface upon which the slides may be placed
measures 24x 6 inches and will accommodate 40 slides at once. It is only necessary to place
the sections in position on the albumen-water covered slides and set the slides on the appara-
tus. The temperature is gradually raised and the paraffine spreads without melting.

With paraffine of 55° Prof. Ranson has obtained the best results when the regulator was
set so that the thermometer in the apparatus registered 43°. The thermometer is placed just
below the block of transite and registers a temperature somewhat less than that on the upper
surface of the block. It is not necessary or desirable to remove the water from the slides after
the sections are flattened as it evaporates in a few minutes. The time required for drying can
be cut down from 24 hours at room temperature to 2 or 3 hours at a temperature just below
the melting point of paraffine. A light, ventilated metal cover is placed over the apparatus
while the slides are drying, to protect them from dust. The apparatus requires no further
attention until the slides are sufficiently dry to go into xylol.

No. 21070 Electric Slide Warmer, complete with metal cover, $33.00 net

ARTHUR H. THOMAS COMPANY

IMPORTERS — DEALERS — EXPORTERS

LABORATORY APPARATUS AND REAGENTS

WEST WASHINGTON SQUARE PHILADELPHIA, U.S.A.

Always mention the JOURNAL OF BOTANY when answering our advertisements.

Bausch & Lomb
Large Dissecting Stand

A new design which meets every requirement by reason of its large size and
wide range of usefulness. Accommodates interchangeably Binocular Micro-
scope body (illustrated), Monocular erecting body using medium high power
objectives, or simple lens in jointed arm—all focused by rack and pinion.
The microscope bodies may be mounted on a sliding track permitting the en-
sire width of the stage to be covered.

Stage, 8x 7 inches, is provided with glass and metal plates, 44 inches in
diameter, and with adjustable background stops beneath.

With 38 and 19 mm. doublet lenses...........ccccceeeeee $26.50
With Binocular Microscope body, 40 mm. objectives and
10X eyepleces. cel. json ee oe ee Deere: $70.00

Write for Complete Illustrated Circular

Bausch £9 Jomb Optical ©.

NEW YORK WASHINGTON CHICAGO SAN FRANCISCO

LONDON ROCHESTER, NY. FRANKFORT

Awarded Grand Prize for Microscopes at Panama-Pacific Exposition

Always mention the JOURNAL OF BOTANY when answering our advertisements.

THE NEW ERA PRINTING COMPANY

LANCASTER, PA.

Is prepared to execute in first-clas8 and satisfactory
manner all kinds of Printing, Efectrotyping and
Binding. Particular attention given to the work of
Schools, Colleges, Universities and Public Institutions

Books, Periodicals
Technical and Scientific Publications
Monographs, Theses, Catalogues
Announcements, Reports, etc.
All Kinds of Commercial Work

Publishers will find our product ranking with the best
in workmanship and material, at satisfactory prices.
Correspondence solicited. Estimates furnished.

THe New Era Printinc COMPANY

Always mention the JOURNAL OF BOTANY when answering our advertisements.

The Diatomaceae of Philadelphia
and Vicinity

By Charles S. Boyer, A.M., F.R.M.S.

A description of the diatomaceous flora of Philadelphia and the region within a
radius of one hundred miles, including the pre- and post-glacial deposits occurring in
the blue clay of the ancient Delaware River has just been issued in book form by the
J. B. Lippincott Company.

143 pages. More than 700 original illustrations, enlarged 800 times,
with complete index. Size 9x12. Bound in Art Vellum cloth.
Sent postage paid $5.00.

CONTENTS

The cell; growth and reproduction; function; motion; classification
and description of more than five hundred species, many of which
are new; appendix containing directions for collecting, preparing
and examining material.

ORDERS MAY BE SENT TO

CHARLES S. BOYER
6140 Columbia Avenue . PHILADELPHIA

The Moss Flora of New York City and Vicinity

By A. J. GROUT, Ph.D.
Member of the Botanical Society of America and Fellow of the A.A A.S.

A carefully annotated list of the mosses found within approximately fifty miles of New York
City, with localities and time of fruiting. It lists over 380 species and varieties. 8vo, 121 pp.,
12 plates.

Nominally a list, in reality it is a local manual carefully prepared and up-to-date. ‘The
key alone will be invaluable to any student of the mosses of northeastern U.S. $1.10 postpaid.

By the Same Author

MOSSES WITH HAND-LENS AND MIcROScOPE. Bound in Cloth, $7.00, in half leather,
$7.50. 8x11 inches, 416 pages, 88 full page plates and 265 figures in the text. Printed by
McFarland on the best coated paper.

MosskEs WITH A HAND-LENS (224 pp., 8vo, copiously illustrated) describes 250 mosses
and 50 hepatics that can be identified by means of the Hand-lensalone. $1.75.

For SALE. Loeske, Studien sur vergleichenden Morphologie, $1.50. Cavers, Inter-rela-
tionships of the Bryophyta, $1.50. MecVicar, Student’s Handbook of British Hepatics, $5.00.
Muscologia Recentiorum Bridel, Tom II, Pars 1, 208 pp. but lacks plates and Index, 1798.
Very rare, $1.00. Catalogue of Plants within 30 miles of Amherst College. Tuckerman and
Frost, 1875, $4.00. Tuckerman, Synopsis N. Am. Lichens, Parts 1 and 2, $15.00. Fink
Lichens of Minnesota, 8vo, 51 full page Plates, mostly photographs, 1910, $3.00. Farlow,
Marine Algae of New England, $2.50. Science, 1898 to date, price on application.

Odd numbers and complete volumes of the Fern Bulletin, Plant World, Mycological Bul-
letin, Muhlenbergia, Bartonia, Bot. Gaz. (running from Vol. 12 to Vol. 52); Bulletin Torrey
Bot. Club, Vols. 1-6 and others, and practically every American botanical publication. Will
also buy and exchange any American botanical publication. Prices on application.

A. J. GROUT, Ph.D., Author and Publisher,
Tel, 412-J New Dotp. New Dorp, Richmond Borough, New York City.

Always mention the JOURNAL OF BOTANY when answering our advertisements.

SIMPLEX WATERPROOF LABELS

FOR ALL HORTICULTURAL PURPOSES = \iwit\
USED FOR PERMANENT AND TEMPORARY RECORDS

Adapted for Botanical Tests and for Conservatory Marking.

These Labels are also Permanent when used for Outdoor Garden

Work.

The writing will last indefi-
nitely if done with ordinary lead
pencil, but may easily be erased
as often as desired. Quickly and
easily marked and always legible.

PLANT LABELS GARDEN LABELS

PRICES, INCLUDING COPPER WIRES
: FOR ATTACHING ON STAKES 18 INCHES LONG
ig 1. ese 3x2 wmehies, 25 cents See $2.00 per 100 No. 19. Card size 23x1} inches, 15 cents each
o. 2. Size 4x{% inches, 50 cents doz., $4.00 per 100 . . ;.
No. 3. Size 5x1 inches, 75 cents doz., $6.00 per 100 gene ancbes,t20c-bte each
Postpaid By express f. 0. b. New York

STEWART & CO., 5 Cortlandt Street, New York

American Journal of Botany
ANNOUNCEMENT

On and after January 1, 1917, the subscription price of the AMERICAN
JOURNAL OF BOTANY, to those not members of the BOTANICAL SOCIETY OF
AMERICA, will be $5.00 a volume, of ten numbers. Single copies, 60 cents,
plus postage.

Subscriptions received before March 1, 1917, will, if accompanied by
an order for back volumes I-III (1914-1916), be entered at $4.00 for one year,

or $16.00 for the four volumes.
The Editorial Committee.

Please address all communications concerning subscriptions and adver-
tising as follows :

American Journal of Botany
Brooklyn Botanic Garden Brooklyn, New York

Always mention the JOURNAL OF BOTANY when answering our advertisements.

NEW SPENCER

CONVERTIBLE BINOCULAR
MICROSCOPE NO. 58

CONVERDIBEE::

For Examining Large Objects on a Table.
For Examining Vertical Surfaces.
For Examining High Objects.

Inclination Joint.

Stage removable.

Long range of upward and downward move-
ment.

Adjustable arm covering entire stage.

Mirror adjustable in all directions.

Accommodation for pupillary distance.

Automaticshutter cutting off light to either eye

Patented solid objective mount holding lenses
in absolute alignment.

ADAPTABILITY - USABILITY - EFFICIENCY

AT ITS BEST

Spencer Convertible Binocular

SPENCER Microscope No. 58

equipped with one pair objectives and
Spencer Lens Comp any one pair oculars in cabinet. $72.00
BUFFALO, N. Y. 54 Discount to schools.

Illustrated circular M | sent on request

Clarke Memow Paper

A Permanent Record Paper with Half ‘Tone
and Color Printing Qualities

CLARKE & COMPANY

731 Brunswick Building 225 Fitth Avenue
NEW YORK

Always mention the JOURNAI. OF BOTANY when answering our advertisements.

: American

colog

rs

oe Geawcna ‘Suggeatig

aah Wi wait to build the latéer house ‘at the top,’ when you can |
‘¥ now build. the smaller’ ‘one at the bottom, and enlarge it later

Roa fs That's our suggestion. Ue ack at ee ten

ee “To which let us add, U-Bar. greenhouses ¢ are e distinctly different |
| “from others. | Different i in construction, laihe in results, ‘t i

”

Office so guclirie tee Ce Leh Sat ih of ee ae

How the U-Bar greenhouse differs i is . told in ‘the eatalog,
for it., Or send for us. Or Doth eae ee a ACCA G Cae

U-BAR GREENHOUS
PIERSON U-BAR CO.

ONE, PADI AVE. NEW YORE.

¥ i al

eS MT
BER aa a
> pil s A | To ri i) ‘ites : ab iol ee
AMG unt y ‘angaroal cae a’ yet “iy i, Mee! ia
annnap? ; qani yar? eR?” hpt A, » - F May ‘apialal |
Rr AA, aii? bP Myar ALTA RT SH TA te Maa a pes Asa:
tO alae {bea wane ota 8? PN TLIFYT Ee Aams aE , te Mane NRG *< -2Py-. | ‘e PR y
| eras Wess Pee tee | | Ty ymae a wtaaeagiye pal || ae ae ait |
4 oF FF i Ps 3 ~ a «4% - hu “iP r
i ay iam Nery “" fal a “ria tent Aaa tn wha sects fl Ly y a Trt tial
y *h rm Rename terry cee aa A Yas cet Mee, - Wen. ha RTT i Bg
\ly RATT 6 Aan So Wace’ wecas | | aa
w A Ah te Asse 0 NDT pants we (any Nira aay ee Apt
Nis Ren Lh) i PTT || Lapeees AnH
h ~ 2h mee ?
yk

hei’. PF Mvngg ° sAngep’

> ® 727 4 eA ye ry. ~

aes apis ay

4 > hm Vee ii
a, <2 sliviiileemr i 1 eee | \euw |g
Attn HEA a nen aad
af ae
~ o> ei >»
ne Nala, n WNusidibey“tRgany, AAR aaay *” “hayes ' Naa”: Lit} | |
ae _ Ma, geste, pone ™ ILL le wane “hy Ayes 5 | triiit =a wh MAMA!
lek wimp, s Rynyyenyare YP WN Aanarrebags. 92: - : F ~~ aut ras “.
hl. p u iS a. 4 - ' ‘ =o
RAR ae rt, sAayesteiat ayes AY rants dannal sr RG eaten tye
Pout. hau . CAMA gs tAara. Se vey, Selannaa) MaRS BI
~ tt nats Dioercets ts piehnanle PYwevey von Away v AP Wenan, 8 Pym ss ae A
tia ac ane i iia iai* a " Tee Hy itt
al ay “ea Saas! ME hanna Mneytean,, Hey trier TET MRBAA Anan, Ro 4, | wy
agAp ha sre N a Te ee a ae we ati p RAAAS “WAya- Sip ebhe | Bip Nines |
ale Ban Ph anes” bo Nees Any A Mi OO ee a
aise ror oo. eauniaibe ol ees A ay © ON Bagge”
Sm i, age, 7A 5 ve 1} a A, b
mn Aa! | | | ot aan aanieitlodl a ) eofiten |||! | |
Reta ey / bib 9 v4n4 w YN m Hts Ny ‘+ aA- pre,
wy ak da Nhe amy DL ele NAP SAA a. iy se -e oY | iched ,
t 4 ~ fp 2 - z a AWA
| Ny * aes Niiapiiipihhnin treet. he a ie areas = ies. tN wenn |
ee ix Ty ayn pipe AAR AKA ENA Wanye rT tbe
rete |||! reacts a- - 4 fh ne LETTE, }
we \e eg PP REPEL dit m sr aetttnaagy aa! NY eet ~te Fy
tax aw Nasr md mm =” Pe
ran aYT TT TC) Teo! |

Satin Mey eh i aqmee 5 PT LAL TAL re Ua it

P an a © Aina .e k ae AY)
ni. vate YT Lolot FFP ea eta ven inn YS vrpalilihet git teled
° ni BM dus AA yams ARR TVET NLT]
} ae Mean Aaah) ene AGB EN eke S - a
| Mya oO Bigag 91 39° anand, WAN ag Duceett en
AN nr. | 1 a yaad A
: a. Be ’ te ge of yan" reo HET a8) Ml Le Serre

~
\Aae-s a7 & Rabe AL A “pense B94 Br

“yews npr ETE Mas 5s aia FR

Ber e ay ry

.

prnrered mers J auee” Siam,

iy piARpe =!
ait Ryeyyny ver’ | A 1}
sau? peared ag 22

wy Vale

ae hone ww Pn A vy yw faataratag eh paaei off £¢ me TILL

AALS] ean hae je Regs Sa
PT me WOR i BiWhs 1 ee eNey ay v..
Whi RdAdsnec’ taeenene ee LUTRAID RR “il
Mele Ty ‘ yale MARY a yeaan.” 1 eet naseenae nalts }
mal (| {ii |) Ste ‘ sarc AAP. ap P Pw ae oR AK ae

phe sdro” ony wate 77h bby AAA ieee

initiate esas sal, wet

a
PO eS a eytgetee syne anh

im Later ww yyw"
aa OA a paP
* Mada Sree Tan oe INNA ot
ao DL ys smalalinlies rianse7Mhe
nae eal TL" maAN Vw
au P| X PETe UTS Agwa, | Y aint Vise 1)
her a Mei tod pervs! #: ap- ub? Aaja”
War piv Worse age yl - i y le seta be”
aims . ww "¢ rm ar
p A AOR yt Tibadtt af” py” VAP ~2 409949) é: we asf 4AUN A\ \ wy es sides.

| it 1 en
y haha. _. Wied pf Resa ; 29QGeOOr fETEP EY habe eb f ale mea Fun 4 \

SS
—_—_— my
——___

—o¥
i Tee er iii,
—r7 F FT , Wat Apa, 24 wma | 1 4) ry a A aaa ;
— edema t AR ea TTR TATRA EREPEE DAR Wit aati
—- rae re Titi ppdgp hort op OPO in easat poo
—s besanttgaan’ 4 . sqapey alse: pandapeerne jyvsrbrp og, gy Mee
3 hae Ohne Bl | ay SRL dada a pia seunrencennthyle pone | Digits '
“hina. Aine eli eMiem | (tT! (Tew ideal hor Pe
bockenen| smb, Ma ry *- stig , > Dela set tH ee en peeien PUTE
G2’ en “ : Seb eal es my me in @ 494 I TN x » PO REBEL a Ws) +i bes |
| 5 3 iy 400 DPubiea AM jyAtane A as, pash as
A wv yee we "Oe &

ne BOP OC aad Ulan

la —
1 Ww

a 5
aa wrser
A Nastia at dg Mae eabecrarcet

, v v «a + *
Ade St AD SIN pre Toly aves wee Ab. SJ wh AL

Nee + dan ty Ww pis Pager We
ent aa sid dined tore

a coe iar [ nee iF ts ae ry a. (el b veers Nee eet rsvanrta
Ad ii nde put? maby: aa Brig Neh, has AA fe ~, Ht

Pyrery eae 228g ¢
mane iy ae <T4ag" TE ARELOR ena Ry ce My
NY cy ALLE Db oe. . ae st M ve rs a a 9 : 3 4 a an 3 : F., fey a \e
: 6. PA AAA a

-
7

or

ae oe cae

‘ie . 4 ——

ahh sauyirmnenen

LH sania vacate ate
hs 1al’ mane, ew “ae 2 ene
epely

! Vypbar nn . eer
ba, “NG Kyi ¥ rary, Posen! es : f-

wall Lyi | MAY wn -
98t Menn mts

™ wey”, Tr ©
fee need Bikaas Wa Ve femieeees

ri tel | sna eResaetieeea: Vyii | aris te
a iiRe p wonanaa sine? a |
orl 8 I BeePcCuDE LLeemenenniEe’ c=

aed 4.1 |

~apneac