/o//_ /9

Vol. IV 1911 No. I

CONTRIBUTIONS

FROM THE

Botanical Laboratory

OF THE

University of Pennsylvania

THE SWEET POTATO

By B. H. A. Groth. A.B., Ph.D.

University of Pennsylvania
Philadelphia
1911

PUBLICATIONS

OF THE

UNIVERSITY OF PENNSYLVANIA

CONTRIBUTIONS FROM THE

BOTANICAL LABORATORY

OF THE UNIVERSITY OF PENNSYLVANIA

Vol. IV. No. 1.

THE SWEET POTATO

BY
B. H. A. GROTH, A.B., Ph.D.

LIBRARY
NEW YORK
BOTANICAL

QAKUfsiN.

UNIVERSITY OF PENNSYLVANIA
D. APPLETON AND COMPANY, Agents, NEW YORK

1911

/■ \

Copyright, 1911
15 Y THE University of Pennsylvania

The John C. Winston Co.
Philadelphia

s

CONTENTS.

PAGE

I. Origin and Histoky

A. Origin

B. History

C. List of References "

OK

D. Historical Summary

E. List of Synonyms "'

33
II. Economic Importance

A. Distribution and Comparative Yield ^^

B. The Sweet Potato as a Starch Producer 36

C. The Sweet Potato as a Sugar Producer 45

III. The Structure of Ipom(EA Batatas

A. Root ^^

B. Stem "^^

C. Leaf ^^

57

IV. Classification of Varieties

A. Popular Varieties •.

B. The System of Classification ^^

C. Key to Varieties '^^

D. Alphabetical List of Formulas ^^

E. Alphabetical List of Varieties 82

85

F. Descriptions

V. List of Illustrations

PREFACE.

This work was undertaken with a \'iew to sift-
ing the historical records of Ipomoea batatas and
furnish a working basis for future experimenta-
tion with its many varieties. It is now recognized
among botanists in general, and physiologists and
agriculturists in particular, that for obtaining reli-
able data, the variety and not the species must form
the basis of experiments.

I wish especially to acknowledge my indebtedness
to Dr. John M. Macfarlane for general directions
and valuable suggestions.

Herewith I present this work as a thesis offered to
the University of Pennsylvania in partial fulfilment
of the requirements for the degree of Doctor of

Philosophy.

B. H. A. Groth.

Philadelphia, June, 1906.

The Sweet Potato.

CD

or

<r

I. ORIGIN AND HISTORY.
A. Origin.

There are tropical shores where cocoanuts and
bananas do not thrive; there are tropical districts
where the sweet potato is unknown. Both are excep-
tions. In the far East, where rice is the daily food
of millions ; on a thousand isles of the Pacific, where
bananas and cocoanuts are the common fare ; in the
lands of the Aztecs and Incas, where beans are the
national dish, the sweet potato is universally culti-
vated. Even in the great country where "Corn is
King" and wheat the "Staff of Life" it has been
for hundreds of years a favorite among the vege-
tables.

But that has not always been so.

The sweet potato has spread over the entire
tropics and sub-tropics within the last four hundred
years. Whence it came is a much-discussed ques-
tion. There are two opinions regarding its origin,
one that it is American, the other that it is Asiatic.
De Candolle, after summing up the arguments in
favor of both, concludes that it is probably Ameri-
can, but does not commit himself. He has often been
criticised as basing his opinion upon inadequate
evidence. With a view to ascertaining what con-
clusions the evidence at hand really warrants, the

2 THE SWEDT potato.

writer has briefly stated below De Candolle 's princi-
pal arguments, with a few additional ones.

The contention that the sweet potato is of Asiatic
origin is based upon the following facts :

1. D'Hervey, Saint Denis (Rech. sur I'Agri-
culture des Chin., 1850, p. 109) states that the
Chinese Encyclop. of Agric. speaks of the sweet
potato and mentions different varieties long before
the discovery of America.

2. Piddington (Index) claims that the sweet
potato has the Sanskrit name ^'ruktalu."

3. Cook, on his first voyage around the world in
1769, found sweet potatoes cultivated at Tahiti, and
in 1770 large plantations of them in New Zealand.
(Low, Cook's Three Voyages Around the World,
pp. 45, 76).

Forster (De Plantis Esculentis, p. 56) in 1783
describes Convolvulus clirysorhyzus as being culti-
vated there, which Hooker (Handbook of N. Z.
Flora) identifies as the sweet potato.

Pickering (Chron. Hist, year 1273) states that he
saw varieties unknown in America cultivated on
Metia, Tahiti, the Hawaiian, Samoan, and Tonga
Islands.

The contention that the sweet potato is of Ameri-
can origin rests on the following:

1. Columbus found it in Cuba on his first voyage
(F. Columbus, 28; Gomara, 16).

2. No reliable reference to the sweet potato, dated
prior to 1492, has been discovered, as

(a) No reference exists which is accompanied
by an accurate description.

(b) The existing references may as well apply
to the yam as to the sweet potato.

(c) The sweet potato mentioned in the Chinese

THE SWEET POTATO. 3

Encyclopedia prior to 1573 was not the cultivated
sweet potato which was introduced into China
between 1573 and 1626. The old Chinese sweet
potato was known as "cliu," the introduced sweet
potato as "an-chu." The information was obtained
by Bretschneider from Chinese records of the six-
teenth century, i. e., the time of the introduction.
(Bretschneider in a letter to De Candolle, after read-
ing "De Candolle" on the origin of the sweet potato.
Quoted in footnote 2, Or. des PI. Cult., end of
article).

3. The word ''ruktalu" seems a Bengalu name
composed of the Sanskrit ' ' alu, ' ' which is the name
of Arum campanulatum, and "rukta," of unknown
meaning and derivation (A. Pictet in a note to De
Candolle). The word "alu' is now employed to
mean yam and potato. Eoxburgh mentions no San-
skrit name; and Wilson's Sanskrit dictionary does
not give "ruktalu."

According to Watt (Diet, of Econ. Prod, of India,
under Ipomcea batatas) the present-day vernacular
contains, among others, the following:

Mita-alu, Sweet potato, Hind. ; Eanga-alu, lal-alu,
lal-shakarkand-alu (the red form) ; shine-alu (the
white), Beng.; Goria-alu (white form), Assam;
Gajar lahori (Lahore Carrot), Sind.; Lardak lahori
(Lahore Carrot) , Pers. All of these names compare
the sweet potato to another tuber or root, which is
certainly a better argument in favor of the opinion
that it was introduced than against it. The author
has not been able to find out the meaning of the
word "rukta," but its combination ^ith the word
"alu" suggests that it describes some quality of the
sweet potato, in which it differs from the arum.

4. The fact remains that the sweet potato was

4 THE SWEET POTATO.

found by Cook and others in Tahiti and New Zealand.
According to native tradition, it is not native to
New Zealand, but was introduced from the direction
of Tahiti or Samoa (Pickering, Chron. Hist., years
1273 and 1740). Its introduction is estimated from
indirect evidence to have occurred at about 1740
(Pickering, Chron. Hist., under 1740).

Captain Cook, on his first voyage, took on board
a native of Tahiti, by the name of Tupia, who could
readily converse with the natives of New Zealand and
other islands of that region. From this it appears
not only that the New Zealanders and Tahitians
were of the same stock, but that their separation
must have been comparatively recent. But if it
was recent, it could not have been accidental or due
to shipwreck, as then both New Zealand and Tahiti
could not have been so well populated as Cook found
them. So it must have been due either to a whole-
sale emigration or to a gradual expansion. But
neither mode of separation would account for the
striking fact that the New Zealanders were, up to
1740, without so important a food plant as the sweet
potato. The significance of this becomes the more
apparent when we consider that bananas, cocoanuts
and yams, which yield staple foods in Tahiti, could
be of little importance in New Zealand. In all prob-
ability, therefore, the Tahitians became acquainted
with the sweet potato only after they had separated
from the New Zealanders for so long a time that
communication between them had practically ceased.

It is possible, of course, that the sweet potato was
native in Tahiti. In that case it must be supposed
that communication by sea between two great
branches of a seafaring people stopped at once after
separation. The only argument known to the

THE SWEET POTATO. 5

author which would support such a supposition is
Pickering's statement that he observed varieties
unknown in America under cultivation on Metia,
Tahiti, the Hawaiian, Samoan and Tonga Islands.

The author has obtained eleven native varieties of
sweet potatoes from Hawaii, some of which do not
differ markedly from some of the varieties received
from Jamaica, although they can be easily distin-
guished from any of the varieties common in the
United States. Pickering gives no description of any
of the varieties. As he was unable to distinguish
from the cultivated plants in Tahiti what he calls,
* ' . . . seemingly the same species springing up spon-
taneously, usually as a weed in cultivated ground, but
distinguished by the natives, and its roots not used"
(Cliron. Hist., under 1273), it seems that he made no
thorough investigation, but made his statements in
regard to the varieties from casual observation.
Moreover, Pickering himself calls the sweet potato
a native of tropical America (Chron. Hist., under
1273).

Should the varieties of Tahiti or Samoa prove
distinct from American varieties, that would in itself
be no proof that these islands should be considered
the native countries of the sweet potato, as the
plant varies readily and had ample time to vary.

There is then really no proof that Tahiti or the
neighboring islands should have originated the
sweet potato, except that it was found there as early
as 1769.

But long before that time it had spread through
tropical America. It had been reported from Cuba
by Oviedo, from other West Indies by Sloane and
Hughes, from Surinam by Merian, from Brazil by
Marcgrav and Piso, from Peru by de Vega.

6 THE SWEET POTATO.

If the sweet potato were a native of Tahiti it
would have to be assumed that the Tahitians, who
did not supply the New Zealanders until about 1740,
had brought the plant to tropical America so early
that it had spread across the continent before 1492.
This seems highly improbable.

It seems very probable, on the other hand, that
the sweet potato was carried from South America
across to Tahiti. If we had any tradition, name, or
other evidence, pointing to an early intercourse
between tropical America and Tahiti, that would
decide the question as completely as one could expect
at this late date. Such proof we have. The writer
is not aware of any tradition in regard to the sweet
potato, but a similarity in names is apparent.
According to Forster, the natives of Tahiti and New
Zealand called the sweet potato ''gumarra," and
according to Markham (Trav. in Peru, p. 234) and
Seemann (Journal of Bot., 1866, p. 328) it is called
"cumar" by the Quichuen tribe near Quito, Ecuador.

Besides, we have ample proof that other economic
plants, such as cocoanuts and bananas, have been
carried from South America to the Pacific Islands,
and vice versa. (The reader is referred especially
to Cook, Bulletin Food Plants of Ancient America,
Bureau of Plant Industry, United States Depart-
ment of Agriculture).

The writer is of the opinion that the evidence pre-
sented warrants fully the conclusion that the sweet
potato, Ipomoea batatas, Lam., is a native of tropical
America.

B. History.

Among the many references to the sweet potato
there are some which have been so often cited and

THE SWEET POTATO. 7

some wliicli sum up so well what was known at the
time about the plant that they must be of great
interest alike to the student and to the reader.
JBelow the writer has given quotations, translations,
and extracts from such important references, in
chronological order.

For over a hundred years after the discovery of
America the sweet potato was rarely mentioned in
botanical literature. Here and there some casual
remarks are encountered about sweet roots, sweet
potatoes or yams, which are not definite enough to
be considered of value here.

The first thorough account is given by Clusius
(Hist. Plant., published 1601, second part, p. 77).

He distinguishes three types: The Camotes,
Batatas, and Inhames Lusitanorum. Under Batatas
he explains that there are three kinds, which he has
found growing in Baetica (a f>i'ovince of south-
western Spain), differing in color. Some were red
or purplish outside (and these were prized most),
others of a paler color, others white. A few had
white meat. They had vines spread diffusely over
the ground like those of the wild cucumber, and
rather thick, succulent leaves, from green to gray, in
shape resembling those of spinach. He could not
learn from anybody whether the plants ever bloomed
or bore fruit.

The roots, which he had seen in London in 1581,
he describes as being usually ''clodrantalis," weigh-
ing a pound or more, uneven, with two or three or
more fibers growing from one head, with roots sim-
ilar to those of "Siseris," thicker at the lower end,
and becoming more slender at the upper end,

Clusius says that the sweet potato originated in
the New World and was brought first over to Spain,

THE SWEET POTATO.

and was at that time iutroduced into many maritime
districts of Baetica. The sweet potato as raised at
Malaca (a city in Baetica) was considered the finest,
and had been previously exported to Cadiz, Spain,
and Ulosipons.

Specimens taken to Belgium would not sprout and
soon spoiled before he could plant them.

He doubts that the plant was known to the
ancients, and knows no Greek or Latin name for
them. He gives the names Batatas, Camotes,
Amotes, and Ajes, which he hears do not differ,
except, perhaps, that the Batatas have longer and
more tender roots.

The Inhame, which he describes next, is certainly
the Dioscorea batatas. He had several specimens
a foot or more in length and four inches in diameter,
and he describes them as being all rough, like the
long roots of Aristolochia, and tasting at first pleas-
ant, when eaten raw, but becoming after a short time
somewhat sharp,

Clusius gives three figures, one of the roots of
''camotes" still hanging together at the top, one of
the Batatas vine and roots, and another of the
Inhame Lusitanorum, and his figure of the Inhame
shows a yam.

In 1623, Bauliin (in his Pinax, p. 91) refers to
the sweet potato as batatas, battades, and potatoes
called camotes in the West Indies. He describes
three types, all radish-shaped, but much larger, dis-
tinguished by their exterior coloring, which may be
either purplish, pale, or pure white. He states that
it serves as a common food to the "negroes" in the
West Indies, is used in Angra, and is mentioned by
Linschoten as occurring in the East Indies, where it
takes the place of fruit and vegetables. Linschoten 's

THE SWEET POTATO. 9

potato was the yam, however (Watt, under Ipomcea
batatas). As his authority, Bauhin gives Joseph
Acosta, lib. 4, cap. 18; Frag. 9, 1, 4, e. 18; 6, 38; 3,
6, 4, cap. 12; 8, 9. '

In 1636, Gerarde (Herball, pp. 925-930) calls the
sweet potato Sisarum Peruviauorum, Batata His-
panorum Potatus, and Potato. His illustration is a
copy of the "Batatas" of Clusius. According to
him, the plant was known as the "skyrrets" of Peru.
A few plants grown in his garden did not flower,
and he states distinctly that "not any said anything
of the flowres." The roots were "many, thicke and
knobby . . . joined together at the top into one
head.'"' His tubers were bought at the Exchange in
London, and he states that the sweet potato grows
in "India, Barbaric, Spaine and the hot regions,"
and that they are common food among Spaniards,
Italians, and Indians. He has quite a selection of
recipes for making palatable dishes from them.
AVhether he meant India or the New World with his
India, the writer does not know.

In 1640 Parkinson gives a pretty full account of
the sweet potato, as then known (Theater of Plants,
pp. 1382-1383). Under "Pappas, batatas, Potatoes,"
he translates a large part of Clusius, as given before,
but he does not at all keep separate the account of
"Camotes" from that of "Inhames," and so he
makes it appear as if Clusius had applied the same
name to both. Later on he states that Lobelius
(in Adversaria) says that the Inhames brought from
Aetliiopia and Guiney were different from the pota-
toes of Spain and the Canary Islands. Parkinson
uses the name "Virginia Potato." He quotes
Scaliger as saying, "The Spaniards know three
other sorts of roots besides the ordinary, which will

10 THE SWEET POTATO.

abide good without perishing for a whole yeare, and
therefore they use to bring them to Sea with them
and call it Igname cicorero. ,The other will last
nothing so long."* The writer has not seen Scali-
ger's statement, but the name Igname coupled with
the statement that it keeps longer than the other
sweet potato, seems to point to the conclusion that
Scaliger refers to the yam. He then relates some-
thing about the methods of culture as practiced at
St. Thomas. He speaks of planting pieces of the
vines, and not tubers, and that the vines ran up
poles, like hops. Perhaps he had the accounts of
the sweet potato and the yam mixed, although some
of the varieties of the sweet potato are excellent
twiners. That he was not any too sure of his ground
is seen from his closing words: "This manner of
planting the Inhame favoureth something of that
of the Manihot or Yucca, whereof the Cassavi is
made, if there be not a mistake, it is wonderful that
the roots should be so propagated."

In 1648 Marcgrav (Historia Plantarum, part II,
p. 16) gives a very good account of the sweet potato,
as cultivated in Brazil under the names of '' Jetica"
and "Quinquoa quianputu." He points out the
variability in the shape of foliage and tubers, and
illustrates it. He is familiar with the habit of the
plant to form tubers at each rooting joint, and states
also that it yields latex. According to Marcgrav the
Brazilians added a little water to the freshly macer-
ated tuber to make it ferment into an alcoholic drink.

He has seen some tubers golden-white without and
clean white within, others red throughout (so much
so that on cooking they colored the hands), the out-
side dark red. Marcgrav remarks that a surface
freshly cut by a knife becomes black like ink. So

THE SWEET POTATO. 11

far as known to the writer, this occurs only when a
rot has attacked the tuber.

In 1686 John Eay (Hist. Phant. Tomus I, p. 728)
describes the sweet potato under: Convolvulus
indicus, Battatas dictus. He quotes Clusius and
Marcgrav largely, and also says that the roots, when
cut, give off latex.- He thinks it must be a species
of Convolvulus, not so much, he says, on account of
the similarity in foliage and vines and in the pos-
session of latex, but because a certain Fr. Hernandez
painted and ascribed to this species, which he (F. H.)
called Cacamotic Tlanoquiloni or Batatas purga-
tivum, malvaceous flowers, i. e., ' ' Caliculorum forma
vel cjinbalorum. "

Ray cites Marcgrav as describing other species
under the names of ''Omenapo," **Yeima Brasilien-
sibus" and "Pararo," the first a white variety,
which turned red on boiling, the last with purple
stems and veins. As he gives no exact reference,
the author has not been able to verify this state-
ment. Marcgrav, however, does not include such
descriptions under the headings of Jetica or Batatas.

In 1688 Eheede (Hortus Malabaricus, p. 95, pi.
50), under Kappa-Kelengu, Batatas (Bramannice)
Cananga, Lusit. Batatas, Belgice Pattates, describes
the sweet potato as blood-red outside and flesh-
colored within, watery, of a somewhat sweetish taste ;
with hairy, rough stems, watery within; petioles
oblong, round, green, falcate, hairy, more or less
rough ; leaf -bases surrounded by roots ; leaves cuspi-
date, incised around base, thin, soft, glabrous, dark
green above, light below; the midrib giving off two
lateral veins, from which others spring, arching in
near the margin, protruding on the lower side.

Eheede states distinctlv that this is the ' ' Camotes

12 THE SWEET POTATO.

Hispanorum" of Clusius and Bauliin (and the Getica
of Piso).

He gives an excellent illustration which tallies
with his description.

In 1696 Plukenet (Almagestum Bot., p. 114) men-
tions the sweet potato under: Convolvulus Indicus,
Batatas dictus, following Ray. .He also gives the
names of Batatas Occidentalis Indiae, Inhame Ori-
entalis Lusitanorum, Park., and Conv. Indie, radice
tuberosa eduli, cortice rubro, Batatas dictus, P. B. P.
326.

In the same year (1696) Commelin (Flora Mala-
barica) mentions the sweet potato, but simply gives
references.

In 1705 Merian (Insectes de Surinam, p. 41)
describes under Battattes a climbing plant, with light
red tubers which taste like chestnuts. Her illustra-
tion shows a sweet potato vine with blue flowers,
twining through a rosebush. Each branch, bending
down to the ground again, takes root. Here then
we have a proof that the twining habit was observed.

In 1707 Sloane (Jam. Nat. Hist., Vol. I, pp. 150-
151) describes the sweet potato thoroughly. In his
reference he quotes Plukenet as naming it Conv.
Malabaricus, which is wrong. Plukenet in the place
cited (Aim., p. 114) calls it Conv. Indicus. He says
that the plant flowers in Jamaica. "The Leaves
stand on five inch long green Foot-Stalks. They are
almost Triangular, having two Ears and a sharp
point opposite to the Foot-Stalk. They are five
Inches broad from Ear to Ear, and three from the
Foot-Stalk's end to the point, having under them
purple Eibs, being soft, smooth and of a yellowish
green color, something resembling the Leaves of
Spinage. The Flowers come out ex alis fol. stand-

THE SWEET POTATO. 13

ing on a three or four luclies long, green, Foot-
Stalk, being monopetalous, Bell-Fashion 'd, not very
open, purple within, and whitish without, having in
the middle some Stamina and a Stylus. After each
flower usually follows one Seed, brown and having
several depressions in it. It is enclosed in a
roundish, brown, membranaceous Capsula, under
which stand five brown capsular withered Leaves,
as in the other Convolvuli.-" This is perhaps the
first description of the fruit.

The red potato, "which differs from the white in
nothing but the color, is as common as the white and
grows indifferently with it." He gives an account
of the culture, which may perhaps best be given in
full :

''They are everywhere planted after a rainy
Season in the Plantations, for Provision by the flip,
a piece of the Stalk and Leaves, being put either into
the plain Field after Rowing or into little liillocks
raised through the Field, in which they are thought
to thrive better. In four months after planting they
are ready to be gathered, the ground being filled with
them, and if they continue therein any longer they
are eaten by worms." We find then that even at
that time there were advocates of both hill and flat
culture.

He proceeds as follows: "They vary very much
as to the figure and bigness of the Eoot, the color
of its Skin being sometimes red and most commonly
white. They are sometimes turbinated, at other
times round and most commonly biggest in the mid-
dle and tapering to both extremes."

"They are boiled or roasted under the ashes and
thought extraordinary good and nourishing Food,
and because of their speedy attaining their due

14 THE SWEET POTATO.

growth and perfection, they are believed to be the
most profitable sort of Root for ordinary Provision."

"They are used in great quantities to make the
Drink called Mobby."

Further on, he says that "They are common at
Velez-Malaga, whence ten or twelve Caravels are
loaded with them every year to Sevil." He gives
Thevet as authority for the statement that "people
feed with them in Trinidad" ; and Lopez de Gomara
for: "These Eoots were by Colon brought from the
West Indies into Europe, in his first voyage, to show
the different Productions of the one and the other."

Under: Convolvulus radice tuberosa, esculenta,
minore, purpurea, Cat., p. 54, batatas Ind. Or., part
6, p. 85, Red Spanish batatas, he describes a very
young plant of another tuber, as large as one's
finger, of deep red or purple color, resembling the
sweet potato in foliage, and containing an abundance
of latex, which dyes of a purple color." No sweet
potato is known to the writer in which the latex is
colored anything but white, and no other reference
has been found to a latex which dyes purple. The
purple color of the purple stems and tubers of some
varieties, however, does stain when they are handled.

Feuillee, 1725 (Hist, des Plantes, Tome HI, p. 16),
accepts Ray's nomenclature. Convolvulus Indicus,
vulgo Patates dictus, and states that the sweet pota-
toes were quite common in Europe. He compares
their flavor to that of the chestnut — which is cor-
rect— and says that they are common and in use
throughout America.

Catesby, in his Nat. Hist, of Carolina, defines the
sweet potato as: "Convolvulus radice tuberoso,
esculento — the Virginian Potato. He resided in the
Carolinas from 1723-1726, and was very familiar

THE SWEET POTATO. 15

with the actual couditions of agriculture there. As
he was commissioued especially to gather useful
information about plants in the colonies, and as he
resided for years in the very district in which sweet
potatoes were a staple crop, his observations are cer-
tainly entitled to the greatest weight. His remarks
are therefore given in full :

"This excellent root seems to merit the prefer-
ence of all others, not only in regard to the whole-
someness and delicacy of its food, but for its more
general use to mankind than any other root, it being
one great part, if not the principal subsistence of
the greater part of Africa, and is likewise in great
use, both in America and in the Southern parts of
Asia. They being of so easy culture, so quick of
growth, and of so vast an increase, that the propa-
gating it seems more agreeable to the indolence of
the Barbarians than cultivating grains, which
requires a longer time, with more labor and uncer-
tainty. In all our Colonies of xVmerica, as well
Islands as Continent, these roots are in great esteem
and use; the common White People, as well as the
Negro slaves, subsisting much upon them, nor are
they thought unworthy a place at principal tables.
In Virginia and to the North thereof they are
annuals and produce no flowers. They plant them
in March and dig them up in October, and to prevent
their rotting, keep them in holes underground near
their fires. In Carolina, where the winters are more
moderate, they are not necessitated to keep them so
warm; and in Bahamas Islands, and other places
between the Tropicks, they are perennial and pro-
duce flowers, yet are annually planted. The most
kinds and best potatoes that I have observed were
in Virginia, and because the names they are called

16 THE SWEET POTATO.

by, in different Colonies, are so various, I shall call
them by those names only by which they are known
here" (Virginia). "I have observed only five
kinds of Potatoes specifically different from one
another: The Common, the Bermudas, the Brim-
stone, the Carrot and the Claret Potatoes.

"The Common Potato is of a muddy red color on
the outside, but being cut appears white with a
reddish cast: they commonly weigh from half a
pound to four, five or six pounds, usually are long,
irregularly shaped and pointed at both ends: this
is an excellent kind and is most planted.

' ' The Bermuda Potato is larger and rounder than
the Common, very white within, and covered with a
white skin: this is a tender kind, requiring more
warmth in keeping, and a different culture from the
rest : this is the most delicate sort, but not so much
planted as the Common Potato, because of its not
keeping so well. This potato only produces a white
flower, the flowers of the other kind being purple.

' ' The Brimstone Potato grows to a large size, and
is shaped like the Common; the colour of it hath
given its name, and in goodness it is esteemed next
to the Common.

"The Carrot Potato is named so from its color
both without and within being like a carrot: these
grow to a very large size, and are of great increase,
though of little esteem, being the most insipid.

"The Claret Potato seems to be propagated more
as a curiosity than for any peculiar excellence it
hath. The colour of it, without and within, is that
of claret, ' '

It should be noted that Catesby based his notes
on tho sweet potato culture in the Carolinas solely
on his own observation. He is the first to describe

THE SWEET POTATO. 17

the method of storing tlieii used, which differs not
very much from tlie metliod of storing empk^yed in
the South at present.

He was aware that tlie plant was really a per-
ennial, and forced into the hal)it of an annual only
by tlie shortness of the season.

Catesby makes the first attempt at a classification
of varieties. From his description it is, of course,
not possible to tell exactly which varieties he picked
out as standards, still his "Conuuon I^otato,"
''Brimstone Potato" and 'H/arrot Potato" may well
have l)een varieties like the l\ed Jersey, Up River
and Pumpkin, respectively. A variety which would
corres])ond to his ''Claret Potato" is unknown to the
wi'iter.

When Catesby tells us that the sweet potato was
one of the princii^al food ]ilants of Africa, he is. no
doul)t, mistaken, as it was not introduced into Africa
until later.

That with such good descriptions it could be pos-
sible to present the figiire which Catesby gives is
hard to understand. He shows a vine with the
flowers of a Convolvulus, with a rough tuber like
that of a yam, leaves like a Smilax (or Dioscorea),
and a tendril at the growing tip. It could not be
the misplacing of a i)icture, as such a picture could
surely not fit any plant. As we shall see later, his
peculiar figure has given rise to Meyer's synonj^m,
which still persists in some books, namely, "Con-
volvulus Catesbaei."

In 1737 Burmann (Thesaurus Zeylanicus), under
Battattas, gives a resume of the literature bearing
on Batatas, but he does not state that he has seen
the sweet potato himself, and some of the references
he cites refer to the vam. He does not decide on

18 THE SWEET POTATO.

any specific name, but accepts that it is a creeping
Convolvulus.

In 1737 Linnaeus (Hortus Cliff ortianus, p. 67)
describes the sweet potato as follows :

' ' Convolvulus. ' '

^'3. Convolvulus foliis cordatis angulatis, radice
tuberosa.

Convolvulus radice tuberosa esculenta, spinaciae
foliis, flore albo fundo purpureo, semine post singulos
flores singulo. (Sloan flor, 53).

Convolvulus indicus. Batatas dictus (Ray, Hist.,
728).

Convolvulus indicus orientalis. Inhame sen Bat-
tatas, Sisarum peruvianoruni sen Battata hispan-
orum (Moris. Hist., 2, p. 11, f. 1, t. 3, f. 4).

Batatas (Bauh., pin., 91; Bauh., Hist., 2, p. 792;
Clus., Hist., 2, p. 78).

Jetica (Marcgr. bras. 16).

Kappa-Kelengu (Rheed, mal., 7, p. 95, t. 50).
Crescit culta in utraque India vulgaris; foliis
variat. ' '

Under 1740, in Pickering's Chronological History
of Plants, appears the following: "In the time of
Teraraku (great-grandfather of Pomare) the chief
seen by us at the Bay of Islands (Hale ethnogr.
Expl. Exp. 146, and Races of Man, IV, 4), the
'kumara' (batatas edulis) sweet potato, brought to
New Zealand in a ' canoe formed of separate pieces '
by Pani and his sister Hinakakirirangi of Hawaii
(Savaii)." The account is confirmed by the *' con-
struction of the canoe, peculiar to Samoan Islands,
by the slender finger-rooted variety, seen by us only
in the two localities," and which a separate tradition
made "the only kind formerlv known in New
Zealand."

THE SWEET POTATO. 19

In 1762 Linnaeus (Species Plantarum) describes it
as follows: "Batatas."

"Convolv. foliis cord., liast. quinque-nervis, caule
repente hispido tuberifero.

Conv. radice tub. escul., Catesb.

Conv. indicus vulgo Patates dictus, Ray.

Batatas, Bauliin. pin.; Kumph., Kalm.

Kappa-Kelengu, liheede.

Habitat in India utraque.

Confer. Con. ind. vulgo Patates dictum.

Few per 3 p. 16 II foliis palmatis."

In 1784 Tliuuberg (Flora Japonica, pp. 84-85),
under Convolvulus edulis, describes it as grown
abundantly around Nagasaki, and states that it was
even then unknown in the higher parts of Japan and
that it had been introduced by the Portuguese. He
describes it as : A Convolvulus with entire and three-
lobed, heart-shaped glabrous leaves, stem creeping,
angular; Flowering rarely — he has never seen the
flowers — but not the same as Ipomoea triloba. The
roots were often the size of one's fist, tuberculate,
flesh-colored like Batatas; esculent, very soft and
well-flavored. The plant differs from Convolvulus
Batatas in having heart-shaped, entire leaves, 3 and
5-lobed, and in not being constricted in the center,
"ut sagittaria evadant."

In 1789 Linnaeus (Amoenitates Academicae, Vol.
VI, p. 121) says: "Conv. Batatas, Indiae occiden-
talis, tuberosa, peregrina. Editur assa sub cineri-
bus, coctione rubicunda evadit. Sapor praecedentis,*
etjam ergajstulocum cibus indis. Tubera per
liyemem ab omni humido studiose praeservanda."

In 1793 Loureiro (Flora Cochinchinensis, p. 107)

* Before goes Dioscorea, of which he says : "Siccior optime sapit
et frequens illis est."

20 THE SWEET POTATO.

describes it under Convolvulus Batatas Klioai
Hoanxy, and states that it is found both in Cochin-
china and in China. He says it flowers annually not-
withstanding Osbeckius's statement to the contrary
(It. ad. Ind. edit. Angl., p. 3117), and that it grows
equally well on both sides of the Ganges.

Eoxburgh (Flora Indica, Vol. II, p. 69), under
Convohnilus Batatas, states that ''the red sort is in
very general cultivation all over the warmer parts
of Asia, and very deservedly esteemed one of their
most palatable and nutritious food. I suspect Con-
volvulus edulis Thumb. Japan. 84 is the same or a
variety. ' '

Choisy (De Candolle, Prodromus, Part IX, pp.
338-339,"^ and Choisy Conv., p. 53), 1824-70, gives the
following descriptions, with numerous references:

"Batatas edulis (Prodromus), caule repente raro
volubili, foliis A-^ariis saepius angulatis etiam lobatis
2-6 pollices longis acutis cordatis petiolatis, pedun-
culis petiolum aequantibus aut superantibus
3-4-floris, sepalis aeuminato-mucronatis raro sub-
truncatis exterioribus i^aulo brevioribus, corolla
campanulata purpurea. Ex India oriental i nata,
fere ubique in tropicis regionibus culta ob radicem
tuberosam edulem."

"Variat:
a 1. radice purpurea aut alba.

2. omni parte nunc glabra nunc hirsuta.

3. caule, petioles et pedunculis purpura-
scentibus.

4. foliis hastatis, anguloso-sinuatis, ([uincpie-
fidis aut 5-partitis, nunc etiam integris.

5. longitudina peduncula rum. ' '

/?. xanthoi'hiza, radice lutes (Bat. zanthorhiza
Boj. h. maurit., ]). 225). Ilanc varietatem el. Bojer

THE SWEET POTATO. 21

vulgo dici i)atate jauiio in siio libello niouet, Patate
junot ill (losiccatis specMmiiiibus ; orta ex ludia, China
aut Cociiiucliina eulta in ins. ^lauritii. Caeterum
a B. eduli non earn sejiingere possuinus (v. s. cult,
eomm. a el. Bojer).

y. platanifolia, foliis palmatim 3-5-fidis, lobis
ovato, rotuudatis, aeuminatis acnmine obtuso g-laber-
rimis; petiolis villesulis, pednneulis longis multi-
floris. In Gnvana britann. legit Schomb. n. 701

(V.O.).

Desci'i])tion (from Clioisy. Conv.) :

Radix tiiberosa, pro alimento in India, Japonica,
China, America colilur. Caulis prostratus. Folia
cordata petiolata acuta 2-6 pollices longa; i)etioIus
sequalis. Pedunculi ascendentes incrassati, pedi-
celli 2-0 lineas longi. Sepala ovato-lanceolata.
Corolla pollicaris glabra. Stamina et stylus dimid-
ium corollae attingentia.

C. List of References.
All references marked with an asterisk (*) have
])een looked up by the author. The year refers to
the publication of the book or the date of observa-
tion, when known. The author does not assume any
responsibility for the correctness of the references
not marked with an asterisk. An (?) denotes doubt
that the reference is to Ipomoea batatas.

lAcosta (C), Aromatum, India, 1582.

Acosta, J., lib. 4, cap. 18. Frag.
tAhimann, Character Plant, naturalis. 211. Yam(?).

Atl-inson, E. I.. Economic Products, Northw. Prov. (India), Pt. V,
p. 19. 1870.

Baden-Pouell, Ph. Pr., 250 (India), 1872.3 (?).
"Bailey, Cyclop, of Am. Hortic. nnder Iponifpa batatas.
"Banhin, Pinax. teatr. bot., 91. 162.3.

Bnuhin, Hist. Plant.. 2. p. 792. 16.50.

22 THE SWEET POTATO.

Benzo, under "Hayas''( '')•

Bertero, Spec. ^Medic. nonnullse.

Biet, p. 334.

Birdwood, Cat. Econ. Prod., Bombay, 1862.

*BIa»co, Flor de Filipinas, Vol. 1, p. 53; Vol. 1, p. 129; Vol. IV,
p. 140, Manila, 1837.

Boissier, Voyage botanique en Espagne, 1839-45.
*Bouton, p. 47, Medic. PL, Mauritis, 1857.
*Boyer, Hort. Maurit, 225, 1837.

Bretschneider, Study and value of Chin. Bot. Works, 1870.
*Bretschneider, Letter to DeCandolle, De. C. Orig. d. P. C. C.

Browne, Nat. Hist Yam, 155 ( ?), 1756.
*Burmann, Thesaurus Zeylanicus, 1737.

Burns (India).

Burr, Field and Gard. Veg., p. 99.

Campbell, Econ. Prod. Chutia Nagpur, No. 8163, 1873.

Candolle (De), Eapport sur les voyages botaniques, 1813.

Candolle (De), Arch, des sc, phys. et natur., 1882.
*Candolle (De), Orig. Cult. Plants, 53, 1882.
*Candolle (De), Prodromus, Part IX, 338-339.

Cardanvs, De Eerum. Var., 189, cap. IV, lib. VII, first part, 1556.

Carey & Wall, II, 69.
*Cateshy, Nat. Hist. Carolina, Vol. II, p. 60, 1731.

*Ghoisy,

*n 1 11 /n 1 Prodromus, 338 (1824-1870).
Candolle (De), ' ^ '

*Choisy, Convolvulaceae Orient, 53, 1834-41.

Clot-Bey (Egypt) (?).
*Clusius, Hist. Bot., 2, p. 77, 1584.

Columbus, F., 28.
*Commelin, Flora !Malabarica, 22, 1696.
*Cook, Three Voyages around World (Low), p. 45, 76.
*Cope d Kingsley, Am. Nat., Vol. XXV, p. 698, 1891.

Crawford, Migrat. of Cvilt. Plants, 1861.

Cruger, Bot. Ztg., 1854; Tap. II.— Trinidad Indust. Exh. Report
1853 (?).

Dale, Sam., Pharmacol. Suplem., p. 150, 1693.

Dampier, p. 10.

Descouriilz, El. Ant.. VIII. t. 545, 1821-9.

D'Hervey, Saint Denis, Eech. sur I'Agriculture des Chinoia, p. 109.

Drury, Useful Plants of India, 72, 1864.

*Engl. d Prantl, imder Ipomoea batatas, TV, t. 3, p. 30.
Esquemeling, p. 54.

THE SWEET POTATO. 23

*Feullee, Hist, des PI. Medic, de Perou., t. 3, p. 16, pi. 77, 1725.
Feullee, Jour. Obser., t. 2, p. 16 (ad calcem libri) ; t. 3(?) p. 16,

pi. 11.
Firm (India) (?).
*Fitz, Sweet Potato Culture. 1903.
Flacourt (De), IIG.

Furshahl. p. 54, Fl. -^^g>-pt. Arab.. 1775.
Forster, George, De Plantis Escul. insul. oc, p. 56, 8, p. 80( ?), 1783.

Gaudichmid-Beaupre, Botanique du Voyage, 1826 or 9 (?).
*Gerarde, Herball, 780, p. 925, 1579.

Gomara, 16.

Grant, Trav. in Africa.

Gray, Port. Voyages (Trav. in W. Africa (?), 1825).
*Grisebach, Flora of British W. India, 1864.

Hale, Races of Man, IV, 4.

Hale, Eth. Expl. Exped., 146.
*Hasselquish, Iter palaestinum, 1766.

Heme (Soudan) ( ?).

Hernandez, Fr.

*Hooker, Fl. Brit. India, 202, 1872.
*Hooker, 1867, Handb. of N. Z. Flora, 760.

Hughes, Nat. Hist, of Barbados, p. 97-256, 1750.

Humboldt, Nouvelle Esp., Ill, p. 81 ; Ed. 2, Vol. 2, p. 470.

Jacquin, Observationum, I, p. 39, 1764-71.

Kalm, It. N. Am., II, 300.
Kochef, Table, p. 48.
Krapf (E. Africa).

''Lamarck, Encycl. Method, VI, 14, 1783-1817.
*Linnaeus, Hort. Cliff., p. 427, 1737.

Linnaeus, Hort. Ups., 39, 1748.
*Linnaeus, Am. Ac, 6, p. 121, 1749-71.
*Linnaeus, Spec. Plant., 220, 1753.
* Linnaeus, Systema Veg.. 344, 1774.

Lishoa, Useful Plants of Bombay, 166.

Lohel, Stirp Ad-versarhia (Pena and Lobelius, 1570).
*Loudon, Encyclop. of Plants, 140, 1829.
*Loureiro, Flora Cochinchinensis, p. 107, 1790.

*Marcgrav, Hist. Plant., Part II, p. 16, 1640.
*Markham, Travels in Peru and India, p. 234, 1862.

24 THE SWEET POTATO.

Mart, Nalurg.

^Martins, | ,,^^^ ,.,.^,^., ^..f',,, ,,^, ■

Meissner, j

Martyr, Eden's History of Travel, 88, 143, 1577.
*Mason, Burmah and Its People.

Mercado, under eaniotes.
*Merian, Insect. Surinam, p. 41. 1705.
*Meyen, Geography of Plants, p. 318. 1836.

Meyen, Grundr. Pflanz. Geogr., p. 373, 1836.
*Meyer, Priniitiie FloriC Esseqn.. 103.
*Michanx, Fl. 13or. Am., I, p. 138. 1803.
"Miqitel, Fl. van Nederl. Indie. 2, p. 599, 1855.
*Miller, Gard. Dictionary, n. 7. 1731 (n. 5) (?).

Monardes, Hist. Medic, de las cosas, 1493-1578.

Morison, Hist. Plant. UniA'.. 2, p. ii; sect, i; t. 3, f. 4, 1G80.

Mozo (Philippines).

■yesMtt, Farm. Bnl. 129, U. S. Dept. of Agrie.
yienhoff, Brazil (?), E. Indies (?).

Oviedo, Eammusins TransL, 1526.

Oviedo, F., Hist. Gen. et Natur., part I, cap. IV, lib. VII.

Osbechitis, Iter ad Ind., edit. Angl.. p. 311.

Par. Bat. Prod. Mus. Zeyl., 325.
^Parkinson, Parad. in sole Parad.. 510. t. 517, f. 2.
"■' Parkinson, Theatrum, 1382 (1629).
"Payen, Des Substa. Alim., p. 178.

Phurmacographia (4th voyage of Columbus), 452, 1879.
*Pickering, Chronologic Hist, of Plants. 1879, years 1273, 1740.

Piddington, Index: (Euktalu), 1832.

Piso, Hist, de Indae utrinque, 254. 16 — .
. PlenJc, Icones Plant, med., p. 9. t. 106. 1788-1812.
""Phikenett, Almagestum Bot., 114. 169(i.

''Kay, Hist. Plant, Vol. I, 728, 1686-704.

Reich, 438(?). See Seemann.
*Rlieede, Hort. ilalab., VII, 95, t. 50, under Kappa-Kelengu, 1678-703.

Roem, Arch. 3, p. 141.
'^Roxburgh, Flora Ind., Ed. C. B. C, p. 162. publ. 1820-24; writ. bef.
1815.

Roxhurgh,. Flora Tnd.. Edit. ^Vall.. II, p. 69.

Roxburgh, edit. 1832. Vol. I. p. 483.

Rumphiiis, Herb. Amb.. V. 368. t. 130, lib. 15.

nil-. swKKi' po'i vm. 25

*Sa(hheck, Kulturgewiichse ikv tUiit-ilieii Koloiiien.

Huyot, Journ. Soc. D'lloitic. de FraiK-e, socoikI ser., Vol. V, 450.

,^aUiihunj, Prodi-.. 123, 170G.

Scnliger, Exeicitat, 181, cap. 17, lib. 15.

Sriiiniiburijk. British GiiyaiU', n. 7(H. lS4tt.

SeenifiiDK Jonni. of But., p. 328, ISlid.

Sloanc, Catal. Tlant. Jam., p. 53.

Sloaiic. Hist., 1, p. 152(?).
*Sl<)<iii»\ Jaiiiaic-a Natural Hist., I. 150, lODG.
*^milh. Dirt, of Eeon. Plants, 399.

kipr. Syst.( ?) I p., t>00.

Stade, H. (Brazil) ('!}.

Steinirt, Vuuy.xh Plants PL. PI.. 150. ISf,!).

Stocls. Aci-ouut of Sind.

Sweert. Florilo<,Miini( "r) . part 2. tal). 3r>. liiI2.

Thevet.
*riiiinh€i(f, Flora Japonica. p. 84. 1784.
*Thurbrr, Am. Woods and Lseful PI., ISf.O.
*Tschirch & Oesterlo Anat. Atlas dor Pliarmak<)<;ii.. 220.

Turpin, Mom. dor ]Mnsec, Vol. XIX, pi. 1. 2. 5(?).

Tussac, Flora Antillanim, 4, t. I; 8T. 5G5, 1808-27.

Vnger (Mexico) ( ?).

Vega (do). P»oy. Com. Ilak. Soc. Ed., II, 359.
*yellozo, Flora Fhimencnsis, t. 57, t. 58, t. 59, t. 60, t. 61, Rio
de Jan., 1827.
Tilmorin, Les Plantes Pot.(?) 401. 1865.

WalUch, Cat. 1356, 1828.

Wallich, Flora Indica (Pvoxb.) II. 69, 1830.
^Wait, Econ. Prod, of India. IpouKva batatas.

Watt, in Sel. Gor't of India, 1888-89, 147-154.

Wiesner, Mikrosc. Untersiich., p. 65, 1872.
*Wies«er, Eolistoffe under Batatas edulis.

Willrs. U. S. Expl. Exped., IV, p. 282, 1838-42.

Wilhl. in Linn. Spec, plant. I, p. 881, 853(?).

D. Historical Summary.
A short summary stating when the sweet potato
was definitely reported from different parts of the
world will doubtless prove of value to future workers

26 THE SWEET POTATO.

on tliis subject. Below has been given such a sum-
mary, as compiled from references found through-
out the literature consulted. The writer has
purposely omitted to state ^tlie sources in which these
were found, as most of them occur in many books,
and it would require much useless labor to deter-
mine which author gave a particular reference first.

The sweet potato was probably first recorded by
Oviedo, as found in 1492 in Cuba, and as being intro-
duced into Spain in 1526. From the West Indies it
was also reported by F. Columbus, Gomara, Sloane
1696, Hughes 1750, Tussac 1808, Descourtilz 1821-29,
Schomburgk 1840, and Grisebach 1864.

In Brazil it was observed by Marcgrav 1640, Piso
16— (?), Vellozo 1827, Martins 1829, Stade (?).
From Peru it is reported by de Vega, Feullee, Hum-
boldt, Markliam, and Seemann. In Mexico it was
seen by linger; Merian, 1705, describes it as culti-
vated in Guiana; Catesby, 1731, and Michaux, 1820,
report it as much grown in the Carolinas.

In Spain it was well known at an early time, and
was mentioned by Oviedo in 1526 as brought from
Cuba, by Cardamus in 1556, and Clusius in 1576.

By the Spaniards it was introduced into the East
Indies, where it was distributed by the Portuguese
(De Candolle, Orig. des Plant Cult.). In India,
where it was brought very early, we find successive
mention of it as follows: Acosta 1582, Osbeckius,
Burmann 1737 (Ceylon), Piso, Lisboa, Roxburgh
1820, Wallich 1828, Piddington 1832, Birdwood 1862,
Drury 1864, Stewart 1869, Hooker 1880, Watt 1872,
Baden-Powell 1872, Campbell 1873, Atkinson 1876,
V. Mueller 1880, Burns (?), Firm (?).

In the East Indies it is mentioned by Eheede 1678,
Commelin 1696, Rumphius 1750, Loureiro 1790,

THE SWEET POTATO. 27

Mason 1851, and Miguel in 1856. From China it is
reported by Bretsclmeider (as of about 1600) and
Loureiro 1790.

Tliuul>erg, 1784, mentions it in Japan, and Mozo
and Pickering, 1740, Blanco, 1837, in the Philippines.

In Hindostan it was observed by Roxburgh
1832 (?) and Burns (?).

In Arabia and Egypt by Forskal, 1<75, Clot Bey,
and in other parts of Africa by Grant (Egypt to
Zanzibar), Krapf (Eastern Africa), Henze (Sou-
dan), Gray (W. Africa) and Pickering (Zanzibar),
1740. From ^Mauritius it was reported by Boyer,
1837, and Bouton, 1857.

In the Pacific Islands it was seen by Captain Cook,
1769, in New Zealand by Hale, where native tradi-
tion puts its introduction back to 1740 (Pickering),
Forster 1783, Bertero, Wilkes 1840, Hooker 1867,
and Pickering. From Hawaii it is reported by
Gaudichaud-Beaupre, Hale, and AVilkes, 1840.

E. List of Synonyms.

In the following the writer has attempted to pre-
sent a complete list of all synonyms by which the
sweet potato has been referred to. Unless the
writer was able to convince himself of the correct-
ness of the synonym by consulting the original
authority, the later authority has been given in
parentheses.

Ahe (Choctaw), Gray (Cope & Kingsley).

Ajes, Clusius. This is doubtless the yam, although it is often given

as a synonym of the sweet potato.
Amotes, Clusius.

Apichu (Peruv.), de Vega (Cope & Kingsley).
Artichaut des Indes. Vilmorin (Cope & Kingsley).
Axe, Pharmacographia (Cope & Kingsley).

28 THE SWEET POTATO.

Batala (Cingalese), Birdwood (Cope & Kiiigsley).

Batala (Sing.), Watt.

Batata de Malaga, Thurber-Darlinglun.

Batata Hispanoium, Geraide.

Batata Lusitanis, Marograv.

Batata purgative, Fr. Hernandez (Ray).

Batatas anthorhiza, Runiphius (Hooker). Bojer (Meyer).

Batatas, Clusius (Malay origin), Birdwood (Cope «t Kingsley).

Batatas, Eatable, Thurber-Darlington.

Batatas edulis, Chois. DeCandolle, Prod., etc.

Batatas Oceidentalis Indiie, Parkinson (Plnkenet).

Batatas platanifolia, Sehomberg (Meyer).

Bataten-^Yinde (German), Thurber-Darlington.

Batates douces, Payen.

Batatas Lusitanorum, Rliccde.

Batatas, Sloane.

Battades, Acosta (Bauhin).

Battatas Oceidentalis India?, Parkinson.

Battatas. Parkinson and later writers.

Boga (red var. Assam), Watt.

Cacamotic, Humboldt. See Camote.

Cacamotic Tlanoquiloni, Fr. Hernandez (Ray).

Canange ( Bramannice ) , Rheede.

Camote, Meyen.

Camote (Aztec Cacamotic) . Humboldt (ileyen).

Camote, Clusius.

Camote Hispanorum. Clusius (Rheede, Ray).

Chelagada (Tel.), Watt.

Chillagada (Telinga), Drury (Cope k Kingsley).

Convolvulus angiilosis soliis, ^Malabaricus radice tuberosa eduli.,

quoted from Plnkenet by Sloane. Not given by Plukenet

under sweet potato.
Convolvulus batata, Lilley & Wait. (Veg. Subst. Used for Food).
Convolvulus batatas, L. Am. Ac, :Mason. Payen, Bhune, Wall

(Hooker Brit. Ind.). According to Hooker, Catesby also uses C.

batatas. Roxburgh uses C." Batatas Willd because Willd wrote

convolvuIace;e in Syst. Veg.
Convolvulus ohrysorhyzus. Forstor (DeCandolle. Or. d. PI.).
Convolvulus coi-datifolius, Vellozo, Fl. Flum.. 60.
Convolvulus edulis. Thunberg. who thinks it differs from C. batatas.

THE SWEET POTATO. 29

Convolvulus esculenta, which DeCandoUe quotes from Catosby, is a

contraction of Conv. rad. tub. esc.
Convolvulus esculentus, Salisb., Prod.
Convolvulus foliis cordatis angulatis radice tuberosa L. Ilort.. Cliff.

(;Meyer).
Convolvulus foliis cordatis hastatis quinque-nervis caulo repente

hispido tuberifero. L. Sp. PL; Mill. Diet., n. 7, Plcnk (Meyer).
Convolvulus indicus orientalis, Morison (Meyer).
Convolvulus indicus radice tuberosa ed\ili. cortifc nibro, Batatas

dictus. Par. Bat. Prod. (Commelin).
Convolvulus indicus. vuljro Patates dictus. By Lin. quoted from

Ray, by Ray from Clusius, who does not give it(?), quoted

by Meyer from Pculli'e. wTio does not give it( ?).
Convolvulus mayor heptapliyllus Sloane (Meyer).
Convolvulus mayor hcptaphyllus tloribus albis, fuixbi imrpureis,

Sloane (Meyer).
Convolvulus radice tuberosa eseulenta, Catesby.
Convolvulus radice tuberosa eseulenta espinachia' fol., llora alba,

Catesby, as quoted by Sloane; not mentioned by Catesby under

sweet potato.
Convohailus radiee tuberosa eseulenta minore purpurea, Catesby, 54,

quoted by Sloane; not mentioned by Catesby under sweet potato.
Convolvulus septangularis. DeCandolle, Prod.
Convolvulus tuberifer Stend., DeCandolle, Prod.
Convolvulus tuberosum, Vellozo, Fl. Flum., 57.
Convolvulus varius. Vellozo, Fl. Flum., 61.
Convolvulus xanthorhiza, DeCandolle, Prod.
Cumala (Otaheite), Cook (Cope & Kingsley).
Cumar (Quichuen). Seeman (DeCandolle. Or. des PI.).
Cumar, Markham (Cope & Kingsley).

Dankali (Soudan), Heuze (Cope & Kingsley).
Doukali (Soudan), Henze (Cope & Kingsley).

Fiasi (^Yanika-land), Krapf. (Cope & Kingsley).

Gajar lahori (Lahore Carrot, Siud.), Watt.

Genasu ( Kan. ) , Watt.

Getica (Brazil), Piso (Rheede).

Goria-alu (white variety, Assam), Watt.

Grasugada (Telinga), Drury (Cope & Kingsley).

Gumalla, Forster (DeCandolle, Origin des PI. Cult.).

Gumara, Forster (DeCandolle, Origin des PI. Cult.).

30 THE SWEET POTATO.

Haias (Benzo), Sloane.

Hantsoa (Chinese at Batavia), Nienhoff (Cope 4: Kingsley).

Hetich (Tupi), Gray (Cope & Kingsley).

Hoanxy (Cochinch.), Loureiro.

Igname cicorero, Scaliger (Parkins) (?). Probably the yam.

IgTiame, Clusius (Parkinson) (?) . Probably the yam.

Imo (Jap.), Thunberg.

Inliame, Clusius ( ? ) . Probably the yam.

Inhame Orientalis Lusit., Parkinson (Plukenet) ( ?). Probably

the yam.
Inhame rubra, Burmann ( ? ) . Probably the yam.
Ipomoea Batatas Lam. ( DeCandolle ) .
Ipomoea Batatas Lam. Flor. Brazil,
var. indivisa, Grisebach.
var. leucorhiza, Grisebach.
var. porhyrorhiza, Grisebach.
var. Xanthorhiza, Choisy.
Ipomoea batatas (L.), Poir. Mohr. Geol. Survey, Alab., July, 1901,

p. 828.
Ipomoea batatas, Poiret. v. Mueller.
Ipomoea Catesbaei, Meyer. This was given by Meyer simply because

Catesby's figure did not agree with Ipomoea batatas while the

description did. Meyer explains that fully in a note, which has

often been overlooked.
Ipomoea heptadactyla mayor, Brown, Jam. (Meyer) .
Ipomoea tuberosa, Willd, Sp. Plant (excl. syn. Pluk.) and Mill.

Diet. n. 5 (Meyer).

Jetica (Brazil), Marcgrav.

Jetica (Brazil), Piso (Commelin).

Jettika (Brazil), Stade (DeCandolle, Orig. des PI. Cult.).

Kan-chu, Bretschneider (DeCandolle, Or. d. PI. Cult.).

Kapa-Kalengu (Malay), Watt.

Kappa-Kelengu, Bheede.

Kara-imo (Japan), Thunberg.

Kazwan (Burmah.), Watt.

Kiasi (Canga). Hoist (Sadebeck).

Kindolo (Canga & Usambara), Hoist (Sadebeck).

Kimhella (Usambara) , Hoist (Sadebeck).

Kitaiti (Usambara), Hoist (Sadebeck).

Kitetta (Usambara), Hoist (Sadebeck).

THE SWEET POTATO. 31

Kumana (N. Zeal.), Wilkes (Cope A Kingsley).
Kuuiara (Tahiti, N. Z., etc.), Hale-Pickering.

Lal-shakarkand-alu (red variety, Beng.), Watt.
Lal-shukur-kunda-aloo (red variety, Beng.), Roxburgh.
Lardak-lahori (Lahore Carrot, Persian), Watt.
Lohita (Sung.), Roxburgh.
Lohitaloo (Sung), Roxburgh.

Maby (Antilles), Descourtilz (Cope <t Kingsley).
Mankutu (Usumbara), Hoist (Sadebeck).
]\Ia\vandres, Flacourt (Sloane).
Mita-alu (Hind.), Watt.
Myonk-ni ( Burniah ) , \\att.

Obi-djawa (Java). Ann. Jard. Buitenz., Vol. I, p. 77.
Onienapo Yeima, Marcgrav (Ray).

Pararo. Marcgrav (Ray).

Pappas, Parkinson.

Patales, Bouton (Sloane).

Patata, Esquemeling (Sloane), Vilmorin (Cope & Kingsley).

Patatas, Red Spanish. Sloane.

Patate Jaune (French) (Thurber-Darlington) .

Patates, Ray (Linn. Sp. PI. and Feuillee).

Patattes, NienhoflF (Cope & Kingsley).

Pattates ( Belgice ) , Rheede.

Potades, Gerarde (Cope & Kingsley).

Potates, Acosta (Bauliin).

Potato, Virginian, Catesby.

Varieties: The Common Potato.

The Bermuda Potato.

The Brimstone Potato.

The Carrot Potato.

The Claret Potato.
Potato, sweet, Gerarde.
Potato, Carolina, Thurber-Darlington.
Potatoes, Gerarde (Cope & Kingsley).
Potatoes, Spanish, Ray.
Potatos. Gerarde.
Potatus. Gerarde.
Pendaloo (Hindustani), Birdwood (Cope & Kingsley).

Quinquoa quianputu ( Congenibua ) , Marcgrav.

32 THE SWEET POTATO.

Ranga (red var., Assam), ^\'att.

Eatali (Mar.), Watt.

Ratalu (Bomb.), Watt.

Eizophora indiea, Burniann ( ?) . ProbaVdy the yam.

Rukta -kunda ( Suug. ) , Roxburgh .

Ruktaloo (Sung.), Roxburgh.

Ruktalu, Piddington (DeCandolle. Dr. d. PI.).

Ruktupindaloo (Sung.), Roxburgh.

Ruktupinduka (Sung.), Roxburgh.

Sakara-kenda (Sant.), Watt.

Sakaria (Guz.), Watt.

Sakka rei-vellei-kelangu ( Tam. ) , Watt.

Shakarkand (Punjab.), Watt.

Schumbalino (Usambara), Hoist (Sadebeck).

Shakarkand (Hind), Watt.

Shakarkand-alu (red variety, Beng.), Watt.

Shakar-kandu (Bombay). Watt.

Shine-alu (white variety, Beng.), ^^aU.

Shukar-kundo (India), Firm (Cope & Kingsley).

Shukar-kundoo-aloo (Bengali), Drury (Cope & Kingsley).

Sisarum Peruvianorum, Gcrarde.

Skirrets (Peru), Xieremljerg (Loudon).

Skyrrets (Peru), Gerarde.

Suffet-shukur kunda-aloo (Beng., white variety), Roxburgh.

Sukkarag-vuUie (Tamil). Birdwood (Cope & Kingsley).

Truflfe douce, Vilmorin (Cope-fc Kingsley).

Ubi (Java), Nienhoff (Cope & Kingsley).

Ubi-castela, Rumphius (Burmanu).

Ubitora (^laley). Xienhoff (Cope & Kingsley).

Umara (South Sea Islands), Forster (DeCandolle, Or. "d. PL),

Vallikilangu (Tam.), Watt.

Veeazee (C. Africa), Grant (Cope & Kingsley).

Ycam (Peru), Clusius(?). Probably the yam.
Yeti (Tupi-Guarani),'Gray (Cope & Kingsley).

Zardak-lahori (Persia), Birdwood (Cope & Kingsley).

THE SWEET POTATO. 33

II. ECONOMIC I.AIPORTANCE.

A. Distribution and Comparative Yield.

The sweet potato is primarily a tropical plant, and
it has found its widest distribution in the tropics.
As with other tropical products, so with this — we
know that it is one of the most important food plants,
but we cannot even approximately estimate the
amounts produced. Even in a country like Mexico
the statistics on agricultural products are so utterly
unreliable that they are of no practical value.

In the tropics the sweet potato is a perennial.
The tubers are constantly dug up without removal
of the parent plants, until replanting seems advis-
able. Since the immature tubers are as good for
use as the mature ones, the natives merely dig up
the largest potatoes, no matter whether they are
ripe or not. The large potatoes can usually be
detected by the cracking of the earth above them.
There is no part of the tropics now where several
varieties of sweet potatoes are not cultivated.

In Mexico, Central, and South America they are
a staple crop in all the states down to Argentine and
Chile.

In Africa they are cultivated largely by the
natives of all the European colonies as well as by
those of the interior. In Mediterranean Europe
they are a well-known crop. In Persia, Hindustan,
India, Farther India, China, Japan, and the entire
Malay Archipelago they form one of the principal
productions, utilized in many ways. In Japan the

34 THE SWEET POTATO.

tubers, cut into small pieces and roasted, can be
bought on the streets, like roasted chestnuts here.
Australia and New Zealand also cultivate the tuber
extensively.

But, of course, what concerns us most is what the
sweet potato is at present and what it ought to be,
in the United States.

In no other country is the sweet potato cultivated
successfully so far north as with us. That is princi-
pally due to our tropical summers. The plant
requires a hot, dry season to mature the tubers, and
such we can offer as far north as central New Jersey
and Illinois. In the Gulf States it continues to crop
throughout the summer and fall, if occasional rains
supply enough moisture. This means that whereas
in New Jersey the sweet potato under favorable con-
ditions has five months and a half in which to mature
the tubers, in the Gulf States it has seven and a half.
As new tubers are started on the roots at the joints
as well as on the main stem throughout the season,
it is evident that those which have not time to mature
in a short season and are so small as to be unsalable
(the ''culls," as they are called) would have matured
in a longer season. So it happens that in a favor-
able season the crop in the Gulf States may be easily
twice as large as in the Northern State, and, as the
quantity of immature tubers at harvest time even in
the South shows, a still longer season would have
matured still more.

Now there can be no question that sweet potato
culture is profitable in the North.

In New Jersey, where sweet potato farming is
practiced on as thorough a plan as am^where, a crop
of 150 to 200 bushels per acre, at an average price
of $0.60 a bushel, and an outlay of $60.00 an acre,
well repays cultivation. As to the crops which can

THE SWEET POTATO. 36

be raised in some of our Southern States, I refer the
reader to the Bulletins of the agricultural stations,
referring to tests of fertilizers and tests of varie-
ties. These can easily be found in the catalogue of
the Department of Agriculture.

We find there that crops of 400-600 bushels per
acre in a good season, and 300-500 in a dry season,
are not rare, and that 800-1,000 bushels per acre
have been obtained in experiment station work. The
writer does not for a moment suppose that 1,000
bushels per acre could be raised everywhere even
in the South, where the season is as long as at
Baton Rouge, but it seems that 500 bushels should
constitute more nearly the average yield of some
varieties than 60 to 120 bushels.

At present the yield per acre in our principal
sweet potato States is far below that.

The following table gives the total yield and the
number of acres cultivated in sweet potatoes, as
given in the United States Twelfth Census Eeport,
and the approximate yield per acre :

Bushels
Bushels. Acres. Per Acre.

North Carolina 5,781,587 68,730 84

Georgia 5,087,674 70,620 72

Virginia 4,470,602 40,681 110

Alabama 3,457,386

South Carolina 3,369,957 48,831 69

Texas 3,299,135 43,561 76

Mississippi 2,817,386 36,169 78

New Jersey 2,418,641 20,588 113

Louisiana 1,865,482 27,372 68

Tennessee 1,571,575 27,372 67

From this we can see that Louisiana, where the
heaviest crops have been raised on one of the
experiment stations, is almost at the bottom of the

36 THE bWEET POTATO.

list when it comes to the actual average crops raised
there. Now, while it is conceivable that the experi-
ment station at Baton Rouge may contain the very
best potato soil in the state, still the difference
between what can be done and what is done is so
enormous that one cannot but think that the average
could be made very much higher if all the potato
fields were given the necessary care. But even at
present the average sweet potato farming is, on the
whole, profitable in Louisiana, or it would soon stop.

B. The Sweet Potato as a Starch Peoducer.

Now, it is true that those varieties which have
given the largest yields are not usually the best
table varieties, although even on that opinions differ.

On the other hand, it is commonly stated that
those coarse varieties which do give enormous yields
are the ones richest in starch. Now the question
of the proportion of starch contained in the tubers
of different varieties grown under the same con-
ditions, and of the same variety under different con-
ditions, has not been investigated thoroughly, and
when we hear from one experimenter that a certain
variety contains 20 per cent, of starch, and from
another that it contains 15 per cent., and from still
another that it contains 30 per cent., that is only
what we should expect as long as our varieties are
not well defined and as long as very little is known
concerning the influence of variety, climate, fertil-
izer, cultivation, and storage upon the starch con-
tents in the sweet potatoes.

As to the percentage of starch in different varie-
ties grown under the same conditions, we find the
following data:

THE SWEET POTATO. 37

Wiesner (Rolistoffe I, under Batatas edulis)
states that the yellow varieties are richest in starch.

In Arkansas Report, 1890, p. 125, Professor Teller,
the station chemist, reports the analyses of ten
varieties of sweet potatoes. Below are given here
only the water-contents, starch, cane-sugar and
glucose :

Water. Starch. Cane Sugar. Glucose.

Per cent. Per cent. I'er cent. Per cent.

Shanghai or California Yam 65.35 20.84 C.94 1.41

Red Nanseniond 71.80 14.52 G.43 1.85

Red Bermuda G9.7G 1G.34 3.60 1.83

Southern Queen 70.00 17.98 6.13 1.15

Yellow Yam 71.13 16.64 5.19 1.25

Poplar Spanish 58.87 19.45 6.08 2.41

Early Nanseniond 75.66 9.79 5.41 1.63

Early Jersey 75.65 11.39 5.68 1.04

South Carolina Bulletins 28 (1897) and 63 (1901)
are devoted almost entirely to the sweet potato as a
starch producer. Mr. Shiver, assistant chemist of
the station, conducted thorough and elaborate experi-
ments, which cannot be given here in detail. The
experiments were made in order to determine to
what extent the starch contents change on storing,
and what effect the method of storing has on this
change. The paper is well worth reading. The
results which interest us are the following :

Analyses of Original Samples Which Had Stood
FOR Some Months After Harvesting.

Water. Starch.

1894 Per cent. Per cent.

Spanish' 55.9.3 29.58

Southern Queen 59.70 25.67

Yams 60.37 22.30

Poor Land 67.62 16.93

38

THE SWEET POTATO.

Water. Starch.

1895. Per cent. Per cent.

African Red 66.38 23.64

Poor Land 67.29 22.82

Southern Queen 66.19 21.74

Water. Starch. Cane Sugar. Glucose.

Dry Season. Per cent. Per cent. Per cent. Per cent.

Southern Queen 63.52 27.36 2.45 .51

Red Bermuda 63.04 28.00 1.81 .78

General Grant 64.36 26.12 2.75 .48

White Bermuda 67.55 23.74 2.27 .62

Georgia Yams 64.04 27.32 2.53 .47

Yellow Jersey 64.60 25.20 2.32 .71

Bunch Yam 63.81 26.42 2.80 .66

Season 1898.

Water. Starch.

Kame ©f Variety. Per cent. Per cent.

Georgia Buck 75.35 13.13

Bunch Yams 72.37 15.12

Bunch Yams from diflferent

sources 67.99 19.58

Horton Yam 70.29 15.06

Georgia Bucks from different

sources 71.56 14.35

Vineless Yams 70.03 16.85

Hanover Yams 76.16 13.61

Georgia Yam 70.01 18.87

Glucose. Sucrose.

Per cent. Per cent.

.77 4.31

1.09 4.45

.56
1.05

.73

.54

1.10

1.00

4.49
6.23

6.61
5.01

4.22
4.08

Season 1899. (C. C. McDonnell, Analyst.)

Water. Starch. Glucose. Sucrose.

Name of Variety. Per cent. Per cent. Per cent. Per cent.

Pumpkin Yam 68.94 17.38 1.08 5.17

Bunch Yam 72.04 13.72 1.38 5.47

Georgia Sugar Yam 67.81 18.41 1.08 5.08

Tennessee Yam 70.42 15.74 1.41 5.02

From Texas Bulletin 36 (1893) is taken the fol-
lowing table, which gives only the percentages of
water, invert sugar, and total sugar:

THE SWEET POTATO. 39

Invert Total

November, 1893. Water.^ ^uga.^ ^Sugar.^

Bunch Yam 70.83 2.14 3.74

Early Bunch Yam 73.26 2.66 4.60

Vineless 66.06 4.16 6.41

Nansemond 71.81 3.33 5.00

Red Nose 77.59 3.27 5.20

Brazilian Yam 67.23 2.52 5.26

Negro Choker 68.23 2.84 7.69

Tennessee 65.83 2.19 2.77

Southern Queen 61.58 5.10 9.20

Red Bermuda 75.81 2.77 5.26

Early Golden 74.70 3.00 6.75

Peabodv 79.04 3.35 6.41

Delaware 78.26 2.08 5.00

Barbadoes 75.44 2.92 6.98

Norton 66.69 4.67 11.90

Pumpkin 69.19 3.76 8.07

Although this table does not give the starch con-
tents, it is plain that in some varieties examined,
as for instance the Peabodv, where water and
sugar alone make 85.45 per cent., after allowance
is made for liber, proteids, and fats, there is not
much left for the starch contents ; while in the Ten-
nessee Yam, where water and sugar together make
only 68.60 per cent., there may be a very high per-
centage of starch.

All of these analyses were made with sweet pota-
toes at harvest time, and they are sufficient to show
that the varieties differ so markedly in starch con-
tents that it is of the utmost importance to choose
the variety carefully if sweet potatoes are to be
raised for starch production. It is obvious that a
reliable definition of the varieties is absolutely neces-
sary before one can begin such a systematic selection.

What effect climatic conditions have upon the
yield in starch has never been accurately deter-
mined.

40 THE SWEET POTATO.

Wiesner states that sweet potatoes may contain as
much as 10 per cent, of sugar, l)ut only about 9
per cent, of starch in the tropics, while in the sub-
tropics they may contain as much as 15 per cent, of
starch and only 3-4 per cent, sugar. From the
analyses quoted above we can see that the last part
of the statement is inaccurate. General statements
of persons holding the same or the opposite view
may be found now and then, but the writer is not
aware that accurate experiments have been made
an^^^iere to settle this question.

it is very likely that the constitution of the fer-
tilizer used should have an influence upon the starch
contents, and particular attention has been paid to
this question by the South Carolina station. It must
be borne in mind, however, that the experiment
extended only over one season, with only one variety
to test. So, while it may suggest certain lines of
experimentation, it can in no way be considered con-
clusive.

Below are parts of the table as given in the South
Carolina Bulletin, 63, 1901, giving the percentages of
water, starch and sugar in diiferent lots of Horton
Yam sweet potatoes grown with different fertilizers :

November 28, 1898 ; Variety, Horton Yam ; Wet Season.

Water. Starch. Glucose. Sucrose.

Fertilizer, Per cent. Per cent. Per cent. Per cent.

ri^';''p""'"'; I 03.81 22.86 .96 5.41

1,000 lbs. Compost j

'1?J'?; ''<^""'' f ^"'"'' ^ 63.77 22.21 1.20 6.10

1,000 lbs. Compost j

Kolhin.-- 62.07 24.58 1.19 5.28

JToo'ii ' r'"'*' f ''''"' \ 64.97 21.63 1.51 5.59

1,000 lbs Compost |

250 ll)s. Silicate of Potash ) , ^,

1 AAA n n f - 65.87 20.70 1.27 6.03

1,000 lbs. Compost j

1,000 lbs. Compost 65.26 20.80 1.41 6.21

THE SWEET POTATO. 41

As far as this experiiiieiit goes, it seems that tlie
given variations in fertiHzer produced a ditt'erenee
in starch percentage, although not nearly as great
as a selection of varieties fertilized in the same way.
If this sliould be confirmed and applied to a variety
naturally rich in starcli, that would, of course, be a
farther means of evolving a first-class starch-yield-
ing variety.

It is not at all improbable that the method of culti-
vation and the soil in which the sweet potato grows
maj' seriously influence the starcli percentage, but,
as far as known to the writer, no experiments have
been made to determine that.

That storing makes the sweet potato sweeter is
well known. By a series of analyses instituted by
the South Carolina and Texas stations, independ-
entlj', it has been determined that an actual increase
of sugar, mainly cane sugar, takes place during stor-
ing, and that this is accompanied by a decrease in
starch. It is strongly suggested by Mr. Shiver that
the starch which is lost is actually changed into cane
sugar. The general trend of the table seems to
favor this view. Still, there are in the analyses
quoted in Bulletin 63 some instances which are so
much out of accord with the averages given that they
strongly suggest that the samples may have been
much less uniform than was supposed. For exam-
ple: In Table XVIII we find the two samples of
Bunch Yam, from different sources, analyzed on
November 28, 1898, as follows :

Water. Starch. Glucose. Sucrose.

Per cent. I'er cent. Per cent. Per cent.

Bunch Yam (No. 1) 72.37 15.12 1.09 4.45

Bunch Yam (No. 2) 67.99 19.58 .56 4.49

42 THE SWEET POTATO.

Samples from the same lot, analyzed January 7,
1899, gave :

Water. Starch. Glucose. Sucrose.

Per cent. Per cent. Per cent. Per cent

Bunch Yam (No. 1) 67.31 13.66 2.02 9.90

Bunch Yam (No. 2) 67.29 13.83 2.40 9.43

Here, then, the contents in sucrose have risen over
100 per cent, in both cases, while the starch contents
have become nearly even, whereas they were very
different at the beginning. The second analysis, in
which all the results correspond closely, suggests at
least that the large dilTerence in the first analysis
may be due to individual variation among the tubers
of the same lot, and that the samples of either
analysis would not have given the other results at
the time of the other analysis. Moreover, there are
cases in the tables where the starch percentage act-
ually rose during a storage of six weeks, while at
the same time both sugar i^ercentages rose, while the
water percentage fell, as in the case of Georgia Buck,
which analvzed :

Water. Starch. Glucose. Sucrose.

Per cent. Per cent. Per cent. Per cent.

November 28 71.56 14.35 .73 6.61

Jamiaiy 7 67.63 14.43 2.12 7.85

The loss of water alone could not have caused a
rise of even 1 per cent, in the total of the other three
compounds, but we find a total increase of 2.61 per
cent. Similarly we find, in Table XXV, that the
starch contents of the Bunch Yam show two breaks
during storing, as follows : November 14, 1899, 13.92
per cent. ; December 14, 9.61 per cent. ; January 15,
12.30 per cent.; February 15, 8.18 per cent., and
March 15, 8.83 per cent. Twice during storing, then,
the starch contents rise.

THE SWEET POTATO. 43

In Tables XXX and XXXI we find another exam-
ple in the storing of Georgia Buck in the usual way
(straw, corn-stalks, etc.). They analyzed:

Water. Starch. Glucose. Sucrose.

Per cent. Per cent. Per cent. Per cent.

November 28 75.35 13.13 .77 4.31

January 7 69.52 13.51 1.74 7.66

March 1 75.80 9.88 3.21 3.77

These three, of course, do not seem to be in accord.
We can easily find quite a number of similar cases
in other tables of the same bulletin. Now, the
writer's purpose is not to criticise the experiments —
they are a step in the right direction — but it should
be emphasized that, with such individual differences
as these occurring in the analyses, we can hardly be
too careful in explaining the results of others which
agree more or less.

It seems clear, however, from the -tables given in
the bulletin, that sweet potatoes lose a large part
of their starch on storing, and that they gain in
sugar. Bulletin 36 of the Texas station gives these
extremes of variation in sugai- contents :

Invert Total
Variety. Time of Analysis. Sugar. Sugar.

Per cent. Per cent.

Tennessee November 1, 1893-94 2.19 2.77

Early Bunch Yam March 6,1894 7.25 19.71

The extremes in starch contents, as taken from the
South Carolina Bulletins 63 and 28, are :

Spanish, some months after harvest, 29.58 per cent.

Georgia Buck, January 7, stored in straw in a cov-
ered house, 7.20 per cent.

In South Carolina Bulletin 28, Mr. Shiver com-
pares the probable yield of starch per acre with that
from wheat and corn, in the following table :

44 IKK SWKE'l' iMjTATO.

Yield Starch

I'er Acre. Starch. I'er Acre.

Pounds. I'er cent. Pounds.

Wheat ] .200 57 684

Corn 1,900 65.5 1.283.8

Sweet Potatoes 12,000 22.0 2.040.00

This is based on a yield of 200 bushels per acre
and 22 per cent, of starch. Two hundred bushels
are even now. with talkie varieties, considered a mod-
erate yield. With a yield of 500-600 bushels with the
varieties which are reported from the South, and
28-29 per cent, of starch from selected varieties, the
yield of starch per acre would be simply enormous.
The Irish potato, which is raised for starch exten-
sively in Germany and France, yields only about
14-21 per cent, of starch, and 150-300 bushels of
tubers per acre.

We see, then, that it is well established that the
sweet potato contains enough starch to be considered
as an excellent source for its manufacture. We now
have to consider the other sides of the question:
What is the quality of the starch? and, Can it be
profitably extracted,?

Wiesner ("Rohstotfe") states that the sweet
potato starch commonly manufactured in Martin-
ique, Guadeloupe, Eeunion, Cochin China, India, etc.,
which has been exhibited at various world's fairs,
is not pure white, but grayish-yellow. By washing
this in pure water he obtained a much whiter prod-
uct, but not pure white. He thinks, however, that
a process might be found by which a jDure white
product could be obtained at once. He cites Gintl
(Appreturmittel, p. 4) for the statement that sweet
potato starch is even now imported to England in
limited quantities from the tropics, to be used for
technical purposes.

THr. SWEET POTATO. 45

Tscliircli and Oesterle (Anat. Atlas tier Pbarmac.
mid Nalirungsm., p. 229) treat sweet potato starch
under "Brazilian arrowroot," and state that it is
manufactured everywhere in the tropics, as, for
example, in India.

We have the statements, then, that sweet potato
starch is actually used for technical purposes now,
and that it is manufactured on a more or less
extended scale by crude methods. As far as known
to the writer, no thorough experiments have ever
been made to manufacture sweet potato starch by
modern processes on a scale sufficiently extensive to
warrant accurate conclusions. This question the
writer intends to take up in the future.

C. The Sweet Potato as a Source of Sugar.

It has often been suggested that the sweet potato
would be a valuable source of cane sugar, and the
sugar beet, in which the sugar contents have been
raised greatly by proper selection, is cited as an
illustration of what can be done. It is true that
some varieties of sweet potatoes, especially after
storing, are much higher in cane sugar percentage
than the sugar beet was at the beginning. The
Texas station Bulletin 36, for example, reports a
percentage of cane sugar as high as 12.46 in the
"Early Bunch "Yam." The great difficulty is that
the 12.46 per cent, of cane sugar are accompanied by
7.25 per cent, of glucose, which is noncrystallizable,
and which, besides, prevents the larger part of the
cane sugar from crystallizing. It may be well to
remember here that sorghum cane, which has a
more favorable cane sugar percentage (up to 14
per cent, of cane sugar with 3-4 per cent, of glucose)

46 THE SWEET POTATO.

than the sweet potato, has been tried and has proved
unprofitable. With the manufacture of sugar from
the sweet potato we may then as well wait until
the making of sorghum sugar is a profitable business.

Still sorghum is an important source for molasses
in the South, and there is no reason why the sweet
potato should not so be utilized. Other starch
crops, such as corn and Irish potatoes, always yield
a big product of glucose, manufactured from the
starch, which could not be separated, and the glucose
already contained in the plant. Here the sweet
potato is certainly superior to either, as the percent-
age of glucose is high and the additional amount
of cane sugar would be a very desirable factor in
manufacturing syrups.

The syrup should be an important product in a
sweet potato starch factory. No reliable data are
at hand to show whether the refuse, after starch and
sugar have been separated, would still be useful as
cattle feed.

On account of this high total percentage of starch,
glucose, and sucrose, the sweet potato should prove
a good source for the manufacture of alcohol.

III. THE STRUCTURE OF IPOMCEA
BATATAS.

In the following is given a detailed description of
macroscopical and microscopical characters of
Ipomoea Batatas. When the macroscopic charac-
ters varied with the varieties, that has been stated;
where differences were found in the microscopic
structure of different varieties, that was also
recorded, but the writer wishes it to be understood
that the material studied microscopically was not
abundant enough to justify him in saying that the
special characters found in particular varieties are
typical of them.

For the sake of convenience the plant is consid-
ered under the headings of Root, Stem, and Leaf.

(A) Root (Figs. 21-26) : Springing from two sides
of a stem at every joint, and from five longitudinal
rows on tubers, which are usually more or less dis-
arranged ; wliite, yellowish, pink or purple, long and
of uniform diameter for a shorter or longer distance,
until thickened into a tuber. Tuber fascicular to
spherical, smooth or veiny, terete or 5, 4, or 6-lobed
in cross-section, with roots coming off in 5, 4, or 6
longitudinal rows, or apparently irregularly; with
large conspicuous or small inconspicuous lenticels,
with dormant shoots usually near the lenticels and
root traces, but often apparently irregularly placed.
Shoots sprout earliest at proximal end of tuber,
breaking through skin, later in centre and distal end.
Color of tuber, white, yellowish-white, pinkish-white,

47

48 THE SWEET POTATO.

golden-yellow, bronze-colored, yellowisli-purple, light
or dark purple ; flesh, white, cream-colored, pinkish-
white, pinkish-orange, with or without purple marks ;
cambium and wood elements, white, dirty-white, yel-
low or orange, very distinct, or indistinct.

Latex abundant or scanty. Flesh soft, so as to
cut very easily, or hard, so as to break better than
cut.

Root (Microscopical) : In young root, xylem shows
five patches more or less separated by medullary
rays. In tuber, groups of xylem elements, or iso-
lated elements, each surrounded by actively dividing
area, scattered throughout the fundamental tissue.
Fundamental tissue and dividing tissue filled with
starch consisting almost entirely of compound
grains. Cambium with ring of xylem elements re-
mains near outside of tuber, with or without xylem
elements radiating towards center. Occasionally five
or more strands of xylem elements run along the
outer surface of the tuber, outside of the cambium,
forming the longitudinal ridges known as "veins."
These may or may not divide up and anastomose
profusely.

(B) Stem (Macroscopical) : Variable in length,
from about two feet in some bunch varieties (Fig.
39) to 20 feet or more in those with running vines.
Variable in thickness, depending on the variety,
from Vs iiich to more than i/4 inch in diameter at
its largest part.

Variable in habit, depending on the variety,
strongly twining or not at all, strongly running or
growing in a clump, prone to fasciation or not.
Variable in color; usually more or less purple col-
ored when emerging from the tuber; when grown,
light green throughout, or purple at the very base

TllK SWEET POTATO. 49

only, or purple below the attaclimeut of the petioles,
or purplish or dark purple all over, or brownish
green on exposed places, often darker on the upper
or lower side.

Lenticels abundant. Hairs usually present in
young shoots from tuber, persistent or not in the old
stem. At every joint roots develop on the two
opposite sides of the stem. These roots may mature
into tubers, depending on the length of season and
tiie variety.

Stem (Microscopical) : (a) Section through tip
(Fig. 13).

Epidermis beset with numerous gland cells, of
which some have much darker contents than the rest,
liairs more or less abundant, depending on the
variety. Epidermis block-shaped, thin- walled; one
or more layers of hypodermis well marked, depend-
ing on the variety; cortex loose, with intercellular
spaces, containing many well-marked latex canals
with usually five, occasionally four or six secreting
cells, largest latex canals toward the interior of the
cortex; endodermis sharply marked, occasionally
with a more or less marked accessory endodermis,
depending on the variety, the cells of which are
intermediate in length (on longitudinal section)
between the short endodermal cells and the long
cortex cells. Below endodermis an interrupted ring
of very thin-walled pericambium, one to three cells
thick; phloem very narrow, or of medium thick-
ness, depending on variety, xylem well developed,
internal phloem consisting of small isolated patches
of dividing cells among pith cells smaller than the
rest ; pith of large cells with intercellular spaces and
numerous latex canals as in cortex.

50 THE SWEET POTATO.

No crystals seen in any young tip, cut where the
folded leaves are about % inch in length.

(b) Section through base of old stem (Fig. 15).

Epidermis with rather few gland cells, with or
without hairs, with numerous lenticels. (The writer
has not been able to decide whether the lenticel-like
patches are made up of cortical tissue, or are pro-
liferations of the epidermis.)

Epidermal cells block-shaped, thin-walled, hypo-
dermis more or less well marked, depending on the
variety, with or without crystal cells, depending on
the varietv; three to five outer lavers of cortex form
a collench^Tua sheath, not very strongly thickened,
latex canals well preserved, or flattened out, or both
kinds in the same section, secreting cells well marked,
or hardly recognizable, depending on the variety.
With or without crystal cells, depending on the
variety.

Endodermis well marked, sometimes accompanied
by slightly-marked accessory endodermis.

Pericambium a ring of strongly thickened fibers,
or occasionally not much thickened, depending on
the variety. Phloem thick or thin, unevenly dis-
tributed in thickness, with or without crystal cells,
depending on the variety, occasionally with a few
latex canals, cambium strongly marked, xylem wide
or narrow, usually with, sometimes without, large
vessels toward external margin, depending on the
variety, these vessels often with an abundance of
tyloses. Xylem cells showing spiral and reticulated
thickenings of various patterns, and bordered pits.
Whenever large vessels are present, they are
arranged only on the two opposite sides of the stem,
one patch on either side, unequal in size. The

THE SWEET POTATO. 51

broader patch is sometimes divided in the middle.
Crystal cells occasionally occur in unliaiiified xvlem.
Xylem cells are not lignified uniformly (as shown by
satfranin stain).

Inside of protoxylem patches there are narrow
areas of tissue simiUir to fundamental tissue, with
the cells smaller than the cells of the pith. These
cells divide the protoxylem from the cambium of
the internal phloem. They are probably part of
the internal phloem.

Internal cambium well marked or not, rarely con-
tinuous around cross-section, depending on variety.
Internal phloem thick or thin, thickest below xylem
jDatches with large vessels, depending on the age of
the stem and the variety; with or without crystal
cells, occasionally enclosing latex canals.

Occasionally a few secondary xylem cells were
found regularly between the protoxylem patches and
the internal phloem, giving rise to bundles of reverse
orientation. Occasionally the internal phloem and
the joining pith cells near such reverse bundles con-
tain large fat globules not observed in any other part
of other varieties.

Fundamental tissue of large thin-walled cells with
large intercellular spaces and many latex canals,
which often show plainly the former plains of divi-
sion. Pith cells connected by pore-plates. Crystal
cells abundant or not, or absent. Pith cells of
normal size and shape, often divided into chambers,
each chamber containing a crystal. Occasionally
crystals cut off in the corner of a pith cell. Ring
fasciation common in some varieties, rare in others.

(C) Leaf (Macroscopical) : Leaf cordate, hastate,
slightly or deeply lobed, or cut (Figs. 74, 72, 70,
52, 31, .34). Apex of leaf and lobes obtuse to acute,

52 THE aWEET POTATO.

midrib prolonged beyond lamina into a short awn-
like tongue, base cordate or truncate, margin entire.
The cordate leaves are the first ones to appear in
most varieties on the young shoots from tubers as
well as on seedlings. They may be retained or be
succeeded by hastate, lobed, or cut leaves. Size of
leaf varying with age of shoot, variety, and food
supply. Normally developed leaves vary in differ-
ent varieties from 2i/o to 6 inches or more in their
widest diameter. Veins more or less palmately
arranged, projecting on both upper and lower. sur-
faces, especially on the lower. They are either paler
or darker green than the lamina, or faintly or deeply
purple on the lower surface or on both surfaces.

Hairs, consisting of several epidermal cells sup-
porting two or three short and one terminal long,
pointed cell with thickened walls (Figs. 12 and 13),
distributed on entire upper surface, with the excep-
tion of a space around the base, or only along veins,
or scattered in a few places around edge and apex,
or absent, depending on the variety; occasionally
occurring on lower surface of veins.

Color of leaf varies with the age, exposure to sun-
light, the variety, and for other reasons unknown,
from light green, through dark green, brownish-
green, light purple, to dark greenish-purple.

Young leaves usually shiny on upper surface, old
leaves less so or not at all. Young leaves folded
along midrib (Fig. 38). In some varieties the sur-
face of the lamina has a tendency to pucker towards
the upper surface (Norton, etc.).

Petiole long, varying with size of leaf and habit
of plant on normally grown leaves from 21/2 to about
9 inches in length, varying in thickness, depending
on variety (Figs. 27, 41). The base usually tliick-

THE SWEET POTATO. 53

gtuhI into a cushion on lower surface, where it bends
sharplv upwards. Cushion sometimes purplish on
green petiole, depending on variety. Base of petiole
twines around support more or less, depending on
variety. Petiole more or less grooved on upper sur-
face, rounded on the lower. Just below the attach-
ment of the lamina there are two small papillae, one
on the right and one on the left side of the petiole,
containing the petiolar nectaries.

Leaf (Microscopical). Epidermis: Consistmg ot
irregularlv shaped, more or less sinuate thin-walled
cells (Figs. 1 to 4), convex towards the outer surfaces
(Figs. 6 to 9), modified in shape when elongated over
Imndle traces (Figs. 1, 2), or radiating from gland
cells (Fig. 2), or surrounding base of hairs (Fig.
1), sometimes showing peculiar wall striations

(Figs. 3, 4).

Stomata very frequently on both surfaces (Figs.
1, 3, 4, 5, 6), most abundant on the lower, often
showing the successive divisions of surrounding cells
and guard cells (Fig. 4), sometimes showing double
guaiTl cell formation (Fig. 5); less abundant in
modified epidermal cells over bundle areas^(Fig. 1) ;
on level with surrounding cells (Figs. 6, 7).

Gland hairs scattered over both surfaces of leaf
(Figs. 1, 3, 6), chiefly the upper, consisting of a
plano-convex basal cell, on which rests a sohd
ellipsoidal mass of 6-8 elongated cells. The basal
cell rests on a somewhat broadened epidermal cell
which is usually sunk (Fig. 6).

Mesoplnfll of 'two or occasionally three layers ot
palisade parenchyma (Figs. 5 to 7), with stomatic
chambers (Figs. 6 and 7), in which may be very thm-
walled cells containing calcium-oxalate crystals.
Among both the palisade and loose parenchyma cells

54 THE SWEET POTATO.

there may be cells containing the same kinds of
crystals (Fig. 8), varying much in size. Loose
parenchyma o-i cells thick, with stomatic chambers
(Figs. 6, 7).

Veins (Fig. 9) protruding on l)oth surfaces; the
upper epidermis modified into narrow, papillate
cells, somewhat more thickened than ordinary,
underlaid by a patch of more or less thickened
collenchyma cells, which may be entirely or partially
sejDarated from the pith by the palisade cells, which
are then usually shorter and at least three deep.
Fundamental tissue with numerous latex canals, the
secreting cells of which are not well differentiated
from the other cells. Bundle crescent-shaped, con-
cave side up, the horns of the crescent of the xylem
connected by a string of separate small patches of
internal phloem, in which there may or may not be
a small patch of xylem, entirely distinct from the
main xylem mass. Fundamental tissue underlaid hj
a band of collenchyma cells, less strongly thickened
than those near the upper surface. Then follows a
single row of small, round hypodermal cells, and then
a modified epidermis, less markedly papillate than
that of the upper surface.

Very small veins enclosed in the loose parenchyma
and surrounded by large non-chlorophyllous cells
(Fig. 7), sometimes accompanied l)y one or two
rows of collenchyma and modified epidermal cells,
even if entirely cut off from epidermis by cliloro-
phyllous parenchyma.

Tip of midril) forming an awn-like tongue built of
non-chlorophyllous cells.

Petiole. Petiolar nectaries (Fig. 10) consist of
invaginations of the epidermis, to form cavities
which are thickly lined with glandular hairs similar

THE SWEET POTATO. 55

to those already described as occurring on the
lamina, but consisting of more cells. That insects
may be attracted by these gland areas appears prob-
able from the fact that one of the sections contains
an insect in the nectary. Fungal hypha^ are com-
monly found in these petiolar nectaries.

Petiole, sectioned one inch from the base (Figs. 11,
12) : Great difference in size, dei)ending chiefly on
variety; in those with larger diametei- all cells are
also proportionately larger, the epidermal cells being
increased least. Epidermis shows gland cells as
in leaf, may show hair as on leaf, usually larger, and
sometimes a peculiar proliferation, which gives the
impression of a lenticel. One to two layers of
modified chlorophyllous hypodermis are present,
and the entire fundamental tissue is enclosed in a
sheath of collenchyma from two to five or six cells
wide, varying in width with the different parts of
the same section, and in corresponding parts of the
sections from different varieties.

Latex canals frequent, irregTilarly distributed
throughout pith tissue, secreting cells well or fairly
well marked, depending on the variety.

Arrangement of bundles in petiole typically in five
parts, the three central ones enclosed by a common
endodermis and 23ericambium, and two upper ones
each with its own. Pericambium at times two layers
thick. Depending on the varieties, the arrangement
of the bundles may be in more than five patches, and
instead of three patches of endodermis, there may
be four. In all cases small patches of internal
phloem are found isolated from the xylem and from
each other, among small pith cells interior to the
xvlem.

In some varieties small groups of dividing cells

56 THE SWEET POTATO.

are found between the separate lower patches of
bundles, in the region of the cambium (Figs. 11, 12).
Crystal cells may be many, few, or absent. They
are formed chiefly in the fundamental tissue near
the bundles and in both external and internal phloem.

IV. CLASSIFICATION.

A. POPULAE VaEIETIES.

It is very commonly stated tliat in the North the
dry, mealy varieties are alone marketable, while in
the South the wet, sugary varieties are preferred.
In looking over the reports of our experiment sta-
tions and other publications the writer finds that
the following varieties are favorites of one section
or another:

According to Fitz (Sweet Potato Culture) the
"Hanover" or "Nansemond Improved" is popular
in the region around Richmond, the "Spanish" for
home use on the eastern shore of Maryland and Vir-
ginia, "Southern Queen" around Baltimore early in
fall and late in winter, the "Nansemond" in Vir-
ginia, New Jersey and Delaware, and the ""V^^ite
Yam" and "Yellow Spanish" in the South for home
consumption. "Yellow" and "Eed Nansemond,"
also "Earlv Jersev," are all old and standard sorts.
(By way of explanation it may be well to repeat that
the name "Yam" is applied to some varieties of
sweet potatoes.)

Mr. Starnes (Bailey's Enc. Ilort.) states that the
Northern markets prefer a dry, mealy potato repre-
sented by the "Jersey" or "Nansemond" strain;
the Southern market a rich, sugary potato, like the
"Georgia" or "Yellow Yam," which is generally
considered to be the standard of excellence. For
Northern shipment the "Jersey Sweet" is prefer-
able; for local sale the "Orleans Eed" (Nigger

57

58 THE SWEET POTATO.

Killer), "Early Golden" or "Bermuda Eed" head
the list.

The Louisiana Bulletins repeatedly refer to cer-
tain varieties as commonly cultivated or standards
of excellence. In Bulletin 13, p. 329, it is said that
the "Georgia Yam," also called "Common Louisi-
ana Yam, " is in general cultivation throughout that
State. "Southern Queen" ranks next to the
"Georgia" and "Sugar Yam" in popularity.

Bulletin 21, p. 648, confirms this, with the addi-
tional remark that the "Peahody," prohably identi-
cal with the "Nansemond," and certainly the same
as "Brazilian," is very popular in N. Louisiana for
hog feed.

In Bulletin 22, p. 709, it .is stated that the cut-
leaved varieties, ' ' Sugar, " " Georgia, " " Spanish, ' '
"Barbado" and "Vineless," are considered best for
table use.

Bulletin 30 of the same station says that the ' ' New
Jersey" and the "Yellow Nansemond" ("Missis-
sippi Yellow") are grown for Northern markets.

In 1898, Bulletin 52, p. 310, we find that the
"Georgia," "Sugar," "Pumpkin," "Hayman" and
"Vineless" are varieties ever\nvhere preferred in
the South for table use.

In 1894, the Georgia Bulletin 25, p. 155, states that
the "Georgia Yam" is the standard of quality.

The Iowa Bulletin 47, of 1900, confirms most of
these statements.

North Carolina Bulletin 112, p. 78, states the same.
Bulletin 132, p. 319, states that the popular potato
in that region (near Ealeigh) is the "Baydus" (cor-
ruption of "Barbadoes," of which there are a yellow
and a white-fleshed variety).

North Carolina Bulletin 74 contains the same

THE SWEET POTATO. 59

statement, with the remark that the bulk of the pota-
toes sold as "Bardos" belong hi reality to the
variety "Southern Queen.

> »

B. The System of CLAssiFiCATioisr.

At present one system of classifying the varieties
of sweet potatoes is in use among experimenters,
but none among growers in general. That used by
experimenters, the foliage system, was elaborated by
Mr. Price, of the Texas station, and reported in
Bulletins 28 and 36 of that station. It classifies the
sweet potatoes in three groups Ijy the typical shape
of their leaves, and then describes each variety
separately. But as no key is given further than that
referring to the foliage, it is of course not possible
to determine a special variety, if the name be
unknown or doubtful; for to comjiare it with all
varieties described under that group and to be still
left in doubt as to whether the variety was repre-
sented at all, is of course an uncertain and unsatis-
factory method.

Mr. Price deserves great credit for introducing
some order into the previous chaos, and the writer
acknowledges that the "foliage system" of Mr. Price
was one of the early stimuli which induced him to
work on the classification of varieties.

To give an idea of how necessary it is at present
to describe certain standards closely may be seen
from the following. Below are given extracts from
various bulletins published at our experiment sta-
tions, the descriptions of one of the best-known
varieties, "Southern Queen."

Mr. Price describes it as follows:

Round leaved, foliage pale green, sometimes
prominently notched on one side, vines very vig-

CO THE SWEET POTATO.

i:)roiis, root profiisel}', length eight feet, tubers
obtuse, medium size, white. Eeliable, much grown
in the South. Soft, damp, late.

Maryland Bulletin 33. — The Southern Queen is a
good yielder, a very handsome and salable potato.

Iowa Bulletin 47.— Southern Queen, medium run-
ner, leaves large, dense mat of foliage, tubers large,
fairly smooth, incline to run to roots. Skin greenish-
brown, rough ; flesh yellow, very wet, coarse ; not very
sweet nor pleasant. Very popular in the South.

North Carolina Bulletin 74. — Southern Queen,
very productive, good keeper, heading the list in
keeping qualities.

xVrkansas Report, 1890. — Southern Queen, heavy
growth of vines, tubers large and smooth. An early
variety.

Georgia Bulletin 25. — Southern Queen. Leaves
shouldered. Foliage deep green. Vines quite vig-
orous. Tuber quite large, both round and ovoid.
Skin white, flesh grayish-white or grayish-yellow;
quality very poor, stringy, coarse, fibrous and taste-
less. Bather early and productive.

Louisiana Bulletin 13. — Southern Queen. ]\Iost
popular in the South, excepting the Georgia and
Sugar Yams. Tubers round and mealy.) Vines
very strong. Good producer. Improve in flavor
by storage.

Louisiana Bulletin 21. — Southern Queen. Large,
round ; light yellow skin and meat, fair quality ; very
early and popular ; a good potato.

Louisiana Bulletin 30.— Southern Queen, white,
rather hard, dry. late. Strong grower. Vines large
and green. Leaves large, broad, rather bluntly
pointed, and have side points.

11 IK SWEET POTATO. 61

Flesh nearly white and rather dry. A very pop-
ular variety and good producer.

Louisiana Bulletin 3G, p. 1266. — Southern Queen is
recommended for both quality and quantity.

Louisiana Bulletin 52. — Southern Queen. One of
the most productive varieties (720 bushels per acre
in last experiment), but not an excellent table
variety. Suggests an enormous amount of hog and
stock food, and should be grown largely for this
purpose.

Fitz (Sweet Potato Culture, p. 9). — Southern
Queen is the earliest of all sweet potatoes. In eating
condition (near Baltimore) by the middle of July
when first dug. too wet in fall and early winter.
Root very large, light color, good keeper, vine vig-
orous, leaves large. Good quality.

So we find that the Southern Queen, one of the
best-known sweet i:>otatoes in the United States, has
been variously characterized as round-leaved and
shouldered, foliage pale green and dark green, vines
very vigorous and medium; tubers obtuse, round and
inclined to run to roots, medium sized, and large,
smooth, and rough ; skin white, greenish-brown, and
light yellow; flesh yellow, grayish-white, grayish-
yellow, light yellow, soft, and hard ; damp, very wet,
mealy, and dry; quality poor, fair, recommendable,
a good potato, good liQg and stock food, flavor coarse,
fibrous, tasteless, yet a very popular table variety in
the South; late, rather early, very early and the
earliest of all sweet potatoes.

The writer has purposely considered chiefly the
work of the agricultural stations, in order that there
could not be the objection that the men giving the
descriptions were unfamiliar or unskilled as regards
such work. These men have been selected as efficient

62 THE SWEET POTATO.

men for that kind of work, and there is no doubt but
that every one of these descriptions fits the sweet
potato known by the particular worker as ' ' Southern
Queen. ' '

The question might well be asked : How can such
statements be reconciled f Much has been said about
the variability of varieties, yet it seems hardly pos-
sible that such changes affecting everything that is
regarded as essential in a sweet potato should occur
in the same variety. That sweet potatoes vary con-
siderably is certainly true. This is just as true,
however, with other plants long in cultivation, as,
for instance, corn, the common bean, the banana,
cabbage, etc. Still in all these we can distinguish
certain types, which may be variable in their repre-
sentatives, but which differ enough from each other
to be easily distinguished. So we have dent corn,
flint corn, sweet corn, pop corn; the drumhead and
Wakefield cabbage, cauliflower, Brussels sprouts,
kale, etc., and similarly different types of the bean
and banana. The sweet potato types certainly are
not nearly as distinct as those of cabbage or corn,
but they are perhaps as well marked as those of the
bean.

Now, what has been done to distinguish these

tvpes ?

' In Farmers' Bulletin 129, 1902, Mr. Nesbit says,
concerning classification :

''Classifications of varieties based on different
principles have been attempted without, as yet, res-
cuing the subject from disorder. The most elabor-
ate system, and perhaps the only one worthy of a
name, is that adopted by R. A. Price, horticulturist
of the Texas Experiment Station in 1893, which he
calls the 'Foliage System.' He divides sweet pota-

THE SWEET POTATO. 63

toes into three groups, having 'round or entire foli-
age,' 'shoulder foliage' and 'split or lobed foliage.'
He says, 'If this foliage system is taken in connection
with a short description of the color of the tubers
and of the vines, there is scarcely a variety which
cannot be distinguished from all other varieties.'

' 'This system has been applied at several stations,
yet it would be quite impossible to recognize some
varieties as known in one section by descriptions
given of them according to the foliage system in
another section if the name were omitted. So
strong are the influences tending toward diversity
that the writer is convinced that no system of classi-
fication can demonstrate much value until the sup-
posed varieties are all brought together and propa-
gated imder uniform conditions for several years."

Before going further into the subject, one might
well ask if there is any urgent necessity for the
classification of the varieties. The writer's reason
will be given later. In the following are quoted
the opinions of others who are practical farmers or
are otherwise interested in the subject:

In Farmers' Bulletin 129, p. 38, Mr. Nesbit, a
practical sweet potato grower of Maryland, gives
some suggestions for future experiments. In regard
to improvement of varieties he has this to say: "If
all the varieties or supposed varieties for which
merit is claimed should be collected and cultivated
for several years under favorable conditions and
with a system calculated to develop excellence,
planters might, at the conclusion of such a course,
be able to select from a few varieties of marked
characteristics such as give promise of special use-
fulness to them. The value of such work in estab-
lishing varieties and determining their relative

64 THE SWEET POTATO.

worth by comparison and in opening the way for
an orderly nomenclature can not be doubted."

In North Carolina Bulletin 74, p. 3, Mr. Massey
says : ' ' There is much confusion in the nomenclature
of* this vegetable. 'Peabody' and 'Early Eed' are
so near alike that they may be regarded as synony-
mous; 'Southern Queen,' 'Hayman,' 'Bahamas' and
'Yams' are the same. 'Norton Yams' and 'Buck-
skins' are also identical. 'African Reds,' 'Black
Spanish' and 'Nigger Killer' are also synonymous,

In the Arkansas Report for 1890, p. 123, Mr. Ben-
nett says : ' ' Much confusion arises with the different
varieties of sweet potatoes by the many different
names by which the separate varieties are known.
They invariably have local names in the particular
locality in which they are grown, and, as might be
expected, there is great similarity between some of
the so-called varieties."

In Georgia Bulletin 25, p. 153, Mr. Starnes: "In
grouping the different varieties of sweet potato we
have followed at the Georgia station the general
custom of arrangement with reference to the leaf,
in default cff a better system. Indeed it can scarcely
be called a system at all, for the reason that the
same vine will sometimes hold half a dozen different
shapes of leaf; and while a distinction appears to
exist between the 'split leaf varieties and all others,
it is by no means easy to determine with some
varieties whether the 'round' or 'shouldered' form
of leaf prevails. Yet when we endeavor to classify
by other forms of resemblance, as shape, size, color
or quality of tuber, we are met by even worse incon-
gruities, and are forced to fall back on the 'leaf
classification,' clumsy and unsatisfactory as it is."

THE SWEET POTATO. 65

Attempts have been made to erect, in addition to
the three forms of leaf generally accepted, to wit:
"Split leaf," "Shouldered" and "Round," a fourth
form, "Semi-shouldered"; but the difficulties are
too great in the way of its adoption, and hence the
regular division into "Split-leafed," "Shouldered"
and "Round" must suffice.

In Louisiana Bulletin 30, p. 1053, Mr. lUiruette
says: "^iucli time and expense have been speut in
trying to properly classify these so-called varieties
and adopt a nomenclature which can be followed
throughout the country, but so far, only with partial
success."

These questions give a fair idea of what others
think. Another good indication of the desire for
a uniform nomenclature are the frequent attempts
to reduce a number of varieties to a group, and the
abundant descriptions of many varieties, given in
various bulletins.

The writer's own reasons for working on a classi-
fication of varieties are as follows : The sweet potato
at present forms one of the staple crops in various
sections of our Atlantic and Gulf States. Experi-
ments liave demonstrated that there is an enormous
difference between the yields of different varieties.
It is well known that there is a great difference in
the quality and market value of different varieties.
Some varieties contain a starch content of as much
as 29 per cent., and could probably be utilized for
starch. But one of the most serious drawbacks to
experiments in any of these lines is the confusion
in nomenclature, "\\1ien one experimenter has
detected a desirable quality, another who tries to
verify the result secures another variety under the
same name and fails ; others having sweet potatoes

66 THE SWEET POTATO.

^yitli the same name growing in their own patches,
observe and condemn, and so experimentation gets
into discredit. As the writer himself intends to
work further on the sweet potato in the future, he
feels the necessity of straightening out the nomen-
clature first of all.

Before advancing any opinion on the character to
be used in the classification, the following deserves
to be carefully considered:

All varieties of the sweet potato are supposed
to have originated from the same plant; they have
varied enormously, and so they are apt to vary more.
Therefore, a variety which fits a certain description
at present may in the course of a few years produce
certain sports which cannot be referred to the
original type. This should be expected and would
not invalidate the system of classification. These
sports belong to a new category and, unless they
have developed into a variety already known, they
form a new one. This principle seems self-evident,
but objections have been made to certain descrip-
tions, because plants in the same patch, the produce
of the same ancestors, did not agree with each other.

Then, also, no system of classification can provide
names for varieties which are in accord with all
names now in use. The example of ''Southern
Queen" and the quotations from the various bulle-
tins already cited amply demonstrate this.

In naming the varieties priority of nomenclature
should be considered. This the writer would gladly
do if the labor involved were not altogether out of
proportion to the results. In 1731 Catesby tried to
reduce the varieties under cultivation in the Caro-
linas to five types. He states, as quoted, that there
were at that time a number of conflicting names for

THE SWEET POTATO. 67

the varieties. lu order to consider priority prop-
erly one would have to go back then to Catesby at
least. The writer has received over seventy so-called
varieties under various names, and finds himself
utterly unable to determine the majority of these
from the scanty description given in previous litera-
ture. Those varieties have been given the names
under which they were secured, unless those names
gave rise to confusion.

It must also be taken into account that certain
varieties will in all probability prdduce normal
tubers only in certain sections, and that some varie-
ties when transplanted from one climate to another
will exhibit different characters. We find the same
to be true of other plants. Sweet corn, growing
six feet high in the Northern States and having the
ear about 21/2 to 3 feet from the ground, will scarcely
grow 4 feet tall when planted in western Texas, and
make the ears right on the ground, in the first season.
Such may be due to one of two things : Either the
variety will not grow under the new conditions, or
it needs to be acclimated. It is, of course, probable
for that very reason, that some varieties with which
the writer has experimented should not do well in
New Jersey. Therefore, unless good typical tubers
were received from which to give the characters, the
tubers of varieties which did not seem to thrive have
not been included in the key.

Primarily, the varieties have been classified for
i^^he writer's own convenience, and the intention is
to continue the work on sweet potatoes in various
directions during spare hours, as opportunity pre-
sents.

In any classification of varieties the following are
essentials ;

\

68 THE SWEET POTATO.

The distingTiisliing characters must be reasonably
permanent, i. e., the great majority of the descend-
ants of the same plant grown together nnder normal
conditions ought to maintain uniformity in those
characters.

The distinguishing characters ought to be so clear
that they can easily be determined.

It is very desirable in a key that it should take
up the characters in such an order that varieties
which differ but little from each other and are very
likely closely related can be easily compared. This
is perhaps best effected by using the same characters
in classifying all varieties.

As it is not always possible to determine all the
characteristics of a varietv at one time, the char-
acters given should be as abundant as possible, so
that the variety may be determined even if not all
parts of the plant are present.

The key, as evolved below, was elaborated to meet
these different requirements.

The distinguishing characters, as utilized in the
key, were determined in the following manner :

Thirty-five so-called varieties were obtained from
the Agricultural Experiment Stations at Washing-
ton; Baton Rouge, La,; Experiment, Ga.; and from
Mr. Rose, of Wilmington, Del., and Mr. Trouncer,
of AVenonah, N. J. All these were planted in the
same patch at W^enonah and studied throughout the
season.

All types of leaves produced by all varieties, as
the season advanced, were collected and pressed.
Photographs were taken of typical hills and vines
of each variety to show characteristic growth and
leaf arrangement. Careful notes were taken repeat-
edly and independently of the comparative length

THE SWEET POTATO. 69

aud thickness of the vines, their exact color near the
tip, the center, the base, on top, and below; the size
of the leaves, the color of young and old leaves, the
length of the petiole, the amount of hairiness on
leaves and stem, the coloration of the veins on both
upper and lower surface, the amount of latex, etc.
After the harvest the tubers were studied in regard
to size, shape, exterior color, distrilmtion and size
of lenticels, roots, and shoots, hardness when raw,
color of flesh, wood elements, and caml)ium, and
relative abundance of latex. Samples were boiled
and baked of all varieties of which there was mate-
rial to spare, and observations made on the color,
odor, softness, flavor, sweetness, and stringiness of
the cooked tubers.

The great difference in the outer appearance ot
foliage, stems, and tubers suggested corresponding
ditferences in their internal structure, and to deter-
mine the extent of such differences, slides were
prepared from fifteen varieties which could easily be
disting-uished, showing the following parts: Upper
and lower epidermis of leaf, cross-sections through
the leaf, petiolar nectaries, petiole, the tip and base
of a full-grown stem, and the upper and lower part
of a tuber, and longitudinal sections through the

stem.

During the winter the writer received an abun-
dance of material from the already named stations
and also from Jamaica, India. Barbadoes, the Sand-
wich Islands, Mauritius, and Colombia. As many of
the tubers were in a rather poor condition on arrival,
the only way to save them was to plant them at
once. Accordingly all the tubers were started in the
greenhouses of the Botanical Garden of the Uni-
versity of Pennsylvania, and the entire lot planted

70 THE SWEET POTATO.

out in May. The varieties, received under more
than a hundred different names, were studied in the
same systematic fashion as the thirty-five varieties
of the previous year.

As was expected, the majority of the characters
noted proved unreliable. The abundance of the
latex and the color of the leaves, for example, while
distinctly variable with the varieties, were not con-
stant enough to be used as a characteristic. By
studying all the plants of every variety planted
(about 15-30 plants on an average), the writer came
to the conclusion that certain characters were suffi-
cient to distinguish all varieties which could be dis-
tinguished at all by the closest macroscopical obser-
vation. Microscopical characters were not utilized
because they could not be easily applied, and because
the writer is not prepared to say that the micro-
scopical differences found are constant.

The character by which the groups can be sep-
arated most readily is, doubtless, the shape of the
leaf. Mr. Price recognized this in his system of
classification, and the author has adopted his terms,
"round-leaved" and '* cut-leaved. " The shape of
the leaves enables one to divide all varieties at hand
into five main groups.

The first group comprises all varieties with "cut"
leaves. As "cut" leaves are counted all those in
which the projecting lobes constitute almost the
entire leaf surface, so that little remains when all
the lobes are cut off at their bases by cuts perpen-
dicular to their median lines. In all but one or two
varieties it is very easy to tell whether the leaf
is cut or not; but even if that should be impossible,
that fact does not interfere much with the determina-
tion of a variety, as will be seen later.

THE SWEET POTATO. 71

The second group is formed by those varieties
which have only '* round" leaves. By a *' round"
leaf is meant a heart-shaped entire leaf.

In the third group have been placed the *4ong-
leaved" varieties. In all well-formed, full-grown
leaves of these varieties the length of the large cen-
tral lobe, as measured along the median line from
the base of the lobe to its tip, exceeds the greatest
breadth of the same lobe.

The "broad-leaved" varieties constitute the fourth
group. In these the greatest breadth, i. e., the base
of the central lobe, exceeds its length from the base
to the tip.

Some varieties invariably have several types of
leaves, so that it is hard to tell which shape is the
most common. All varieties which normally have
well-developed, full-grown, ''round" leaves, as well
as ''long" or "broad," have been classed as "mixed-
leaved" in group five.

In the key there has also been provision made for
varieties in which both "long" leaves and "broad"
leaves are frequent.

It should be kept in mind that all references to
leaves, unless otherwise stated, apply only to full-
sized leaves developed at least two months after
planting, for the first leaves developing from a tuber
are usually "round" leaves, even in varieties which
later never produce them. The writer has even
observed them occasionally on plants with normal
"cut" leaves.

Almost as characteristic as the shape of the leaf
is the size of the leaf. Weather and soil conditions
exert a certain influence on the size of plant parts ;
still, the author has convinced himself that neither
the difference between the compost soil used in

72 THE SWEET POTATO.

greenhouses and the soil of a sandy New Jersey-
field, nor the difference between the comparatively
dry growing season of the summer of 1905 and the
extremely wet and hot season of 1906, affected the
value-of the size of the leaf as a characteristic in the
least.

All varieties begin the season's growth with small-
sized leaves. The largest leaves do not appear until
after two or three months of growth, and therefore
all references to size apply only to those later large
leaves.

The ''small-leaved" varieties never have normal
leaves measuring over four inches in width from
tip to tip. In fact, most of their leaves are less than
three inches wide. These "small-leaved" varieties
are, however, most given to fasciation, and fasciated
branches may have larger leaves. These have not
been considered in the key.

The "large-leaved" varieties have many of the
later large leaves measuring over four inches at
their widest part. Some varieties commonly grow
leaves six to seven inches across.

A character as definite and as easy to determine
as the shape and size of the leaves is the length of
the stems.

The bunch varieties and other "short-stemmed"
varieties never have stems measuring as much as
four feet long, even in a wet season.

The color of the stems is another very helpful
character which is easily determined.

Some varieties have entirely green stems with at
the most a few blotches of brownish sunburn. These
are classed together as "green-stemmed." Others
liave green stems which bear small purple marks
around the axils of the leaves, i. e., around the base

THE SWEE:T potato. 73

of the leaf-stalks, with perhaps occasional other
spots of purple. Still others have stems which are
for the most part purple, but shade into a dull
greenish-brown, and from that into green at the base
and tip and in various otlier places, so that while a
large portion of the stem is actually purple, the dull
greenish-brown color is also quite evident. Thin or
young stems are often entirely green. These stems
have been called ' ' greenish-brown to purple. ' ' Then
there are varieties in which the stems are unmis-
takably purple. It may be that even in these a few
inches at the base and tip are green, but there are
no entirely green young stems, and there is no green-
ish-brown color in evidence. These varieties are
termed "purple-stemmed."

Almost as constant as the width of the leaves is
the thickness of the stems. In the '"thin-stemmed"
varieties the stems never measure more than one-
eighth inch in diameter; the diameter is determined
by lajdng the stem across a rule and looking with
one eje. The "thick-stemmed" varieties measure
at least three-sixteenths of an inch at the thickest
part of their full-grown stems, and some of them
even measure more than one-quarter inch.

Another character, which can be seen at a glance,
is the purple star-shaped spot which some varieties
have on the upper surface of the leaves at the point
where the basal primary veins spring from the
petiole. This purple spot has been called the
"star." It may be present or absent.

The lower surface of the midrib and the other
large veins affords another easy means of distin-
guishing between varieties. In some varieties the
lower surface of the primary veins is more or less
deeply colored purple. In others some of the older

74 THE SWEET POTATO.

leaves, usually those wliicli happen to be supported
by darker petioles than the rest, have a faint streak
of pink running part-way up the midrib only.
Other varieties have a small purplish area on the
lower surface of the base of the midrib, i. e., just
before the petiole divides into the primary veins.
Still others have the lower surface of the veins
green.

The leaves offer another distinguishing feature in
the arrangement of the hairs on the upper surface.
Some varieties have the leaves entirely covered with
hairs. Others have the hairs only along the veins
and perhaps over a more or less extended area on
the tip of the central lobe. Still others have no
hairs on the upper surface of the leaf.

Although the root, or "tuber," is the most impor-
tant part of the plant, it offers few characters of
stability which might be used in a practical key.

The outside color of the tuber, however, is as
easily determined as it is important. Some varie-
ties have grayish-white or popularly called "white"
tubers. Others have yellowish, golden, or bronze-
colored tubers. Still others have them colored a
yellowish-red, or a pinkish-yellow, and some have
dark red or purple ones.

Although not quite so constant as might be desired,
the coloration of the cut surface of a freshly cut
normal-sized tuber is characteristic enough to be
used as a distingniishing character. On cutting some
varieties one finds that the flesh is colored pure
white. In others the flesh is cream-colored or yel-
lowish-white. Still others show a pinkish hue dif-
fused through a white or yellowish mass color. A
few have the flesh colored a deep pinkish-orange,

THE SWEET POTATO. 75

while some have part or the entire heart of the tuber
stained with pnrple.

About as definite a character as the color of the
flesh is the distinctness of the bundles which are
scattered irregularly throughout the mass of the
starch-bearing tissue of the tuber. In a few varie-
ties a freshly cut tuber shows at a glance the loca-
tion of these bundles. Small, sharply-marked spots
dot the surface, standing out "distinctly." In most
tubers a freshly cut surface shows to a certain extent
the location of the bundles, but the bundles appear
only as slightly darker areas or watery specks not
sharply cut off by a line from the surrounding tissue.
The bundles appear "blurred." Still other tubers
show no indication of the position of the bundles on
a freshly cut surface. In these the bundles are "not
visible."

It is certainly desirable that other important
characters, such as dryness or dampness of the
flesh of the cooked tubers, flavor, earliness, etc.,
should be considered in a comparative classification.
So far the writer has not succeeded, however, in
establishing standards of flavor, and he was not in a
position to make tests of the dr^mess and earliness
of the tubers. The description of the varieties, as
given later on, will therefore only serve to distin-
guish them, and not as an index to their market value.

C. Key to Vameties.
With all the above facts in his possession, the
writer began to evolve a key. The style of key that
naturally suggested itself was the one commonly
used in botanical text-books for the determination
of species. After careful thought the writer
decided to use the key given below, although it

76 THE SWEET POTATO.

requires somewhat more work for the determination
of some varieties. The reasons will be given later.

A. — Shape of Leaf.

1. Cut.

2. Eound.

3. Long.

4. Broad.

5. Mixed (round and lobed).

B. — Size of Leaf.

1. Small (less than four inches across).

2. Large (more than four inches across).

C. — Length of Stem.

1. Long (more than four feet long).

2. Short (less than four feet long).

D. — Color of Stem.

1. Green (with or without brownish areas).

2. Green, with purple around the axils of the
leaves.

3. Greenish-brown to purple.

4. Purple.

E. — Size of Stem.

1. Thin (less than ^/,s inch in diameter).

2. Thick (more than i/g inch in diameter, often

t\ or more).

F. — Presence of Star.

1. Star present.

2. Star absent.

G. — Color of Lower Surface of Veins.

1. Veins purple.

2. Midrib pinkish in some old leaves.

THE SWEET POTATO. 77

3. Purple spot at the base of the midrib.

4. Veins all green.

H. — Arrangement of Hair on Upper Surface of Leaf.

1. Hair all over.

2. Chiefly on tip and along veins.

3. Absent.

I. — Outside Color of Tubers.

1. White.

2. Yellow, golden or bronze.

3. Yellow-red or pinkish.

4. Red or i^urple.

J. — Color of the Flesh of the Tubers.

1. White.

2. Cream-colored or yellowish-white.

3. Pinkish-white or pinkish-yellow.

4. Pink-orange.

5. Marked with pnrx^le.

K. — Distinctness of Wood Elements in Tuber.

1. Distinct.

2. Blurred.

3. Not visible.

In the determination of most varieties known to
the author there is no difficulty in referring all their
characteristics to the different alternative numbers
expressing them. In some varieties, however, it is
hard to tell whether the leaves reach a greater width
than four inches, as there are such few leaves that
do, and those might be exceptions. A similar diffi-
culty will sometimes present itself in other places.
All that is necessary is to express by the figures the
actual condition. Thus : A. 3-4 means that both long

78 THE SWEET POTATO.

and broad leaves are frequent on the plant; B. 1-2,
that most large leaves are below four inches in
width, but that quite a number measure more than
four inches between their extreme tips from side to
side. Similarly, E. 1-2 means that the stems occa-
sionally attain a diameter of A of an inch, but that
most of the stems are less than A of an inch thick,
and a few even as little as or less than l^ inch.
These peculiarities of some varieties will rather help
than hinder in the work of determining them.

Hoiu to Use the Key.

The plants should be examined in the field, under
normal conditions. The tubers may as well be
examined after they are mature, as they show the
details better than immature ones.

Given a plant which was received under some
unfamiliar name, and it is to be determined whether
it is an old variety appearing under a new name.
The leaves are cut. That gives us A. — 1. Most of
the later, full-grown leaves measure more than four
inches across at their widest points, i. e., from tip
to tip of the most spreading lobes. That means
B. — 2. The full-grown stems are shorter than four
feet. So we put down C. — 2. As the stems are
green with purple marks around the axils of the
leaves, we have D. — 2. The full-grown stems meas-
ure y\ of an inch at their thickest point, which is
expressed by E. — 2. The purple star is clear and
the lower surface of the vein is purple, which gives
us F. — 1 and G. — 1. As a few of the younger leaves
show a few scattered hairs on the midrib, and some
of the older leaves have no hair at all, we express
that by H. — 2-3. The tubers we find to be white
outside and inside, which means I. — 1 and J. — 1.

THE SWEET POTATO. 79

The bundles, or wood elements, are not visible on the
freshly cut surface. That gives us K.— 3. Thus our
formula runs : Al— B2— C2— D2— E2— Fl— Gl—
H2-3— II— Jl— K3. AVhen we refer to the alpha-
betical arrangement of the formulas we find that it
is the formula of Ticotea. To make entirely sure,
we compare our plant with the photograph of Ticotea
and with its description. If all these points agree,
we may be certain that the variety is Ticotea.

Advantages of This Key Over the Ordinary
Botanical Key. ♦

At the beginning of the chapter the writer prom-
ised to give his reasons for choosing this key. This
key is preferred on account of the following advan-
tages :

1. It is a key and a classification combined. All
varieties are determined by the same eharacter^^ and
can be readily compared, while in the ordinary key
each variety is thrown out at an opportune moment
by a character which may be common to many others
already separated by some other character. One
comes to associate that particular character with
that variety, while in reality it depends entirely on
the arbitrary arrangement of the key whether the
character is used at all. Similarly there is little
attempt made in the arrangement of the ordinary
key to keep together the varieties most resembling
each other, while in this key they stay together nat-
urally. For example, Bronze Spanish, as the name
indicates, has a peculiarly colored tuber. In the
ordinary key this character would very likely be used
at an early stage to separate it. Black Spanish has
a tuber which is similar to several others, and it
would be hard to separate it by that character. Yet

80 THE SWEET POTATO.

the two cannot be told apart in the field, unless one
dig for the tubers.

2. It can be used with incomplete specimens. If
a certain character cannot be determined from the
material on hand, the space reserved for it may be
left blank, and the next character taken up. One is
not continually before alternatives which may be at
the time unanswerable. It is likely that the deter-
mination is possible even without that character.

3. It is more convenient to the non-scientist. In
the determination of all varieties the same process
is gone through, and once that process is learned by
heart it need not be changed again for the deter-
mination of the next variety.

4. It is flexible. The writer is certain that he
has not studied all varieties existing. New varie-
ties can be easily catalogued and inserted in the list
by anyone. In the ordinary key the advent of a new
variety necessarily causes confusion, as it not only
could be wrongly determined itself, but would also
interfere with the determination of other varieties
which might agree with it in the critical characters.
A considerable number of new varieties would make
an ordinary key useless, while this key provides for
all. Should other characters be found which would
aid in the determination, they could simply be
entered under subsequent letters, L — , M — , N — ,
etc., without interfering in the least with the work-
ing of the key.

D. Alphabetical List of Formulas.

Al B2 Cl Dl El F2 G4 HI 13 J3 Kl = Belmont.
Al B2 Cl Dl El-2 F2 Gl -f- 3 H3 12 J2 K3 — Kala.
Al B2 Cl Dl El-2 F2 G4 HI 12-3 J3 Kl = Georgia,
Al B2 Cl D2 E2 Fl Gl H2 II .Jl K3 = White Gilk.
Al B2 Cl D2 E2 Fl Gl H2 12 .J2 K3 = John Bvirnet.

THE SWEET POTATO. 81

Al B2 CI D2 E2 F2 G3 H3 13 Jl K3 = Kawelo.

Al B2 CI D4 El-2 Fl Gl H2-3 13 J3 K3 = Huaiiioa.

Al B2 CI D4 E2 Fl Gl H3 14 Jl K3 = Ihumai.

Al B2 C2 Dl E2 r2 G3 H2 14 J3 K2 = Vincentoniaii.

Al B2 C2 D2 E2 Fl Gl H2-3 11 Jl K3 = Ticotea.

*A1 B2 C Dl E F2 G4 H2 13 Jl K3 = Gros Gandia.

A2 Bl C2 Dl El-2 Fl Gl HI 11 J3 K3 = Minnet.

A2 i31-2 C2 Dl El-2 F2 G4 Hl-2 11 Jl K3 = Vinele«s Beech.

A2 B2 CI Dl El-2 F2 G4 Hl-2 II Jl K2-3 = Caroline Lee.

A2 B2 CI Dl E2 F2 G4 H2 11 J2 K3 = White Sealy.

A2 B2 CI D2 E2 Fl Gl HI 14 J2 K3 = Trinidadiau No. 1.

A2 B2 C2 D2 E2 Fl Gl H2 11-2 Jl Kl = Shanghai.

A2 B2 C2 D3 E2 Fl Gl HI 14 J2 K3 = Brass Cannon.

♦A2 B C D4 E2 Fl Gl Hl-2 14 Jl K3 = MuflFard.

A3 Bl CI D4 El Fl Gl HI 12 J2 K3 = Laiakona.

A3 Bl CI D4 El Fl Gl H2 14 Jo K3 = Kapo.

A3 Bl-2 C2 D4 E2 Fl Gl H2 11 Jl K2 = Thompson's Favorite.

*A3 B2 CI Dl El Fl Gl H2 14 Jl K = India Red.

A3 B2 CI D2 El-2 Fl Gl H2 14 J2 K2 = Governor.

A3 B2 CI D2 E2 Fl Gl H3 II Jl Kl-2 = Roosevelt.

A3 B2 CI D2 E2 Fl Gl H3 12 J2 K3 = Pilipili.

A3 B2 CI D2 E2 Fl Gl H3 14 Jl K3 = Kauaheahe.

A3 B2 CI D4 E2 Fl Gl H2 13-4 J2 K3 = Yellow Red.

A3 B2 CI D4 E2 Fl Gl H2 14 Jo K3 = Black Spanish.

A3 B2 CI D4 E2 Fl Gl H3 12 Jl K3 = Bronze Spanish.

A3 B2 CI D4 E2 Fl Gl H3 12 J3 K3 = Pikonui.

A3 B2 CI D4 E2 Fl Gl H3 13 J2 K2 = Key West.

*A3 B C Dl E2 Fl Gl H2 12 J2 K3 = Sulla.

*A3 B C Dl E2 F2 G4 H2 13 J2 K2 = Trinidadian Xo. 2.

*A3 B C D2 E Fl Gl H3 14 Jl K3 = Chazal.

A4 Bl CI D3 El-2 Fl G4 H2 12 J2-3 Kl = Yellow Colombia.

A4 Bl CI D3 El-2 Fl G4 H2-3 II J2 Kl = White Colombia.

A4 Bl CI D3 El-2 F2 G2 H2 11-2 J2 K2 = Southern Queen, weak

type.
A4 Bl-2 C2 D2 E2 Fl Gl HI 13 Jl K2 = Red Sealy.
A4 B2 CI D3 E2 r2 G2 H2 11-2 J2 K2 = Southern Queen, strong

type.
*A4 B2 C2 D2 E2 Fl Gl HI 12 J K = Florida.
A4 B2 C2 D4 E2 Fl Gl H3 14 Jl K3 = Halonaipu,
A4 B2 C2 D4 E2 F2 G2 + 3 H3 12 J3 K3 = Pu.
*A4 B CI D2 El Fl Gl H3 14 J5 K3 = Violette Rouge.

82 THE SWEET POTATO.

A4 B C D2 E Fl Gl H3 13 J5 K3 = Violette Blanche.

A3 + 4 Bl CI El F2 G3 HI 13-4 J2 K2 = Peabody.

A3 + 4 Bl-2 CI Dl El F2 G3 HI 12 J2-3 Kl = Carolina Extra

Early,
A3 +4 B2 CI Dl E2 F2 G4 HI 12 J3 Kl = Norton.
A4 + 3 Bl CI Dl El F2 G4 HI 13 J4 K2 = Pumpkin,
*A4 +3 B2 C Dl E2 F2 G4 HI 13 J3 K3 = Fire Brass.
Ao Bl CI Dl El F2 G3 HI 12 J2 K2 = Up River.
A5 Bl CI Dl El F2 G3 HI 12 J3 Kl = Yellow Jersey.
A5 Bl CI Dl El F2 G3 HI + 2 14 J2 K2 = Pved Jersey.
A5 Bl-2 CI Dl El F2 G3 HI 12 J2 K2 = Pepper's Choice.
A5 Bl-2 CI D3 El-2 Fl Gl H(l-)2 11-2 Jl-2 Kl =:Yellow

Straussberg.
A5 Bl-2 CI D4 El-2 Fl Gl H2 11-2 J2 K2 = Alabama.
A5 Bl-2 CI D4 El-2 Fl Gl H2 14 J2 K2-3 = Red Nansemond,
A5 Bl-2 CI Dl El-2 F2 G3 HI 14 J3 Kl = Van Ness Red.
A5 B2 CI D4 E2 Fl Gl H2 11-2 J2 K2 = Early General Grant.
A.5 B2 C2 Dl E2 Fl Gl HI 13 J3 K2 = Nancy Hall.
"A B C D2 E Fl Gl H3 12 J2 K3 = Yellow Manritius.

E. Alphabetical List of Varieties.

Alabama.— A.5 Bl-2 CI D4 El-2 Fl Gl H2 11-2 J2 K2.

Archer's Hybrid. — See Southern Queen, strong type.

Belmont.— Al B2 CI Dl El F2 G4 HI 13 J3 Kl.

Big Stem Jersey — See Yellow Jersey.

Black Spanish.— A3 B2 CI D4 E2 Fl Gl H2 14 J5 K3.

Brass Cannon.— A2 B2 C2 D3 E2 Fl Gl HI 14 J2 K3.

Brazilian. — See Southern Queen, strong type.

Bronze Spanish.— A3 B2 CI D4 E2 Fl Gl H3 12 Jl K3.

Carolina Extra Early.- A3 + 4 Bl-2 CI Dl El F2 G3 HI 12 J2 3 Kl.

Caroline Lee.- A2 B2 CI Dl El-2 F2 G4 Hl-2 II Jl K2-3.

Chazal.— A3(?) B(?) C(?) D2 E(?) Fl Gl H3 14 Jl K3.

Dooley. — See Ticotea.

Early Carolina. — See Up River,

Early General Grant.— A5 B2 CI D4 E2 Fl Gl H2 11-2 J2 K2.

Early Golden. — See Southern Queen, weak type.

l>lipse Sugar Yam. — See Georgia.

I'ire Brass.— A4 + 3 B2 C(?) Dl E2 F2 G4 HI 13 .J3 K3.

!• Inrida.— A4 r.2 C2 D2 E2 Fl Gl HI 12 J K.

* Plants wore too young to be fully determined, if the full
formula is not given.

THE SWEET POTATO. 83

Florida. — See Southern Queen, strong type.

Florida Bunch. — See Southern Queen, strong type.

Florida Yam, No. 2. — See Southern Queen, strong type.

General Grant. — See Florida.

General Grant. — See Southern Queen, weak type.

Georgia.--Al B2 CI Dl El-2 F2 G4 HI 12-3 J3 Kl.

Greorgia. — See Alabama.

Gold Coin. — See Southern Queen, strong type.

Gold Skin. — See Southern Queen, strong type.

Governor.— A3 B2 CI D2 El-2 Fl Gl H2 14 J2 K2.

Gros Gandia.— Al B2 C( ?) Dl E( ?) F2 G4 H2 13 Jl K3.

Halonaipu.— A4 B2 C2 D4 E2 Fl Gl H3 14 Jl K3.

Hamburg. — See Southern Queen, strong type.

Hayman. — See Southern Queen, strong type.

lluamoa.— Al B2 CI D4 El-2 Fl Gl H2-3 13 J3 K3.

Ihumai.— Al B2 CI D4 E2 Fl Gl H3 14 Jl K3.

India Red.— A3 B2 CI Dl El Fl Gl H2 14 Jl K.

John Burnet.— Al B2 CI D2 E2 Fl Gl H2 12 J2 K3.

Kala.— Al B2 CI Dl El-2 F2 Gl -f 3 H3 12 J2 K3.

Kapo.— A3 Bl CI D4 El Fl Gl H2 14 J5 K3.

Kauaheahe.— A3 B2 CI D2 E2 Fl Gl H3 14 Jl K3.

Kawelo.— Al B2 CI D2 E2 F2 G3 H3 13 Jl K3.

Kentucky White. — See Southern Queen, weak type.

Key West.— A3 B2 CI D4 E2 Fl Gl H3 13 J2 K2.

Koali. — See Kawelo.

Laiakona.— A3 Bl CI D4 El Fl Gl HI 12 J2 K3.

McCoy. — See Southern Queen, weak type.

Miles Yam. — See Southern Queen, strong type.

Minnet.— A2 Bl C2 Dl El-2 Fl Gl HI II J3 K3.

MuflFard.- A2(?) B(?) C(?) D4 E2 Fl Gl Hl-2 14 Jl K3.

Nancy Hall.— A5 B2 C2 Dl E2 Fl Gl HI 13 J3 K2.

Nigger Choker — Killer. — See Black Spanish.

Norton.— A3 + 4 B2 CI Dl E2 F2 G4 HI 12 J3 Kl.

Peabody.— A3 + 4 Bl CI Dl El F2 G3 HI 13-4 J2 K2.

Pepper's Choice.— A5 Bl-2 CI Dl El F2 G3 HI 12 J2 K2.

Pierson. — See Southern Queen, weak type.

Pikonui.— A3 B2 CI D4 E2 Fl Gl H3 12 J3 K3.

Pilipili.— A3 B2 CI D2 E2 Fl Gl H3 12 J2 K3.

Polo. — See Southern Queen.

Pu.— A4 B2 C2 D4 E2 F2 G2 3 H3 12 ( ?) J3 K3.

Pumpkin.— A4 + 3 Bl CI Dl El F2 G4 HI 13 J4 K2.

Red Bermuda. — See Red Nansemond.

84 THE SWEET POTATO.

Red Jersey —A5 Bl CI Dl El F2 G3 111-2 14 J2 K2.

Red Nansemond.— A5 Bl-2 CI D4 El-2 Fl Gl H2 14 .J2 K2-3.

Red Sealy — A4 Bl-2 C2 D2 E2 El Gl HI 13 Jl K2.

Roosevelt.— A3 B2 CI D2 E2 Fl Gl H3 11 Jl Kl-2.

Shanghai.— A2 B2 C2 D2 E2 Fl Gl H2 11-2 Jl KL

Southern Queen, strong type.— A4 B2 CI D3 E2 F2 G2 H2 11-2 J2 K2.

Southern Queen, weak type.— A4 Bl CI D3 El-2 F2 G2 H2 11-2 J2 K2.

Spanish (Yam). — See Black Spanish.

Strassburg, Straussberg.— See Southern Queen, weak type.

Sugar Yam. — See Georgia,

Sulla.— A3 B{?) C(?) Dl(?) E2(?) Fl(?) Gl H2 12 J2 K3.

Thompson's Favorite.— A3 Bl-2 C2 D4 E2 Fl Gl H2 II Jl K2.

Ticotea.— Al B2 C2 D2 E2 Fl Gl H2-3 II Jl K3.

Trinidadian, No. 1.— A2 B2 CI D2 E2 Fl Gl HI 14 J2 K3.

Trinidadian, No. 2.— A3 ( ?) B ( ?) C ( ?) Dl E2 ( ?) F2 G4 H2 13 J2 K2.

True Parson Prince. — See John Burnet.

Up River.— A5 Bl CI Dl El F2 G3 HI 12 J2 K2.

Up River. — See Southern Queen, weak type.

Van Ness Red.— A5 Bl-2 CI Dl El-2 F2 G3 HI 14 J3 KL

Vincentonian.— Al B2 C2 Dl E2 F2 G3 H2 14 J3 K2.

Vineless Beech.— A2 Bl-2 C2 Dl El-2 F2 G4 Hl-2 II Jl K3.

Violette Blanche.— A4 ( ?) B ( ?) C ( ?) D2 ( ?) E ( ?) Fl Gl H3 13 J5 K3.

Violette Rouge.— A4(?) B(?) CI D2 El Fl Gl H3 14 J5 K3.

White Colombia.— A4 Bl CI D3 El-2 Fl G4 H2-3 II ( ?) J2 Kl.

White Gilk.— Al B2 CI D2 E2 Fl Gl H2 II Jl K3.

White Gilkes, 3 mo.— See White Gilk.

White Gilkes, 6 mo.— See White Gilk.

White Sealy.— A2 B2 CI Dl E2 F2 G4 H2 11 J2 K3.

White Yam. — See Southern Queen, weak type.

Yellow Bean.— See Southern Queen, strong type.

Yellow Colombia.— A4 Bl ( ?) CI D3 El-2 Fl G4 H2 12 J2-3 Kl.

Yellow Jersey.— A5 Bl CI Dl El F2 G3 HI 12 J3 Kl.

Yellow Mauritius.— A ( ?) B( ?) C( ?) D2 E( ?) Fl Gl H3 12 J2 K3.

Yellow Red.— A3 B2 CI D4 E2 Fl Gl H2 13-4 J2 K3.

Yellow Straussberg.— A5 Bl-2 CI D3 El-2 Fl Gl H ( 1-) 2 11-2 Jl-2 Kl.

In addition to the above varieties a number of
seedlings and young plants too young for full
determination have been left by the writer in the care
of the United States Experiment Station at AVash-
ington.

THE SWEET POTATO. 85

The writer wishes it to be clearly understood that
lie elaborated the key for the convenience of workers
in this field, and that therefore he welcomes criti-
cisms and corrections.

For sweet i^otato material the writer feels espec-
ially indebted to Mr. ]^eattie, of the Washington
Experiment Station; Mr. Starnes, of the Georgia
Station; Mr. Bnrnette, of the Baton Kouge Station;
Mr. Rosa, of Milford, Delaware; Mr. Pittier, of
Washington, and to the Botanic Gardens of Jamaica,
]»arbadoes, Calcutta, Mauritius, and Honohilu. The
writer wishes to express his sincere thanks to Dr.
Macfarlane for the facilities for study olfered in the
]^otanic Gardens of the University and the many sug-
gestions made; to Mr. Trouncer, of Wenonah, N. J.,
for the generous otfer of his field, and to Mr. Maisch,
l^hotographer of the Philadelphia Museums, for the
care he has taken with the photographs.

F. Desceiptions.

Alabama (Georgia). Southern States.
Formula :

A5 Bl-2 CI D4 El-2 Fl Gl H2 11-2 J2 K2.

Stems long, between % and ^ inch thick, bright light purple.

Star clear, but small. Lower surface of veinss dark purple. Hair
chiefly along the veins, scant.

Type leaf Fig. 71.

Differs from Early General Grant only in its less vigorous growth.

Belmont. Washington. United States.
Formula :

Al B2 CI Dl El F2 G4 HI 13 J3 Kl.

Stems thin, about ^^ of an inch in diameter, green. Often broAvnish
or even faintly purplish where exposed to the sun. Hairy.

Large grown leaves, about 4-4^^ inches across from tip to tip.

Type leaf figure omitted by mistake, resembles Georgia very
closely.

86 THE SWEEl POTATO.

Black Spanish. Various States of the Union.

Formula :

A3 B2 CI D4 E2 Fl Gl H2 14 J5 K3.

Stems very long, fg inch in diameter, very dark purple. Petioles
also purple, except those near the tip, if shaded.

Star bright and large. Leaves dark colored, i^owcr surface of
veins purple. Leaves often 5 inches long and broad. Type leaf
Fig. 50. Latex very abundant, dripping off when stem is cut.

Hair mostly on tip only, or continued down on larger veins.

Brass Cannon. Jamaica.

Formula:

A2 B2 C2 D3 E2 Fl Gl HI 14 J2 K3.

Stems short, thick (t\ inch), greenish brown, to purple on upper
side.

Star large, lower surface of veins strongly marked purple. Hair
thick all over the upper surface of the leaf.

Type leaf Fig. 40.

Bronze Spanish. Washington, D. C.

Formula :

A3 B2 CI D4 E2 Fl Gl H3 12 Jl K3.

Stems long, fg inch in diameter, or a little thinner, but never
thinner than % inch, very dark purple. Petioles also purple.

Star bright and large. Leaves dark colored. Lower surface of
veins purple. Leaves often 5 inches or more long and broad.

Hairs practically absent.

Tj-pe leaf Fig. 51.

Carolina Extra Early. Georgia, etc.

Formula:

A3 + 4 Bl-2 CI Pl El F2 G3 HI 12 J2-3 Kl.

Stems long, less than % inch in diameter, green. Faint purplish
spot at the lower surface of the base of the midrib.
Type leaf Fig. 63.

THE tSVYEET POTATO. 87

Caroline Lee. Louisiana.

Formula :

A2 B2 CI Dl El-2 r2 G4 Hl-2 II Jl K2 3.

Stems longer than lour feet, rarely more than ^ inch thick, light
green.

Hair almost all over young leaves, all over tip and along the
veins in old ones.

Type leaf Fig. 36.

Slightly shouldered leaves occur occasionally, even among the later
leaves.

Early General Grant. Various States of the Union.

Formula :

A5 B2 CI D4 E2 Fl Gl H2 11-2 J2 K2.

Stems long, i% inch thick, purple. Star clear, veins strongly
marked purple below.

Hair scant along the veins.
Type leaf Fig. 75.

Fire Brass. Jamaica.

Formula:

A4 B C Dl E F2 G4 HI 13 J3 K3.

Stems light green to brownish, but without purple.
Hair all over leaf.

Florida. Various States of the Union.

Formula :

A4 B2 C2 D2 E2 Fl Gl HI 12 J K.

Stems about 1 foot long. Plant very bushy. Stems ^g to %
inch thick, green, with purple spots around the axils of the leaves.

Star very large. Vein strikingly marked purple below. Leaves
very much sunk in the center, deeply saucer-shaped.

Type leaf Fig. 59.

88 THE SWEET POTATO.

Georgia (Eclipse Sugar Yam.) United States.

Formula :

Al B2 CI Dl El-2 r2 G4 HI 12-3 J3 Kl.

Stems long, often ^ thick, but usually between -^^ and -^ inch
thick, long, green, with brownish sunburnt patches where exposed.
Large leaves 6 inches across from tip to tip.
Type leaf Fig. 27.

Governor. Jamaica.

Formula :

A3 B2 CI D2 El-2 Fl Gl H2 14 J2 K2.

Stems not much longer than 4 feet as a rule; about ^ inch thick
at thickest point, but often thinner; greenish brown with purple
around the axils of the leaves.

Star rather faint, lower surface of veins ralher weakly colored
purple.

Type leaf Fig. 45.

Halonaipu. Sandwich Islands.

Formula :

A4 B2 C2 D4 E2 Fl Gl H3 14 Jl K3.

Stems about 3 feet, -^ inch in diameter, purple.
Star plain, veins purple below.
Hair absent.
Type leaf Fig. 60.

Large leaves, over 4 inches across, but most leaves 3-4 inches in
diameter.

Huamoa. Sandwich Islands.

Formula :

Al B2 CI D4 El-2 Fl Gl H2-3 13 J3 K3.

Stems long, thin or medium-sized, purple. Star large and clear.
Type leaf Fig. 32; little over 4 inches in diameter.
Hair very scant on veins, or almost entirely absent.

THE SWEET POTATO. 89

Ihumai. Sandwich Islands.
Formula :

Al B2 CI D4 E2 Fl Gl H3 14 Jl K3.

Stems not much longer than 4 feet, extremely thick, sometimes
measuring over V^ inch in diameter at the thickest point. Very
dark purple.

Star extending half-way up the veins or farther, bright. Under
side of veins brilliantly set off in purple in younger opened leaves.

Hair absent from surface of the leaf, present only at the edge.

Type leaf Fig. 33.

India Red India.

Formula :

A3 B2 CI Dl El Fl Gl H2 14 Jl K.

Stems long, whitish-green, with pinkish-purple blotches around
the axils of the leaves and occasionally in other places; thin, only
rarely over % inch in diameter.

Star large, bright red.

Type leaf Fig. 44; largest leaves over 4 inches across, but most
full-grown leaves are 3-4 inches across.

Hair scattered along the larger veins.

The young leaves, just opened, are characteristically colored
pinkish-brown.

John Burnet. (True Parson Prince). Jamaica.

Formula :

Al B2 CI D2 E2 Fl Gl H2 12 J2 K3.

Stems little over 4 feet, -^ inch thick, green, with purple marks
around the axils of the leaves and purple sunburns on the upper
surface, where exposed.

Star small, but plain. Lower surface of veins well marked with
purple, out to the margin of the leaf.

Leaf like type Fig. 30, not cordate at the base. Hair rather far
down the central lobe, and along larger veins.

Large leaves, about 5 inches across from tip to tip.

Latex very abundant.

90 THE SWEET POTATO.

Kala. Sandwich Islands.

Formula :

Al B2 CI Dl El-2 F2 Gl + 3 H3 12 J2 K3.

Stems little longer than 4 feet, green, with here and there brownish
or slightly purplish patches, ts inch in diameter or less.

No star. On the lower surface of the base of the midrib there
is a purple spot, usually noticeable only in young leaves. In some
leaves even the lower surface of the veins is purplish.

Type leaf Fig. 28.

Kapo. Sandwich Islands.
Formula :

A3 Bl CI D4 El Fl Gl H2 14 J5 K3.

Stems long, % inch in diameter, very rarely thicker; dark purple,
lighter purple at the tip, but even there dark purple around the
axils of the leaves.

Star very small. Lower surface of veins rather light purple.
Latex very abundant, dripping ofT from a cut stem.

Type leaf Fig. 42.

Kauaheahe. Sandwich Islands.
Formula :

A3 B2 CI D2 E2 FI Gl H3 14 Jl K3,

Stems long, green, with purple around the base of the leaves and
near the tip. The tips are either entirely purple for about six
inches, or green with purple around the axils of the leaves.

Star small, but clear. Veins rather weakly marked with purple on
the lower surface.

Hairs absent on surface, scattered on the edge of the leaf.

Type leaf Fig. 48.

Kawelo. (Koali), Sandwich Islands.

Formula :

Al B2 CI D2 E2 F2 G3 H3 13 Jl K3,

Stems long, i% inch or more in diameter, whitish, with purple
around the axils of the leaves, and occasional sunburnt patches on
the upper surface. Leaves very light yellowish-green.

Type leaf Fig. 31.

Especiallj' in young, full-grown leaves there is a faint purple spot
on the lower sxirface of the midrib, where it springs from the petiole.

Hair absent.

THE SWEET POTATO. 91

Key West, ^^'ashington, D. C.

Formula :

A3 B2 CI D4 E2 Fl Gl H3 13 J2 K2.

Steins very long, fg inch diameter or more at thickest point, bright
purple, but often green at the very base.

Star large, extending up the veins; lower, surface of veins
strongly marked purple.

Type leaf Fig. 53.

Leaves often 6-8 inches broad; hair absent on ^-urface, present
along edge.

•o

Laiakona. Sandwich Islands.

Formula :

A3 Bl CI D4 El Fl Gl HI 12 J2 K3.

Stems little more than 4 feet long, seldom more than % inch in
diameter, purple.

Star small, lower surface of veins strongly marked purple. Hair
almost all over the upper surface of leaf.

Type leaf Fig. 41.

Nancy Hall. Louisiana.

Formula :

A5 B2 C2 Dl E2 Fl Gl HI 13 J3 K2.

Stems less than 4 feet long, f^ to % inch in diameter, entirely
green. The plant has a bunch habit. Star plain, veins purple below.
Most leaves are slightly lobed on the sides, but ma«y are without
projections.

Type leaf Fig. 74.

Hair all over upper surface of loaf.

Norton. Washington, D. C.

Formula:

A3 + 4 B2 CI Dl E2 F2 C;4 HI 12 -73 Kl.

Stems long, -fs inch or more in diameter, green, with occasional
patches of brownish in exposed places.

Hair very thickly covering the entire upper surface of the leaf.
Type leaf Fig. 64.

92 THE SWEET POTATO.

Peabody. Louisiana.

Formula:

A3 + 4(?) Bl 01 Dl El F2 G3 HI 13-4 J2 K2.

Stems long, % inch in diameter, green. Leaf yrith a purplish spot
on the under side of the base of the midrib.

Hair thick all over upper surface.

Both long-lobed and broad-lobed leaves frequent. Leaves about 3
inches across.

Type leaf Fig. 62.

Pepper's Chmce. Delaware.

Formula :

A5 Bl-2 01 Dl El F2 G3 HI 12 J2 K2.

Stems long, less than % inch in diameter, light green.

Leaves round and lobed, of about equal frequency, or the lobed
ones predominating. Most leaves are less than 4 inches across, but
some are larger.

In young leaves there is a faint purple spot on the under side
of the base of the midrib.

Hair all over the upper surface of the leaf.

Type leaf Fig. 69.

Pikonul. Sandwich Islands.

Formula :

A3 B2 01 D4 E2 Fl Gl H3 12 J3 K3.

Stems not much longer than 4 feet, ^ inch thick or thicker.
Star very small, but bright.
Veins purple below. Hairs absent.
Type leaf Fig. 52.

Pilipili. Sandwich Islands.

Formula j

A3 B2 01 D2 E2 Fl Gl H3 12 J2 KS.

Stems long, ^ inch in diameter, light green, with small purple
spots in the axils of the leaves. Star very small, but clear.
Veins purple below.

THE SWEET POTATO. 93

Leav€3 range from light yellowish-green to rather dark green
on the upper surface.

The largest leaves on long stems are the darkest.

Type leaf Fig. 47.

Hairs absent from upper surface of leaf, present on the edge.

Pu. Sandwich Islands.

Formula :

A4 B2 C2 D4 E2 F2 G2-3 H3 12 J3 K3.

Stems about 2 feet, ^ inch thick, pinkish-purple. Leaves, which
are on purplish petioles, have the lower surface of the midrib with a
hue of pink. Sometimes there is a purplish spot at the base of
the midrib on the lower surface.

Type leaf Fig. 61.

Pumpkin. Various States of the Union.

Formula :

A4 -f 3 Bl CI Dl El F2 G4 HI 13 J4 K2,

Stems long, % inch in diameter, green, with occasional brownish
places, where exposed to the sun. Exceptionally a few leaves have
a hue of pink on the lower side of the midrib.

Hair thick all over upper surface.

Type leaf Fig. 65.

Red Jersey. Washington, D. C.

Formula :

A5 Bl CI Dl El F2 G3 Hl-2 14 J2 K2.

Stems long, less than % inch in diameter, green, brownish in
patches.

A purplish spot on the under side of the base of the midrib.
Leaves rarely more than 3 inches across, both round and lobed.

Hairs all over young open leaves, but chiefly on the veins in the
oldest ones.

Type leaf Fig. 68.

iU THE {sWEET POTATO.

Red Nan»®inni®nd i Ked NanceiiKm.l— lU-.l Bermiula > . Louisiana.

l''omiula :

A5 Bl-2 CI U4 El 2 K I Gl IJ2 14 J2 K2-3.

Stems long, between Vs and & inch thick, purple.
Star clear, lower surface of veins strongly marked purple.
Hair scant along veins and li[i of median lobe..
Most leaves a little less than -I: inches across, especially the round
ones, but many measure more ihnii 4 inches.
Type leaf Fig. 72.

Red SeeAy. Jamaica.

I'ormula t

A4 P.1-2 C2 D2 E2 Fl Gl HI 13 Jl K2.

Stems about :i feet long. Plant bushy. Stems i\s inch in diameter,
green, with a purplish crescent around the base of each petiole, which
sometimes extends slightly over the side of the stem.

Star small, but bright. Veins purple below. Hair all over upper
surface of leaf.

Type leaf Fig. 57.

Roosevelt. Jamaica.

Formula :

A3 B2 CI D2 E2 Fl Gl HS II Jl Kl-2.

Stems long, ife inch in diameter at the thicke-t point, greenish-
brown, with purple around the axils of the leaves.
Star small and bright.

Lower surface of veins well marked purple.
Hair absent.
Type leaf Fig. 46.
Large leaves often over 5 inches across.

Shanghai. Washington, D. C.

I'^ormula :

A2 B2 C2 D2 E2 Fl Gl H2 11-2 Jl Kl.

Stems less than 4 feet long, -f.j inch thick or thicker, green, with

)>urple marks around the axils of the. leaves and in other places.

Star very small, veins purple on lower surface in old leaves.

THE SWEET POTATO. 95

Hair plentiful along the veins, rarely between the veins, except
at the tip.

The young open leaves are typically purplish, with the veins darker
purple on the upper surface.

Type leaf Fig. 39.

Southern Queen, Strong type.

(Miles Yam, Hayman, Archer's Hybrid, Gold Coin, Florida Bunch,
Florida, Brazilian, Florida Yam No. 2, Gold Skin, Caroline Lee,
Yellow Bean, Polo, Hamburg.)

Formula:

A4 B2 CI D3 E2 F2 G2 H2 11-2 J2 K2.

Stems long, ys-V* inch thick, greenish-brown to purple.
Midrib of old leaves, e52>eeially such as have purplish petioles, is
often pinkish below.

Leaf often 5-6 inches in its widest diameter.
Hair chiefly along the veins.
Type leaf Fig. 58.

Southern Queen. Weak type.

(Early Golden, McCoy, White Yam, Kentucky W^hite, General
Grant, Up River, Pierson.)

Formula :

A4 Bl CI D3 El-2 F2 G2 H2 11-2 J2 K2.

Leaves sometimes over 4 inches, but usually 3 inches in diameter.
Otherwise like the strong type of Southern Queen.
Type leaf Fig. 56.

Thompson's Favorite. (Jamaica).
Formula :

A3 Bl-2 C2 D4 E2 Fl Gl H2 II Jl K2.

Stems short, rarely over 18 inches, ^ inch thick or thicker, dark
purple.

Star very large and striking, running up the veins in thick
purple lines. Lower surface of veins purple. All older petioles are
entirely dark purple. Hair scant on the tip and along the veins.

Type leaf Fig. 43.

]\Iost leaves are less than 4 inches across, but some measure more
than 4 inches.

96 THE SWEET POTATO.

Ticotea (Dooley). Southern States.

Formula :

Al B2 C2 D2 E2 Fl Gl H2-3 II Jl K3.

Stems short, -i^ inch thick or thicker, green, with purple patches
around the axils of the leaves. The plant forms thick bunches.

Star bright and large, veins very strongly marked purple on lower
surface. Hair scant along the veins, or almost absent.

Type leaf Fig. 35.

Trinidadian No. 1. Jamaica.

Formula :

A2 B2 CI D2 E2 Fl Gl HI 14 J2 K3.

Stems usually less than 4 feet long, but occasionally more, very
hairy all over down to the base, ^ inch thick at thickest point,
purplish or purple.

Star large and clear, running part way up the veins in very thin
lines. Practically all leaves are round.

Hairs thin all over the upper surface. Lower surface of veins
strongly marked purple.

Type leaf Fig. 38.

Up River (Early Carolina). Various States of the I niuii.

Formula :

A5 Bl CI Dl El F2 G3 HI 12 J2 K2.

Stems long, thin, less than % inch as a rule, never as much as
^ inch thick, green.

A faint purplish spot on the \inder side of the base of the midrib.

Both round and lobed leaves frequent. Leaves usually less than
3 inches across.

Type leaf Fig. 66.

Van Ness Red. Wa&hington, D. C.

Formula :

A5 BI-2 CI Dl El-2 F2 G3 HI 14 J3 Kl.

Stems long, less than % inch in diameter, with some stems a
little tliicker, green, with brownish patches where exposed.

Leaves mostly less than 4 inches in diameter, sometimes a little

THE SWEET POTATO. 97

larger. Small purplish spot on the louver side of the base of the
midrib. Hair all over upper surface of leaf.
Type leaf Fig, 73.

Vincentonian. Jamaica.

Formula :

Al B2 C2 Dl E2 F2 G3 H2 14 J3 K2.

Stems usually less than 3 feet long, -^ inch thick or more, green,
with occasional blotches of sunburn.

Plant of bushy habit.

Veins faintly purple for a short distance on the lower surface, or
only at the place of junction of the primary veins.

Hair over a large part of the tip of the center lobe, continued
down along the veins.

Type leaf Fig. 34.

Vinelets Beech. Louisiana.
Formula :

A2 Bl-2 C2 Dl El-2 F2 G4 Hl-2 II Jl K3.

Stems short, rarely over 3 feet, green, between y^,, and -^s inch
in diameter.

Hair covering a large part of the apex of the leaf, and continued
down along the veins.

Latex very abundant. All leaves are round.

Type leaf Fig. 37.

Eesembles Caroline Lee, but differs from it in the length and
thickness of stalk, and size and shape of leaf.

White Colombia. Colombia.
Formula :

A4 Bl CI D3 El-2 Fl G4 H2-3 II J2 Kl.

( Plants not old enough to be certain about all the measurements ) .

Stems long, between % and ^ inch thick, green to purple. Star
very large in young leaves, disappearing in old ones; at times not
present in young leaves. Color of leaves ranging from deep purple
to green.

Hairs few along the veins, or absent.

Type leaf Fig. 55.

So far no difference has been observed between this and Yellow
Colombia.

98 THE SWEET POTATO.

White Gilk. Southern States and West Indies.

= (W. Gilkes, 3 months.)
(W. Gilkes, 6 months.)

Formula :

Al B2 CI D2 E2 Fl Gl H2 II Jl K3.

Stems not much longer than 4 feet, fg inch thick, greeii', with
purplish spots around the axils of the leaves and in other places.

Star conspicuous, veins strongly marked with purple on the under
surface. Hair scattered along veins and on the tip.

Type leaf Fig. 29.

Differs from John Burnet in having the star smaller in the young
leaves and larger in the old ones.

White Sealy. Louisiana and West Indies.

Formula :

A2 B2 CI Dl E2 F2 G4 H2 II J2 K3.

Stems long, thick, green, no purple on entire plant.

Type leaf omitted by mistake.

Hair scattered along midrib and principal veins.

Yellow Colombia.

Formula :

A4 Bl( ?) CI D3 EI-2 Fl G4 H2 12 J2-3 KI.

(Plants not old enough to be certain about all the measurements.)
See description of \^Tiite Colombia.
Type leaf Fig. 54.

Yellow Jersey. (Big Stem Jersey). Washington, D. C.

Formula :

A5 Bl CI Dl El F2 G3 HI 12 J3 Kl.

Stems long, less than % inch in diameter, light green.
Purple spot under side of the base of the midrib.
Type leaf Fig. 67.

THE SWEET POTATO. 98

Yellow Red. Louisiana.

Formula :

A3 B2 CI D4 E2 Fl Gl H2 13-4 J2 K3.

Stems long, ^s inch in diameter at the thickest point, dark purple.
Star well marked, lower surface of veins purple.
Hairs on tip of middle lobe and along large veins.
Type leaf Fig. 49.

Yellow Strau8«berg. Various States of the Union.

Formula :

A5 B1-2C1 D3 El-2 Fl Gl Hl-2 11-2 Jl-2 Kl.

Stems long, between % inch and ^ inch in diameter, greenish-
brown to purple, but purple for most of their length.
Halberd-shaped leaves prevailing.
Many leaves over 4 inches in diameter, but most leaves less than

4 inches across.

Star faint, veins faintly purple on the lower surface.
Hairs scant along the veins.
Type leaf Fig. 70.

V. LIST OF ILLUSTRATIONS.

Fig. 1. Surface view of lower epidermis of the leaf of jSTorton.

a. Normal cell.

b. Cells modified over bundle traces.

c. Cells modified around gland cells.

d. Gland cells.

e. ytomata.

Fig. 2. . Witlldra^fn.

Fig. 3. Striations on lower epidermis of leaf of Norton.
Fig. 4. Striations on upper epidermis of leaf of Norton.

Fig. 5. . Withdrawn.

Fig. 6. Cross-section through a leaf of Up River,
a. Gland cells.

Fig. 7. . Withdrawn.

Fig. 8. . Withdrawn.

Fig. 9. Section through midrib of Van Ness Red.

a. Modified epidermis.

b. Collenchyma.

c. Latex canals.

d. Internal phloem.

e. Abnormal xylem inside of internal phloem.

Fig. 10. Perspective section of a petiolar nectary of Yellow Jersey.
Fig. 11. Section of petiole of Georgia, two inches from the base, to
show arrangement of bundles.

a. Internal phloem.

b. Dividing cells between bundles.

c. Crystal cells.

d. Collenchyma.

Fig. 12. Section of petiole of Up River, one inch from the base, same
magnification as Fig. 11.

a. Endodermis.

b. Hair.

Fig. 13. Section through tip of stem of Van Ness Red.

a. Hairs.

b. Gland cells.

lOO

THE SWEET POTATO. 101

Fig. 13 — Continued.

c. Latex canals.

d. Endoderniis.

e. Pericambium.

f. Protoxylem.

g. Internal phloem.

Fig. 14. Diagram of cross-section of base of stem of Florida.

a. Epidermis.

b. Cortex.

c. Endodermis.

d. Pericambium,

e. Phloem.

f. Cambium.

g. Xylem.
h. Vessels,
i. Tyloses.

k. Cambium of internal phloem.

1. Fundamental tissue.
FitT. 15. Cross-section of old stem of Southern Queen.

a. Epidermis.

b. Hypodermis.

c. Cortex.

d. Latex canals.

e. Endodermis.

f. Pericambium,

g. Phloem,
h. Cambium,
i. Xylem.

j. Vessels,

k. Protoxylem.

1. Internal cambium,

m. Internal phloem,

n. Fundamental tissue.
Fig. 16. Development of reversed bundle in cross-section of old

stem of Red Jersey.

a. Normal metaxylem.

b. Protoxylem.

c. Internal phloem. (?)

d. Abnormal metaxylem.

e. Internal phloem.

f. Oil cells in fundamental tissue.

g. Crystal cells.

102 THE SWEET POTATO.

Fig. 17. Longitudinal section of old stem of Georgia.

a. Epidermis.

b. Hypodermis,

c. Cortex.

d. Latex canals.

e. Endodermis.

f . * Pericambium.

g. Phloem.

h. Vessel with reticulated thickenings,
i. Vessel with bordered pits,
j . Protoxylem.
k. Internal phloem.
1. Fundamental tissue.
Fig. 18. Crystal cells in longitudinal section of pith of Southern
Queen, weak type.
(In course of foi'mation) (?).

a. Crystals.

b. Protoplasm.

c. Nucleus.

d. Nucleolus.

e. Double crystal.

f. Crystal, isolated in the corner of a pith cell.
Fie. 19. Latex canals in Van Ness Red.

a. In cortex of tip of stem.

b. In pith of old stem.

Fig. 20. Lenticel or proliferation of the epidermis of old stem of

Georgia.
Fig. 21. A tuber of Red Jersey, fascicular type, smooth, with roots
coming off in vertical rows near lenticels. Lenticels not
conspicuous.
Fig. 22. A tuber of Florida, spherical, five-lobed, with roots coming

off in the sunken areas. Lenticels conspicuous.
Fig. 23. A tuber of Up River, cylindrical, smooth, with lenticels

inconspicuous.
Fig. 24. A tuber of Shanghai, with many conspicuous lenticels and

long and large roots.
Fig. 25. a. A tuber of Belmont, with five longitudinal veins.

Roots deeply sunk between the veins. Lenticels
scattered, inconspicuous,
b. Diagram of cross-section through tuber, showing
the five exterior bundles.
Fig. 26. A tuber of Pumpkin, showing anastomosing veins. Lenti-
cels conspicuous.

THE SWEET POTATO. 103

Fie. 27. Georgia. All photographs of branches are reduced on the
same scale, so that the photographs show exact compara-
tive size.

Fig. 28. Kala.

Fig. 29. White Gilk.

Fig. 30. John Burnet.

Fig. 31. Kawelo.

Fig. 32. Huanioa.

Fig. 33. Ihuniai.

Fig. 34. Vineenlonian.

Fig. 35. Ticotea.

Fig. 36. Caroline Lee.

Fig. 37. Vlneloss Beech.

Fig. 38. Trinidadian No. 1.

Fig. 39. Shanghai.

Fig. 40. Brass Cannon.

Fig. 41. Laiakona.

Fig. 42. Kapo.

Fig. 43, Thompson's Favorite.

Fig. 44. India Red.

Fig. 45. Governor.

Fig. 46. Roosevelt.

Fig. 47. Pilipili.

Fig. 48. Kaiiaheahe.

Fig. 49. Yellow Red.

Fig. 50. Black Spanish.

Fig. 51. Bronze Spanish.

Fig. 52. Pikonui.

Fig. 53. Key West.

Fig. 54. Yellow Colombia.

Fig. 55. \Miite Colombia.

Fig. 56. Southern Queen, weak type.

Fig. 57. Red Sealy.

Fig. 58. Southern Queen, strong type.

Fig. 59. Florida.

Fig. 60. Halonaipu.

Fig. 61. Pu.

Fig. 62. Peabody.

Fig. 63. Carolina Extra Early.

Fig. 64. Norton.

Fig. 65. Pumpkin.

Fig. 66. Up River.

104 THE SWEET POTATO.

Fig. 67. Yellow Jersey.

Fig. 68. Red Jersey.

Fig. 69. Pepper's Choice.

Fig. 70. Yellow Straussberg.

Fig. 71. Alabama.

Fig. 72. Red Nansemond.

Fig. 73. Van Ness Red.

Fig. 74. Nancy Hall.

Fig. 75. Early General Grant.

Bot. Contrib. Univ. Penn.

Vol. IV, Plate I.

Fig. 1.

Fig. 3.

Fig. 4.

Fig. 6.

Bot. Cotitrib. Univ. Pom.

Vol. IV, Plate IL

Fig. 9.

—r

Fig. 11.

Bot. Contrib. Univ. Penn.

\o\. IV, Plate III.

Fig. 12.

Fig. 13.

Fig. 16.

Fig. 14.

Bot. Contrib. Univ. Perm.

Vol. IV, Plate IV.

i,.,'i'.^7ri7^w

Fig. 15.

^Bi

□

Fig. 18.

Fig. 10.

&n^ot

Fig. 17.

Fig. 20.

Bot. Contrih. Univ. Penn.

Vol. IV, Plate V.

fl

Fig. 25b

^:

^ae^y*-

■5*.

-;i^^-

/

^

Fig. 22.

^ Fig. 23.

Fig. 21.

1

\

«-^

/

Fig. 24.

Fig. 25.

1^

Fig. 26.

Bot. Contrih. Univ. Penn.

Vol. IV, Plate VI.

Fig. 27 — Georgia.

Bot. Contrib. Uiik-. Pcnn.

Vol. IV, Plate VII.

Fig. 28— Kala.

lot. Contrib. Ciiiz: Pcnn.

Vol. IV, Plate VIIL

Fig. 29 — White Gilk.

Bot. Contrib. Univ. Perm.

Vol. IV, Plate IX.

Fig. 30 — John Burnet.

Bot. Contrib. Univ. Pcnn.

Vol. IV, Plate X.

Fig. 31 — Kawelo.

Bot. Contrib. Unh: Pcnn.

Vol. IV, Plate XL

Fig. 32 — ^Huamoa.

Bot. Contrib. Univ. Pcnn.

Vol. IV, Plate XII.

Fig. 33 — Ihumai.

Bot. Co)itrib. Univ. Penn.

Vol. IV, Plate XIIL

Fig. 34 — ^Vincentonian.

Bot. Contrib. U)iiz: Penn.

Vol. IV, Plate XIV.

Fig. 35 — Ticotea.

Bot. Contrib. Univ. Pciin.

Vol. IV, Plate XV.

Fig. 36 — Caroline Lee.

Bot. Contrih. Univ. Penn.

Vol. IV, Plate XVL

Fig. 37 — Vineless Beech.

Bot. Contrib. Univ. Pcnn.

Vol. IV, Plate XVII.

Fig. 38 — Tkinidadian Ko. 1.

Bot. Contrib. Unii'. Pcnn.

Vol. IV, Plate XVIIL

Fig. 39 — Shanghai.

Bot. Contrib. Univ. Pcnn.

Vol. IV, Plate XIX.

Fig. 40 — Brass Cannon.

Bot. Coiitrib. Univ. Pcnn.

Vol. IV, Plate XX.

Fig. 41 — Laiakona.

Bot. Coutrib. Univ. Pcnn.

Vol. IV, Plate XXI.

Fia. 42— Kapo.

Bot. Contrib. Unii'. Penn.

Vol. IV, Plate XXII.

Fig. 43 — Thompson's Favoeite.

Bot. Contrib. Univ. Pom.

Vol. IV, Plate XXiil.

Fig. 44 — India Red.

Bot. Contrih. Univ. Penn.

Vol. IV, Plate XXIV.

Fig. 45 — Go\'ERKor.

Bot. Coiitrib. Univ. Penn.

Vol. l\\ Plate XXV.

Fig. 46 — Roosevelt.

Bot. Contrib. Uiiii: Pcnii.

Vol. IV, Plate XXVI.

Fig. 47 — Pilipili.

Bot. Contrib. Univ. Penn.

Xo\. IV, Plate XXYIL

Fig. 48 — Kauaheahe.

Bot. Contrib. Univ. Pcnn.

Vol. IV, Plate XXVIII.

Fig. 49 — Yellow Eed.

Bot. Contrib. Univ. Penn.

Vol. IV. Plate XXIX.

Fig. 50 — Black Spanish.

Bot. Contrib. Univ. Penn.

Vol. IV, Plate XXX.

Fig. 51 — Bronze Spanish.

Bot. Contrib. Univ. Pciiii.

Vol. IV, Plate XXXI.

Fig. 52 — Pikonui.

Bot. Contrib. Univ. Pciui.

Vol. IV, Plate XXXIL

Fig. 53 — Key West.

Bot. Contrib. Univ. Pom.

\'ol. IV, Plate XXXIII.

Fig. 54 — ^Yellow Colombia.

Bot. Contrib. Uiih'. Pcnn.

\\A. IV, Plate XXXIV.

Fig. 55 — White Colombia.

Bot. Contnb. Univ. Pcnn.

Vol. IV, Plate XXXV.

Fig. 56 — Southern Queen, Weak Type.

Bot. Contrib. Unir. Penn.

Vol. IV, Plate XXXVI.

Fig. 57 — Eed Sealy.

Bot. Contrib. Univ. Penn.

Vol. IV, Plate XXXYIL

Fig. 58— Southern Queen, Strong Type.

Bot. Contrib. Uniz: Pom.

Vol. IV, Plate XXX\ III.

^^^^^^^^^^^^m .^^^^^^^^^^^^^^^^Hl i

.^^i^^^^^^^^^^^l

^^^^I^^HI^R^^^^^^H

VH

Fig. 59 — Florida.

Bot. Contrib. Univ. Penu.

Vol. IV, Plate XXXIX.

Fig. 60 — Halonaipu.

Bot. Contrib. Unh: Pcini.

Vol. IV, Plate XL.

Fig. 61— Pu.

Bot. Contrib. Univ. Penn.

Vol. IV, Plate XLI.

Fig. 62 — Peabody.

Bot. Contrib. Univ. Perm.

Vol. IV, Plate XLIl.

Fig. 63— Carolina Extra Early.

Bot. Contrib. Univ. Pciui.

Vol. IV, Plate XLIII.

Fig. 64 — Norton.

Bot. Contrib. Univ. Penn.

Vol. IV, Plate XLIV.

Fig. 65 — Pumpkin.

Bot. Contrib. Uiih: Pciui.

Vol. IV, Plate XLV.

<^'J

Fig. 66 — ^Up Riveb.

Bot. Contrib. Univ. Penn.

Vol. IV, Plate XLVI.

Fig. 67 — Yellow Jersey.

Bot. Contrih. Univ. Pciiii.

Vol. IV, Plate XLYll.

Fig. 68 — Red Jersey.

Bot. Contrib. Univ. Penn.

Vol. IV, Plate XLVIIL

Fig. 69 — Peppeb's Choice.

Bot. Cofitrib. Univ. Perm.

Vol. IV, Plate XLIX.

Fig. 70 — Yellow Stbaussberg.

Bot. Contrib. Univ. Pciiii.

Vol. IV, Plate L.

Fig. 71 — Alabama.

Bot. Contrib. Univ. Pciiii.

Vol. IV, Plate LI.

Fig. 72 — Red Nansemond.

\

Bot. Contrib. Univ. Penn.

Vol. IV, Plate LII.

Fig. 73 — Van Ness Red.

Bot. Coiitrib. Univ. Penn.

Vol. IV, Plate LIIL

Fig. 74 — Nancy Hall.

Bot. Contrib. Univ. Pcnn.

Vol. IV, Plate LIV.

Fig. 75 — Eaely Genebai, Grant.

Vol. IV 1919 No. 2

CONTRIBUTIONS

FROM THE

Botanical Laboratory

OF THE

University of Pennsylvania

University of Pennsylvania

Philadelphia

1919

CONTENTS OF VOLUME IV, No. 2

Water Content {in4 Transpirfition of Leaves. By A^sh^ W. Clark, PH.p 105

Chasmogamic and Cleistogamic Flowers of Impatiens Fuiva. By Franidin E.

Carroll, M.A., ph.d. (With Plates LV-LVII) 'l44

The Comparative Histology' and Irritability of Sensitive Plants. By D. Waltear

Steckbeck, a.m., ph.d., (With Plates LVIII-LXV) 185

Observations oh the Beach Plum. By John Young Pennypacker, a.m., ph.d.

(With Plates LXVI-LXX) 231

Production of New Cell Formations in P lants. By William R. Taylor, B.s.,
M.S. (With Plates LXXI-LXXVIII) 271

The Peanut, its History, Histology, Physiology and Utility. By Ralph A. Wal-

dron, B.S., M.S. (With Plates LXXXIX-LXXX) : 301

The Comparative Morphology, Taxonomy, and Distribution of the Myricaceae
of the Eastern United States. By Heber W. Youngken, a.m., m.s., ph.d.
(With Plates LXXXI-XC) 339

A Comparative Study of Floerkea Proserpinacoides and Allies. By Alice M.

Russell, B.S., M.S. (With^ Plates XCI-XCII) 401

Biochemical Studies of Insectivorous Plants.

I. The Work of Previous Investigators on Nepenthes. By Joseph S.

Hepburn, a.m., m.s.,* ph.d 419

II. A Study of the Protease of the Pitcher Liquor of Nepenthes. By

Joseph S. Hepburn, a.m., m.s., ph.d 442

III. A Bacteriological Study of the Pitcher Liquor of Nepenthes. By Jo-

seph S. Hepburn, ph.d., and E. Qumtard St. John, m.d 451

IV. Antiproleases in the Larvae of the Sarcophaga Associates of Sarracenia

flava. By Joseph S. Hepburn and Frank M. Jones 460

Vol. IV 1919 No. 2

CONTRIBUTIONS

FROM THE

Botanical Laboratory

OF THE

University of Pennsylvania

University of Pennsylvania

Philadelphia

1919

SEASONAL VARIATION IN WATER CONTENT

AND IN TRANSPIRATION OF LEAVES

OF FAGUS AMERICANA, HAMAMELIS

VIRGINIANA, AND QUERCUS ALBA.

BY
Arabel W. Clark, Ph. D.

With Chart Figures I— XXXIII.

(Thesis presented to the Faculty of the Graduate School
in partial fulfilment of the requirements for
the Degree of Doctor of Philosophy)

CONTENTS

I. The Problem 106

II. Work Previously Done 106

III. Available Methods 106

A. Transpiration 106

B. Water Content 107

IV. Preliminary Problem 108

A. Location 108

B. The Problem 108

C. Apparatus 108

D. Detailed Method 108

E. Data and Results HO

F. Conclusions 115

V. Present Problem 115

A. Location and Ecological Conditions 115

B. Variation and Description of Apparatus 117

C. Variation and Description of Methods 117

D. Data and Results 120

E. Conclusions .•. 129

F. Practical Application 129

0> VI. Summary 129

,rj>, A. Present Problem 130

^~ B. General Results 130

"^ C. Conclusions 130

^'^ VII. Acknowledgments 130i

106 CLARK— ON WATER CONTENT

I. The Problem

The problem is, to determine the water content and transpiration
in the leaves of Fagus americana, Hamamelis virginiana, and Quercus
alba during successive seasonal changes. Fagus americana, Hamamelis
virginiana, and Quercus alba are all native species of our Pennsylvania
hardwoods, and in this connection, have been studied, as far as possible,
under normal conditions in their natural habitat.

By "water content" is meant, the percentage of the whole fresh
leaf that is water; it includes both the water within the walls of the
leaf cells and that held in the intercellular spaces. "Transpiration"
is used in the sense of Burgerstein^ to mean the amount of moisture
given off in the vapor form from the surface of the leaf ; it includes both
epidermal and stomatal evaporation. The work started with the
earliest leaf development in the Spring of 1915, extended through the
Summer, and ended with the leaf -fall in the Autumn months.

II. Work Previously Done

Much work has been done along the lines of transpiration. It has
been studied extensively from morphological, physiological, ecological,
and purely physical standpoints, and detailed research has been carried
on showing the relation between transpiration and sap flow. While
Miinsch^ has shown that water content in plants is a determining fac-
tor in the appearance of fungous diseases, little connected work has
ever been done heretofore along the lines of transpiration and the water
content of leaves.

III. Available Methods
Transpiration

As pointed out by Burgerstein^ any of the following methods are
.available for determining the amount of water lost in transpiration:
a. Physical methods.

1. direct weighing. This consists of weighing specimen and
apparatus both before and after a certain transpiration
period.

2. collecting and weighing water vapor in a closed room.

^ Burgerstein, "Die Transpiration der Pflanzen," 1904, p. 3.

^Munsch, " Untersuchungen iiber Immunitat und Krankheitempfanglichkeit
»der Holzpflanzen. " Naturwiss. Zeitschr. f. Forst. u. Landw. 1909, 7:54-75, 87-114,
129-160.

^ Burgerstein, "Transpiration der Pflanzen," 1904, p. 4-28.

AND TRANSPIRATION OF TREES 107

3. self-registering apparatus made especially to determine
amount of transpiration.

4. calculating the amount of water absorbed by plants and
considering this the equivalent of the amount transpired.

b. Chemical methods.

1. determination of increase of water absorbed by such
substances as calcium chloride and concentrated sulphuric

acid.

2. color change in paper impregnated with palladium chloride
and cobalt chloride.

The method used was that of direct weighing, which, according to
Burgerstein is the most reliable.

Water Content

For determining water content the following methods are offered:

a. Physical methods. Any physical property which changes with
the water content offers a possible basis for experimentation.

1. electrical properties, for example, electrical resistance-

2. thermal properties, such as thermal conductivity and
capacity for heat.

3. mechanical properties, such as weight and tensile strength.

4. optical properties, such as coefficient of absorption and
reflection of radiation. It is known, for instance, that
water very strongly absorbs infra-red radiation. The
amount of absorption would probably indicate the amount
of water present in the leaf.

b. Chemical methods.

1. qualitative test for water. This test sensitive to less
than 0.1 mg. is easily and quickly made by bringing the
substance to be tested into contact with calcium carbide
in the presence of an acetylene solvent. This is then
decanted or distilled into an ammoniacal solution of
cuprous chloride.^

The method selected was the one which had been satisfactorily used
in preliminary experiments carried on in the summer of 1914 at the
University of Michigan Biological Station. It was based upon the
mechanical property of weight.

♦Journal Franklin Institute, March, 1916. p. 408.

108 CLARK— ON WATER CONTENT

IV. Preliminary Problem
Location

The Michigan Biological Station is situated on the south side of
Douglas Lake, about twenty miles south of Mackinac Island, in the
northern part of the Southern Peninsula of Michigan. Immediately
surrounding the camp is a xerophytic association of aspens; a mile and
a half east of camp lies a bog, in which are found the same species of
aspen in hydrophytic habitat.

The Problem

The problem in this preliminary work was to determine the water
content in the leaves of Populus tremuloides and Populus grandidentata
under both typically xerophytic and hydrophytic conditions for both
clear and cloudy weather. The topic thus naturally resolved itself
into the following heads:

1. Populus tremuloides = xerophytic = clear

2. Populus tremuloides = xerophytic = cloudy.

3. Populus tremuloides = hydrophytic = clear.

4. Populus tremuloides = hydrophytic = cloudy.

5. Populus grandidentata = xerophytic = clear.

6. Populus grandidentata = xerophytic = cloudy.

7. Populus grandidentata = hydrophytic = clear.
8; Populus grandidentata = hydrophytic = cloudy.

Apparatus
The apparatus used was in all cases simple and unoriginal. It
consisted of 144 homeopathic vials with paper labels and tight-fitting
corks; chemical balances, accurate to .5 of 1%; a gas oven; a Clements'
photometer, loaded with Solio paper; and an egg-beater psychrometer,
fitted with wet and dry bulb Fahrenheit thermometers.

Detailed Method
For purposes of curtailing the work, the first and fifth, second and
sixth, third and seventh, and fourth and eighth divisions of the problem
were carried through simultaneously. The work thus resolved itself
into four different experiments, readings for which were taken every
hour for 24 consecutive hours. As far as possible, leaves of apparently
the same size were selected from the same trees. Experiments for
similar conditions were not duplicated and because of weather changes,
readings for xerophytic cloudy conditions were got for 12 hours only.

AND TRANSPIR-'VTION OF TREES 109

The three series of 48 bottles, labelled and tightly corked, were
marked in ink on both label and cork: It, Ig, 2t, 2g, etc., and placed in
order in a shallow box of convenient size. The number in each case
stood for the reading of the experiment, and the letter indicated the
species of leaf, whether tremuloides or grandidentata. This system of
labelling not only indicated the species of leaf, but proved itself most
valuable in associating any cork always with its proper bottle, a device
to reduce the number of weighings. Each bottle with cork was then
weighed accurately, registered according to label, and arranged in series
according to number. Three such series were kept in readiness, each
designated by a special mark (check, comma, cross) to avoid any pos-
sible cork or bottle confusion.

When the experiment was started, the first reading was taken by
removing three to six leaves (depending upon size) from Populus tre-
muloides, rolling them quickly, and corking them tightly in the bottle
marked It. Immediately the psychrometer was placed vertically with
its bulbs in the position formerly held by the leaves pulled off. The wet
and dry readings were tabulated. The photometer was placed in a
similar position and exposed to the light. The number of seconds re-
quired for suitable coloration was recorded together with the number
of the disk exposed. A similar photometric reading was taken immedi-
ately in the open, and the whole process repeated for Populus grandi-
dentata.

After the readings were completed, the corked bottles with their
contained leaf rolls were weighed. The object of the tight cork was to
prevent the escape of evaporated water before weighing. All results
were tabulated and the corks were then removed and arranged in series
for future convenience. The 48 corkless bottles with their leafy con-
tents were placed in shallow boxes and put in the gas oven. In order
to prevent carbonization of the leaf tissues, the oven was regulated so
as never to exceed 100° C. They were kept in the oven until there
was no loss in weight, and this required on the average about 24 hours.
Upon removal from the oven, the bottles were again fitted with their
respective corks, and another weighing was taken, this time of bottle,
cork, and dried leaves within.

The following data were thus obtained:

Wl = weight of labelled bottle+ labelled cork

W2 = weight of labelled bottle-j- labelled cork -f fresh leaves

W3 = weight of labelled bottle-f labelled cork-f dried leaves

1 1 0 CLARK— ON WATER CONTENT

W2 - Wl = weight of fresh leaves

W3 - Wl = weight of dried leaves

(W2 - Wl) - (W3 - Wl) = weight of water in fresh leaves

W2 - W3 = weight of water in fresh leaves

W2-W3

r— — ZZ7- X 100 = water content, based on weight of the fresh leaves.

Wz - W 1

For every hour of the 24 consecutive hours, water content for both
Populus tremuloides and Populus grqudidentata, based on the weight
of the fresh leaves, was reckoned.

Due to lack of anemometers at the Station, only general observa-
tions were made as to air movement; no data were got for air pressure;
measurements of water content of soil were eliminated; and general
weather conditions were roughly summed up as cloudy, clear, rain, etc.
Percentages of relative humidity were obtained from the "Marvin
Psychometric Tables." Percentages of light intensity were calculated
according to the following formula:

— = light intensity, where
bx

x = number of seconds required to produce a good tint on the

Solio paper in the shade to be tested.
y= number of seconds required to produce a good tint in the

open at the same time,
a = number of seconds required to produce on the standard a

tint to match x.
b = number of seconds required to produce on the standard a

tint to match y.

The standard was made by exposing the Solio strip for 2, 4, 6, 8,
etc., seconds at noon on a bright, sunny day, July 4. All Solio strips
were kept carefully in a dark, dry place and were later developed in a
solution consisting of a teaspoonful of Hypo dissolved in 4 ounces of
water.

Since these figures in light intensity show only the percentage of all
available light that is received at each particular reading, the results
are practically valueless so far as daily or seasonal consecutive work
in water content and transpiration values are concerned.

Data and Results

From readings of the various phases of the experiments, the follow-
ing tables were compiled:

AND TRANSPIRATION OF TREES

111

<

1

a

33BJ3AV

60.1

65.2
62.9

59.2
61.7

On

1

^
p^^^

uinuiiu!i\i

lO lO

00

auitx

A. M.

5:30

Midnt.
12
A. M.

P. M.

10:30

A. M.

1:30

Midnt.
12

uinuiixBj^;

\0 \0 O

Tt< NO
NO NO

NO

1

a

<

00

38BJ3AV

lO O) tj; ^_

ON t--i u-j O
uo lO O vO

rv) ■^ NO

od 00 00
IT) in *n

00

3UIIX

^ U-. CrH *-i »-< O

pH '-H HH E-H l-H

<:::<2p;^

uinujiuiiv

IT) -O o ^

Tji iW fN Tt

IT) irj o o

NO lO

t^ CN rjJ
IT) lO "^

NO

I

(See Fig

3UIIX

< p^ < < <

It-' lO '«ri *-l lir< O >rH O
^ ^ r^ le^ ,t^ (^O ^ rO

1-) z

h o
U
Pi

H

uiniuirEj^

I~-. o ■^ o

O 00 C> C;
vO lO O O

r^ rvi NO

C^ O '-H

1/-; NO NO

NO

i

J!

-r

33BJ3AV

lo o o

00 o

lO NO

1-H

3UIIX

^^ ^ ^ S

p:=^ p."*

lunuiiuij^

00

3U1JX

P.M.
10:30

A.M.

A.M.

8:30

P. M.

10:30

A. M.

1:30

Midnt.
12

uinuiixBj^

^ so
u-i On Cn

(^ NO NO

TjH NO

■^ lO

NO NO

fO

NO

1
1

Clear
Cloudy

Clear
Cloudy

rt 3 nS
<U O m

u G u

>N

-?

O

U

X X X K

XXX

>, bb bi) ^ bi) *P
3 Ph' 3 pLi ^ Ph'

s _^ _^ .^ .

^N*-* bb -^^ bi)*"! i*^*^

^PhSP^^Ph^P-i

H-^ <; < 1— >

o

>,bX)

112

CLARK— ON WATER CONTENT

bc

pa -5

H ■-'
&

H

<

H

«

2

3
O

31U1X

uinmiuTj^

lO 1^ 00 ^ r^ CO

3UIIX

ph-2 ^"'^^ ph':^ <^<^

umuiix^j^

-H O 'f 00 o -^
00 O 0\ r^ O On

s
pi

to

1

•a
<

90

U35(BJ 10^

aiuix

P.M.

2:15

P. M.

3
Noon
12
P. M.

4:30

P.M.

2:15
P. M.

2
Noon
12
P. M.

4:30

uinuiiuij^

3«i!X

^ ro ^ S ^ ro § .—( S ^ S fO

a
<

in

1

pi

00

n35[Bl 10>J

3«I]X

p. M.

2:15

Noon
12

P. M.
4:30

P. M.

2:15

Noon
12

P.M.
4:30

inntornijAf

5 ^ lO 'i' 5 "D

3«i!X

uinunrBjif

O O rj< OO O •^
OO O 0\ t^ O O*

■,-1 y-<

Conditions

Clear
Cloudy
Clear
Cloudy

Clear
Cloudy
Clear
Cloudy

S -^ o s;, '^ .'^ .^ ."^ .
>,*; bc*j 55i,*J >,*i >> tb (i ti bb W)^ t^

AND TRANSPIRATION OF TREES

113

bfi

(IH

<u

»— (

1)

C/2

t-H

^

W

><

H

i-J

m

a

<

a

H

p

ffi

^

»-H

H

<

1-1

U

P^

so

aunx

uinuituTj^

3UIIX

s s
&:^(i:°°

in

00

O
O

ON

00

a:2

<"'<"^

<^<"^

uinuiirBj^

3UIIX

uinuiiuiiv

00

o
o

00

o

o

br-< lO >M C >Br< O

^ -- ^ O f< <r3

. c^ . po -5 r^ - . Tj"

icH "1 krH

Oh PL|

C
O

I o
1 rn

I "^ f^ «■ ■

On

o

^ lO

auiix

wnniixBi^

annx

uinunuixv

15 O «^ "^ k-O
f^rn i< f^ i< <r)

Si2S

<5

<5 ro

..^•^•OO-.C^

o
o

10

o

00

o
o

10

o

00

S ro

ON

5 10

suiix

uinunx-Ej«i

S ^
<"=<"'

^ f^

&^i

S ^
<"^<"'

00

o
o

On

00

O .*

O On

a
o

o
U

>>

t_l

-d

t-l

:S

3

rt

u

0

(U

rj

u

U

C

u

-c

OJ

0)

X

X

K

3
O

3
O

3
O

CJ U U U

X X

X X

-0 T3

July 16

P.t.
Aug. 17

P.t.

NO

bi

3

P.t.
July 30
P.t.

July 16

P.g.
Aug. 17

P.g.

NO

bh

3

P.g.
July 30

P.g.

114

CLARK— ON WATER CONTENT
TABLE IV (Results)

Conditions

Results

Proportion

Difference (%)

Xer. clear
"Xer. cloudy
hyd. clear
hyd. cloudy

P.t. > P.g.
P.g.>P.t.
P.t.>P.g.
P.t.>P.g.

20:4

9:0

24:1

24:1

1.2
1.2
5.3
3.7

Species

IP.t. (day)

P.t. (night)

P.t. (day)

P.t.

P.t.

P.g. (day)

P.g. (night)

P.g. (day)

Pg.

Pg.

Xer. clear > xer. cloudy
Hyd. clear > hyd. cloudy
Hyd. cloudy > hyd. clear
Hyd. clear > xer. clear
Hyd. cloudy > xer. cloudy
Xer. cloudy > xer. clear
Hyd. clear > hyd. cloudy
Hyd. cloudy > hyd. clear
Hyd. clear > xer. clear
Hyd. cloudy > xer. cloudy

8:1

7:2

10:5

23:1

9:0

6:3

6:2

14:2

17:5

9:0

2.3
2.2

.4
5.5
9.3

.2
1.1
2.1
1.4
2.5

It is seen from Table I that there is no definite time for maxima
and minima water contents. Fig. 1 shows however that, as a general
rule, in spite of numerous fluctuations, there is a gradual increase in
water content by night, and a gradual decrease by day. It shows
likewise that the curves of the two species of Populus, when compared
under similar conditions of habitat and atmosphere, follow each other
rather closely. This fact indicates the influence of a common factor.
By comparing Tables I, II, III and Figures I, II, it is evident that water
content is independent of both temperature and relative humidity.

For both Populus tremuloides and Populus grandidentata under
similar atmospheric conditions, water content is greater under hydrophy-
tic than under xerophytic habitat. Under xerophytic cloudy conditions,
day readings only were obtained, and with this one exception, water
content for Populus tremuloides was found to be greater than for Populus
grandidentata. Day readings under both xerophytic and hydrophytic
situations were found for both species to be greater under clear than under
cloudy conditions. One exception to this was for Populus tremuloides
under xerophytic conditions. By night for both species, under hydro-
phytic situations, water content for clear was greater than for cloudy
conditions. Irregularities in general results occur only in P. tremuloides
under xerophytic conditions, and might probably be accounted for in
scarcity of moisture available in the dry soil.

AND TRANSPIRATION OF TREES US

Conclusions

From above results the following conclusions may be drawn:

1. Average water content for Populus tremuloides is 61.8%.

2. Average water content for Populus grandidentata is 59.6%.

v3. There is no definite period for maxima and minima water content.

4. With fluctuations, there is a general increase in water content
by night, and a general decrease by day.

5. Under like conditions curves for both species follow each other
more or less closely, indicating a common influential factor.

6. Water content is independent of both temperature and relative
humidity.

7. Under like atmospheric conditions, for both species, water
content under hydrophytic is greater than under xerophytic
conditions.

8. With the exception of xerophytic cloudy results, where day
readings only were obtained, Populus tremuloides water content
was greater than that for Populus grandidentata.

9. With the exception of Populus tremuloides under xerophytic
situations by day, water content for cloudy conditions was
greater than that for clear.

10. By night for both species, water content for hydrophytic clear
was greater than for hydrophytic cloudy conditions.

11. Irregularities are prominent in Populus tremuloides under xero-
phytic conditions.

V. The Present Problem
Location and Ecological Conditions

While the main part of the problem at hand was to determine the
water content in leaves of Fagus americana, Hamamelis virginiana,
and Quercus alba during the seasonal changes, parallel experiments
regarding the amount of transpiration for these leaves were run simul-
taneously.

The work was done on a Brandywine farm located 6 miles southwest
of West Chester, Pa., and 30 miles west of Philadelphia. The experi-
mental field was confined to the northwest corner of a twenty acre
woodlot, whose timber had been removed for the first time, 25 years
ago in lumbering processes. Here is now a splendid second growth of
hardwoods with a few towering relics of the original stand.

The plot selected for the experimental work was irregular in outline:
its northern boundary line was a twelve foot cart-road leading into the

5)
J)

116 CLARK— ON WATER CONTENT

woods, its southern one was a shallow stream 10 feet in width, fed by a
spring 300 feet farther east; its western boundary was a fence, dividing
the woods from a broad, open pasture; its eastern border was an ima-
ginary line about 20 feet long running between the cart-road and the
stream. The plot thus irregular, had its greatest length of 70 feet, and
its greatest width of 30.

During the previous summer (1914) a bhghted chestnut tree had
been removed from the western end of the plot, leaving a large, solid,
two-foot stump. The removal process had resulted in the destruction
of most of the secondary growth of the plot, as well as of many young
beech saplings. In the plot were found the following species:

(1) Trees:

Fagus americana 15 individuals

Quercus alba 13

Cornus florida 7

Hamamelis virginiana 7

Liriodendron tulipifera.... 6

Castanea dentata 4

Acer rubrum 4

Prunus serotina 3

Fraxinus americana 1

Nyssa sylvatica (60 it.).... 1 "

(2) Shrubs (few and scattered:)
Vaccinium stamineum

Rhus toxicodendron
Viburnum acerifolium
Diervilla lonicera

(3) Herbs (few and scattered :)
Lysimachia punctata
Polygonatum biflorum
Impatiens pallida

The forest floor was covered with leaf-mold to a depth of 10-15
inches. During the whole season there was no scarcity of water supply,
for not only did the stream run on two sides of the plot, but the season
was one of exceptionally great and frequent rainfall, as the following
table shows :^

«U. S. Department of Agriculture, Weather Bureau, Monthly Meteorlogical
Summary, 1915, Philadelphia, Pa.

AND TRANSPIRATION OF TREES
TABLE V

117

Precipitation

Month

Total

N0RB1.\L

No. Clear days

'. ilay

June

4.12
3.45
5.02
6.84
0.46
1.98

3.20
3.30
4.33
4.61
3.38
3.10

10
8

ulv

7

August

6

Seotember

12

October

14

The average level of the forest floor above water line was 3 ft. 2. in.

Variation and Description of Apparatus
As in preHminary experiments, the apparatus used for getting water
content was plain and simple. The psychrometer was home-made-
constructed from a Dover egg-beater: tin clamps were attached to
hold the thermometers firmly in place, and certain unnecessary wires
and plates were removed to avoid interference. A spirit-level was in
constant use in adjusting the chemical balances and the improvised
table on which they rested. The Clements' photometer with Solio
strips was again used in light tests The homeopathic vials were re-
placed by plain, business-sized envelopes. For drying purposes, the
warming oven of the kitchen range in the farm-house proved adequate
and satisfactory. For transpiration tests three homeopathic vials,
vaseline, medicated cotton, a pint tin cup, and shears were provided.

Variation and Description of Methods
A heavy, well-seasoned, 2-inch plank, 2^^ feet long by 2 ft. wide,
was placed, leveled, and permanently fastened on the above mentioned
chestnut stump. On this was placed on end and leveled, a small, heavy
dry goods box which sheltered the balances. The box was completely
covered with a thick, woolen blanket, over which was spread a pliable
oil-cloth. These coverings served to protect the scales from air currents
during weighing processes.

The leaves were pulled from the tree, were weighed immediately,
and were placed in a series of dated envelopes marked 8B, 8W, 80,
9B, 9W 90, etc. The label indicated the hour of the day and the kind
of leaf— beech, witchhazel, or oak. After each weighing, psychrometric
and photometric readings were taken, and general weather conditions
noted. Such readings were made hourly from 8 a.m. to 5 p.m. either
weekly or semi-seekly from the first of May to the middle of October.

118 CLARK— ON WATER CONTENT

At the end of each day's work, a paste board box containing the
various envelopes with their leaves was placed in the warming oven.
No constant temperature could be maintained here. A wood fire was
made up 3 times daily for getting meals, and often, depending upon
household needs, it was kept burning for the greater part of the day.
At its hottest, the oven registered 118° F. After a week the leaves were
taken from the oven, removed from their respective envelopes, and
weighed. From the fresh and dry weights, the water content, based
upon the fresh weight, was calculated. (See p. 112.)

For transpiration tests, attached twig ends were bent about a foot
from the tip into a pint cup filled with water. Since wetting of leaves,
according to Haberlandt,*' Wiesner,^ and Burgerstein,* hastens tran-
spiration rate, great care was taken that the leaves did not come in con-
tact with the water. Following the method of De Vries,^ the twigs
were cut off under water, and then without permitting air to come in
contact with the cut surface, they were transferred to homeopathic
vials containing water.

The vials were about three-fourths full of water, and a cork was made
by wrapping cotton closely around the stem of the twig and pushing it
tightly into the mouth of the bottle. The cotton cork was covered
with a thick layer of vaseHne to exclude all air, and the vaseline was
covered with a thin layer of cotton to prevent the leaves above from
coming in contact with the vaseline. Normal transpiration would have
been interfered with had the leaf surface received a coat of vaseline.

These experiments were prepared by 7 a. m. and the vials containing
the twigs were placed always in the same order in the same shallow
wooden box, which was kept always in the same position on top of the
oil-cloth, covering the above mentioned dry-goods box. Thus placed,
the twigs were left for an hour to overcome the shock of separation from
the tree and to become adjusted to new conditions.

* Haherlandt, F. Das Austrocknen abgeschnittener und benetzter, sowie abge-
schnittener und nicht benetzter griiner Blatter und Pflanzenteile. (Wissensch. prakt.
Unters, auf dam Gebiete des Pflanzenbaues, herausg. von Fr. Haberlandt, Bd. II,
Wien 1877, p. 130.)

' Wiesner. Studien iiber das Welken von Bliiten und Laubsprossen. Ein
Beitrag zur Lehre von der Wasseraufnahme, Saftleitung und Transpiration der Pflan-
zen. (Ebenda, Bd. LXXXVI, 1882, p. 209.)

* Burgerslein, Die Transpiration der Pflanzen, 1904, p. 68.

' De Yries. Uber das Welken abgeschnittener Sposse. (Arb. d. Botan. Inst.
Wurzburg, Bd. I, Leipzig 1874, p. 287.) Burgerstein. Die Transpiration der
Pflanzen, 1904, p. 73.

AND TRANSPIRATION OF TREES 1 19

At eight o'clock the first weighings were made and tabulated. The
readings which were taken hourly, consisted of the weight of the bottle,
water, cork, and twig. At the end of the day the leaves were removed
from their twigs, placed in labelled envelopes, and dried in the warming
oven.

Wl = weight of bottle, cork, water, twig, at 8 a. m.

W2 = weight of bottle, cork, water, twig, at 9 a.m.

W3 = weight of dried leaves.

Wl-W2 = weight of water lost in transpiration between 8 and 9 a. m.

Wl - W2

X 100 = percentage of water lost in transpiration, based on

"^^ dry leaf weight, between 8-9 a. m.

Thus hourly water loss was calculated for beech, witchhazel, and
oak for each day's experiment.

Because the experiments for water content and for transpiration
were made at the same time, no separate data for light and atmospheric
conditions were got for transpiration. Light intensity percentages
proved valueless for successive daily and seasonal work. Slight error
in transpiration rate might have crept in for the early hours of the day
when the leaves in the morning had been covered with dew. This
occurred only once, August 9, and then the dew was carefully absorbed
by blotters before the experiments were started.

According to Haberlandt,^" and Klebahn," water is lost in transpira-
tion by the lenticels. The present method used did not take into
consideration the water transpired by stems.

Selection of foliage of the beech was limited in the early summer
by. the presence of numerous egg-cases of some unknown insect. These
egg-cases were found on the ventral side only of the leaf. Oak galls,
on the oak leaves, were evident, but not numerous. On the witch-
hazel were many galls, containing nymphs of Hormaphis hamamelidis^^
which were parasitic on the leaves. Within the galls were also unknown
larvae, which fed upon the nymphs, and later emerging, fed upon the
tissues of the leaf. As a result, the foliage of the witchhazel was greatly
damaged, and selection of leaves, in the late summer, was made most
difficult.

^^ Haberlatidt, G. Beitrage zur Kenntnis der Lenticellen. (Sitzb. d. K. Akad.
der Wissensch. in Wien. Bd. LXXII, 1875, p. 175.)

" Klebahn, tJber die Struktur und die Funktionen der Lenticellen, etc. (Ber.
d. Deutsch. Botan. Gesellsch. in Berlin, Bd. I, 1883, p. 113.)

^^ Pergande, Theodore. Technical Series, Bulletin, number 9, U. S. Bureau of
Entomology — 1901.

120

CLARK— ON WATER CONTENT

Ji

a,

B
o
u

m

I

I

a>

7i

CO

>H H-l

2

I/)

d

-a

(S3

s

O

i

<

>>

'B

aSBjSAV

1— (

"0

65.6

58.2

3"i!X

rf<

CN

CN

uinuiiuip\[

^

NO

00

aiuix

CK

CN

-H CM

On

UinUlK'BJ^

00

NO

=5;

NO

9;

NO

o

(-(

Oh

s

H

33BJ3AV

ID

d

ON

o

d

CO
00

NO

auiiX

oo

00

00

00

lunuiiuij^

in

NO

NO

§

3UIIX

CN

CD

•^ Tf ir>

*

uinuiixBj^

V2

1^

CO

■*

a

.2

"a

en

d

33-BjaAV

■rt_ UO -^

IT} ■*_ ro
■t-h' CN TH

CN CN NO

CO0C5

auiix

C^l CN 0\

lO uo ON

O ON UO

On "0 0^

uinuira!j\;

lO "0 CN

00 CN

f^tp

LO ■*_ CO

sui'i

O <N ^

CN ^ O

^—1

T-H »-H ^— (

uinuiixBj\r

CN tN ON
•^ CN

00 ■*
^ CO CN

t--. CN ro

>— 1 t^ NO
CN ^' ^

c

d
o
U

1

JU3JU03 J3}BA\ aSBJSAy
UIOJJ

uop-EUBA agBjaAy

rO t^ C^-I
vq 00 CN

CN-* rrj

MO O CO
Tt; CN NO

00 OO CN
CS CN T-H

■■-H ' CO

CO 00 NO
CN COOO

^ CN

33EI3AY

On 0(3 r^

^' irj d
oc f~ r^

ON CN -^

r<~ r^ irj
t^ r^ j-^

00 CN lO

oo' d o^

lO 1^ NO

u-)^ 00

CN NO LO

LO NO NO

3^!X

T-H (MC

rr;ir>ir2

CN CN CN

CO ■* t}<

uiniuiuij^

oq t--; On

t-~ lO NO

CN -^ 00_
CN lO CO

1^ r^ t^

CN rvj CN

NO On CO

u^ O UO

r~ On

>D NO NO

3"i!X

On CD CN

CN CN ro

OOiOO

lO 00 LO

uiniuixBi^

CN r- t---

O O CO

OO 00 00

CN 0C3 T^H
lO OC! NO

r^ t^ t^

CN 00

-H CO '*

nO t^ <^

■r}<t^ vq

•.-h' CN 0C>

vO r~ O

CO

>, pq^O

AND TRANSPIRATION OF TREES

121

0^

1^

sd

:2 1
1

sq
■>*

ID

lO

■*

fO

■*

»— t

PO "*

OO

s

"0

Os

ID

5

00

00

00

00

Os

S

00

O

so

ID

SO

•«*<
g

00

so

00

00

00

00

00

00

00

t

3

so

■*

»^ <M <~0 Tf lO

•>*"

»^fD

pDt*<

00

o

00

OS

00

00

so

^00

V— 1

SO rC

'^

rt< CN SO

so sO t^

u-j -^ lO

OS OMO

Tj< OS CN OS

rO OS

Os

ID ID ID

ID ID -H

NO t-; CN

t^ Tf lO

00

liOO;

J-. ro

t^

O; fsj T-<

O'-iO

*-H v-^ »-H

1— (

^H *-*

^^

^-H

»-C ^H »-l

■^OO

fO -^ ■^

O O 00

to

Os J-~

' csi

rsi psj

ID

fD tN 00
<>i <N CN

<si rt' -H

.8
2.44
1.03

<>) O; p
(TJ CN -rt'

Os

00 PC
sC ID

Os

csi

^^ rsi o

so o^p

PD Csl so
Os_ •*_ OS

■t— 1 ^-H

59.1
64.3
59.6

O lO ro

\0 SO sO

00

00

"0

OS OS

o o^

sOu->

00 <>J
ID O

SO

fD ID t^

oo'-ld

ID so so

tN PO ""I

so so so

■<* fO-H

PO Os ■*

O r^ Tt

— 1 fD

-H rD PC

ID OS PC

rs ■*_ -*
r-^ 00 t-~"

t--. (N lO

Tt OS -h'

lo lio SO

ID

IT) ID

SO r^
ro ID
ID ID

sO_
ID

t)< Ti; Os
r--.' 00 r^

ID ID ID

CN OS PD

ID OO' Os
ID ID ID

00 00 OS

00 00

00 >D

00

CN CN Os

tJ< (N O

fN "* CN

•-H 00 tN
\0 \0 vO

so Psi 00

so Os'^t
t^ so O

so

so

sd 00

so so

fDJ--;

So

sd

sO'^ 00
Os sd tN

ID sO SO

***. ^. ^

•^ PC •*'
so so so

3

o

3
1— >

3
1— >

^o

00

3

O

>s«^0

3

>— >

122

CLARK— ON WATER CONTENT

3SBJ3AY

■o

NO

fN

NO

3™!X

CN

fO

CN

1

lunuiiuij^

to

NO

lO

NO

3"J!X

00

lO

On

umuiixBj^

lO

00

00

O

o

■>->

c

3

a

33BJ3AV

On

o

00

aiuix

00

00

00

uinminip^

<NO

amjx

CN rr)

PO'*

ro

•T3

ranuiixBj^

00

NO

00

3

a

-«^

§

>

d

sSBjaAV

O oq On

i~~ l~-. Tl*

^—1 ^-H

^_p "O

auiix

On On C^

On lO CN

On -^ CN

mnmiuij^

Tt;-* ro

00 (^ Tj<

O t^ CN

^— t

3i"!X

CN '- fN

*— ( ^— 1 ^-H

^ -H ro

uinmix-Bj^

lO Tt On
Cvi -h' CN

t-~ fD t^
fS ^ CN

li-> vq On

c

CI

o
U

Ul

«

5U3JUOD jajBM 33BJ3AV
UIOJJ
UOIl-EUBA 33BJ3AV

r-l NO Tf
pr^ On

NO NO

PO t^ t^

1— ■* O

lO ro p

' CN i-i

3SBJ3AV

'~; ^ '^

O <T/ '— '

NO NO NO

ro "0 ■*

NO NO O

oq NO oc)

O fN) (-^

NO O "-)

3UIIX

UO CN lO

(^ ■rt ro

Tf< Tf* CN

uinui!nij\[

t~~ NO 00

00 t^ On
lO lO lO

t--.' -r^ On

lo NO lo

^ ■* 00

On' 00 iW

lO ir> lO

3UIIX

O ONOO

uinuiix-EjAl

lO 1~~ li")

«' r^ cn)

NO NO nO

r-. ri< On
tN 00 rvj

NO NO NO

t^ lO ro

NO NO NO

(
1

o

"3

1—,

00

3

>— >

AND TRANSPIRATION OF TREES

123"-

d

3

d
1^

in

NO

f^

•— t

ro -^

CNl

r^

CM-*

1^5

1

PNI

NO

in

00

OO

00

00

lO

00

g

o

On

o

ON

NO

00

oo

00

On

<*5

d

00

d

00

00

00

00

00 ON

00

00

ON

9^

NO

s

in

fO

fo

Ol ro rt

tN

^ <V| rO

rO

^

00

00

00

00

O \q 1--;

■^ *^ ^

J-~

O lO

Cn t^ NO

vqp

csoot-^

0\ 0\ o^

Cs 0\ 0\

lO

ir> lo

■*OnOn

CN

On lO

On On "^

Ti< rq (>q

re tN C^

-1

"^ '-;

NO lO Tj<

NO

fvlTJ-

in ro CN

*— ( ^-H *— <

i-H 1— 1 ^-H

^—t ^-H O

^^^

^H ^-H ,-1

T-H «-l cs

^— ) T— (

ii-> Ov fO

00 CN OO

♦-H *-H »-H

P*^ 0\ On

00 lO

00 TlH (>J

T-H T-H ^-H

Tj< 00 "*
(N 0\ oo

Ov "-I O
t^ t-- CN

NO to

p lO >-<

^ PO »^

On
CO

00 00

On t^ T-i
^ tN ^

00 '-;

00 r'l t-^
«0 vO lO

00 r-; -^_

00 On t^
uo lO lO

NO
ID

rNJ oo"

NO lO

f^ r'i On
^-I^^'nO

in NO lO

On

ON

in

^in

■^' NO

NO in

NO On NO

in lo lo

lO •* U-)

t}< -^ "0

o

00 ro

fNj -H LO

rt*

T-H

■njc^ 00
ir> \o m

t^ O^ LO

to T^ U-)

NO

CN NO
00 LO

in lo

^-. 00
in o T)^
in in lo

in

oqcNq

NO ro
LO m

in \0 -^
CN in' cnJ

LO lO LO

■^ 00 00

O 00 00

ON

fO ■^ 00

in

CM CN On

T-H

00_-H ro

»-< t-^ 00

\0 vO LO

fo r^ 00
\0 vO lO

O^

in

^_00_

no' "^

NO NO

NO 04 00

0^ '— ' On

in r^ lo

NO
NO

in 00
vD m

■*. '~'^. "*.
d to' oo'

NO NO LO

3

•— >

3

T-H

3
>— >

^o

<

ON

^O

T-H

<:

124

CLARK— ON WATER CONTENT

t3

0)
§

>

w

m
<

a3BJ3AV

0

0

10

suiix

ro

■^ r<)

P^ CO-*

XI

3

iunuiiu!J\[

•0

10

w ■
&5

3™!1

IT)

00

CO

uinuiix-B;^

00

9i

vO

00_

o

J3
C
0)
M

-4-)

a

H

a3BJ3AV

d d
00 1^

CO

31UIX

1

00

00

00

uinuiiuipM

3t"!X

r^ Tti

•^ rq f^ Tf

(N re*

uinuiixBj^

00

00

.2
a

'a

C

a

3SBJ3AV

vo r^ t--;

t--. 00

<r)

00 1--;

3iu!X

OMO 10

«0

0 CN

10

10 10

uinuiiuiH

Ch r^l -H

"1

'*.'*.

^.

^. ^.

8UI1X

CN »-< CN

tN 0

1— 1 T-H

uinuiix'E]^

^ ^.00

1

OS fO

*->

a

CJ

§

t-l

;U3}U03 J3}-BA\ 32BJ3AV
UlOiJ
UOpBU-BA 33BJ3AV

r/^ ro 0
CN ro ■^'

CO

(>1 ^

y—t

0^ 10

3S13J3AV

oi 0 <^

06 00 10

10 10 to

0
10

CN to

06 ro

10 10

0

d

-0

1

O; "0
CN t-^

vO"0

8U1!1

—1 Tl"*

in ■<+

CN On

00

0 <M

iunuiiuii\[

■*. ^ "^.

T*' CN --^
10 ir-, 10

10

ro On
•0 ^

00'
10

\0 CN
10 10

auiiX

»-i 00 Ov

00

Tt — 1

10

rOTf

uinuiix-Ej\[

cn' 10 c^

■o 0 uo

00

10

00 0
00 0

•oio

00 10

vOO

<

0

<

^0

rO

3
<

^0

AND TRANSPIRATION OF TREES

125

o

o

00

OC

o

<5

CN r<i ■>*

O

re O ^^

Ov lO vo

uo lO >0

CO t-~ <M

ro ■^ fO

O O^ t^

CO vO lO

CN rj< 00

o lo ^— '

00 0\ t^
ID lO lO

r^ r/^ <r3

r<i o '— I
10 10"^

Ov

1-H

CO

o

o

o

o

\0

o

00

o

re

00

o
oo

rn
00

o

00

00

CO

oo

LO •--< p <^. "^^ '^

OS 0\ Ov

tJ< C^ '^ ^ <^. "^

re

ir5 OTO

0\ 0\ 0^

lO re CN '*. '^ ^

'"' "1 "1

<N T-i '-I

^ o o

O 00 re

CN ■* 00
O ■^ fN

t^ r-J O

IT) O "0

lO r^ tN

tN re to

Tt VO lO

OS Os lO

CN CN CN

ON On

oo 00 r^
t^ lO O

CO fN ""'

irj nc lo

re 0\ re

Tt< CM P<1

00 -^ •^.

liO lO CM
IT) IT) iTi

CM CS 00

On CO On
O NO NO

ID -* -*

T* CM CN

CM On 00_

NO On re
lO lO lO

CM -H 00

t^ O ■*

lo NO lo

T^ On !>;

oo' O ^

CM CM CM

CM On

ID On •— '

NO O) On
ID ID ■^

Tt 0C3 •*.
Tt r^ '-i

ID •D ID

■^ re

•^ NO O
NO NO NO

.-H CM CM 00

ID -* NO
ID NO ID

ID re CM

ID On CM

■^ CM NO

NO

ID

t- o

ID ID

0\ CO On i— I

NO NO
CO ^ O

ID NO NO

NO
CM

to

m^o

C/3

m^o

ipq^O

t-~ re 00

O NO 00
\0 NO ID

00 CM nO_

00 re NO
ID nO "D

00

ID

CM

en

;m^O ^w^O

C/2

On
ID

On On
CM t^
NO ID

'^M ^O

o

126

CLARK— ON WATER CONTENT

Humidity

a3BJ3AV

3'^JJ.

^— (

P^ r^ Tl<

mnuiiuij^

"0

On

3i"!X

o

00

luniuixcj^

00
00

o '

3
■♦->

a

s

H

aSwsAY

auijx

00

00

mniuiuij^

O

Th

00

in

3i"!X

^-H

ro

1

uinuiixBj^

fO

lo

O

1

•4->

2
'a

tn

C

H

sSBjaAV

t~~.ir)in

ro

T-H

O)

SI"!!

O^ Os 0^

OS

OS

Os

>

w

pa

uinuiiuij^

ro ro (^

q

-

^-H

3"^!X

rt r/:) ^

fO

•*

PO

ainuiixBj^

On t-- r-;

^

CsJ

rs

c

c
o
U

;u3;uo3 J3J13M aSBjaAV

UIOJ}
UOp-BUBA aSBJSAV

Ov O; PO

s;

33BJ3AV

o ^ ^d

vO O lO

00
OS

SO

"1

ai"!!

,-1 T-H rri

»— '

CO CN lO

uinuiiuij^

lo \q ov

l-~! ,-H ro
ir> o >0

SO
ID

OS

fO

3U1IX

■rt 00 CN

▼— t

tN

OS

mnuiix'Bi^

^ 00 c^

O O U^y

Os_

(vi

OS

oo'
>o

sq

00

o

O

^

o

AND TRANSPIRATION OF TREES 127

For daily readings the maxima and minima values of water content
occur at no definite periods. As a general rule the two minima for
transpiration occur early and late in the day, and the maximum at
1 p. M. The minimum temperature is at 8 a. m. and the maximum at
2-3 p. M. Maximum percentage for relative humidity is at 8 A. m. and
minimum at 2-3 p. m. Weather conditions are on the whole regular
and constant. While daily variations in transpiration, temperature,
and relative humidity are fairly constant, we find no connection between
these and water content. Water content therefore must be indepen-
dent of the influence of transpiration, temperature and relative humidity.

According to Dixon^' the loss of water due to evaporation "can only
be made good by drawing in water from adjacent tracheae, and this
pull acting on the upper ends of the cohering columns of sap is propo-
gated downward through the tree. We may then regard secretion or
evaporation as the force which actually exerts the tension on the sap,
and this tension is transmitted through the leaf cells to the sap in the
conducting tracts. " Also the same author states that " the amount of
transpiration falls off as water is exhausted from the soil." It would
seem therefore that so long as the available soil moisture remained
constant, there would be no change in the water content of the leaves.

Because of the presence of a stream flowing on two sides of the plot
and because of unusually great seasonal rainfall, there could have been
no scarcity of available soil moisture, and daily water content should
have been constant. Reference to Table VI shows that the average
variation from the average water content for each day lies between
.4 of 1% for Fagus americana (May 9) and 4.87% for Hamamelis vir-
giniana (May 2). For the whole season the average variation from the
average water content was 1.5% for Fagus americana, 2.4% for Hamame-
lis virginiana, and 1.5% for Quercus alba. These differences run so
low that on the grounds of probable error they may be disregarded.
We may consider that from 8 a. m. to 5 p. m. the water content for
Fagus americana, Hamamelis virginiana, and Quercus alba is constant.

In the final results of the preliminary experiment it was found (Table
rV) that the difference in average water content between the species
of Fopulus under various conditions was generally negligible. Under
hydrophytic conditions, both cloudy and clear, the water content for
Fopulus tremuloides was always greater than for Fopulus grandidentata
(5.3% and 3.7% respectively.) Here there is probably the presence of

'^^ Dixon, H. Transpiration and the Ascent of Sap in Plants, 1914, p. 118-140.

128 CLARK— ON WATER CONTENT

a species difference. The water content for Populus tremuloides was
always greater under hydrophytic clear and cloudy than under xerophy-
tic clear and cloudy situations (5.3% and 9.3% respectively.) This
difference may be accounted for on the grounds that under xerophytic
conditions there is probably a scarcity of available soil moisture. The
amount of moisture lost in transpiration would exceed that taken in by
absorption, and the water content would decrease.

Average transpiration was \.2\%, .83%o, and .43% for Fagus ameri-
cana, Hamamelis virginiana, and Quercus alba respectively. These
results agree with those of von Hohnel/* who stated that transpira-
tion in beech leaves was greater than in oak. Rate of transpiration
in all three species was highest during the developing season in May,
and early June. This confirms Holtermann's^^ work on Paradenyia.
Wiesner^'' found that falling leaves transpired more rapidly than those
that fell later. Fig. XXXIII shows that in Fagus americana, Hamame-
lis virginiana, and Quercus alba, on the contrary, transpiration decreases
steadily from September 1st to leaf-fall. Water content is in no way
influenced by the factors of transpiration, temperature, or relative
humidity.

Average water content for the season was 59.9%, 63.1%, and 59.2%
for Fagus americana, Hamamelis virginiana, and Quercus alba respec-
tively. Fig. XXXIII for seasonal variation shows that for Fagus ameri-
cana, Hamamelis virginiana, and Quercus alba, during the developing
period in May and early June, there is a steady decrease in water con-
tent. With fluctuations the water content is constant through June.
During July and August there is a slight gradual decrease, and from
September 1st to leaf-fall, there is a gradual rise in water content.
Benedict^'' finds that in Vitis vulpina L. and certain other plants senile
changes are evident. Seasonal variation in water content in Fagus
americana, Hamamelis virginiana, and Quercus alba is probably due to

i^ow Hdknel, F. v., tJoer das Welken abgeschnittener Sprosse." (Wissensch.
prakt. Unters. auf dem Gebiete des Pflanzenbaues, herausg. v. Fr. Haberlandt, Bd.
II, Wien 1877, p. 120.)

" HoUermann, K., Anatomisch- physiologische Untersuchungen in den Tropen.
Die Transpiration der Pflanzen in den Tropen. (Sitzb. d. kgl. preuss. Akad. d.
Wissensch. Berlin, Bd. XXX, 1902, p. 656.)

^^Wiesner, J., Untersuchungen uber die herbsthche Entlaubung der Holzge-
wachse. (Sitzb. d. K. Akad. der Wissensch. Wien, Bd. LXIV, 1871, p. 461.)

" Benedict, Senile Changes in Leaves of Vitis Vulpina L. and Certain Other
Plants. Cornell University Agricultural Experiment Station. 1915. p. 281.

AND TRANSPIRATION OF TREES 129

structural changes in the leaves during successive periods of develop-
ment, maturity, and senility.

Conclusions
From above results the following conclusions may be drawn:

1) Daily maxima and minima values of water content occur at no
definite periods.

2) Dailv results for transpiration, temperature, and relative humidi-
ty follow more or less definite courses.

3) Water content is independent of the influences of transpiration,
temperature, and relative humidity.

4) From 8 a. m. to 5 p. m. water content is constant. This is pro-
bably due to the fact that there was no scarcity in available soil-moisture.

5) Average transpiration for the season was 1.21%, .83%, and .43%
for Fagus americana, Hamamelis virginiana, and Quercus alba respec-
tively.

6) Rate of transpiration was highest during the early developing-
season, and lowest at time of leaf-fall.

7) Average seasonal water content was 59.9%, 63.1%, and 59.2%
for Fagus americana, Hamamelis virginiana, and Quercus alba respec-
tively.

8) Water content was highest during the early developing-season,
decreased gradually through May and early June, remained more or
less constant through middle and late June, decreased slightly through
July and August, and increased steadily from September 1st to leaf-fall.

9) Seasonal variations in water content are probably due to seasonal
structural changes.

Practical Applications

Miinsch's^^ work has shown that the water content is a determining
factor in the appearance of fungous diseases in plants. While his work
dealt with the water content of the stem, it does not seem improbable
that the water content of certain leaves might sometime prove valuable
in determining conditions favorable or unfavorable for parasitic fun-
gous growths.

VI. Summary

The object of the present work was to determine seasonal variations
in water content and transpiration of leaves of Fagus americana, Ha-
mamelis virginiana, and Quercus alba.

^^Miinsch, ' Untersuchungen uber Immunitat und Krankheitempfanglichkeit
der Hokpflanzen.' Naturwiss Zeitschr. f. Forst. u. Landw. 1909. 7:54-75, 87-
114, 129-160.

130 CLARK— ON WATER CONTENT

Results show (1st) that there is no connection between water content
and transpiration, temperature, and relative humidity; (2nd) that from
8 A. M. to 5 p. M. there is practically no variation in water content, but
that variations are regular and constant for transpiration; (3rd) average
water content was greater for Hamamelis virginiana, than for Fagus
americana and Quercus alba; (4th) average transpiration was greatest
for Fagus americana and least for Quercus alba; (5th) water content is
highest in the Spring, lowers during the Summer, and rises again in the
Fall; (6th) transpiration is greatest in the Spring and lowest in the Fall.

A general conclusion of results may be stated as follows:

1) Under the same conditions, water content and transpiration
differ in different species.

2) Water content is independent of transpiration, temperature, and
relative humidity.

3) Water content is constant from 8 a. m. to 5 p. m. Large supply
of soil-moisture may be an influential factor here.

4) Water content varies during seasonal changes. This may be
accounted for by structural differences in the leaves from the stages of
early development, to those of senility.

VII. Acknowledgments
Great pleasure is taken in expressing to Dr. H. A. Gleason of the
University of Michigan, my appreciation of his able and helpful gui-
dance while engaged in the preliminary experiments at the University
Biological Station. To Dr. John W. Harshberger of the University of
Pennsylvania I owe greatest thanks for his generous and profitable
assistance along various lines of the present problem. For construc-
tion of simple apparatus and for valuable suggestions I feel deeply
indebted to Dr. Enoch Karrer of the Physical Laboratory of the United
Gas Improvement Company of Philadelphia. To Mr. and Mrs. WilUam
H. Clark of Lenape, Pa. I wish to express my gratitude for the privileges
and hospitality of their Brandywine farm during the experimental
phases of the subject. My appreciation is also extended to Miss Mabel
M. Thackara of the Commercial Course of the Germantown High School,
whose valuable manipulation in clerical details has brought the present
work to its close.

Philadelphia, Pa. April 6, 1916.

AND TRANSPIRATION OF TREES 131

Explanation of Figures
Fig. I.

Water Content of Popidiis trcmuloides and Popuhis grandidentata.

(1) xerophytic, clear and cloudy

(2) hydrophytic, clear and cloudy

Fig. II.

Temperature and Relative Humidity for Popuhis tremuloides and Populus

grandidentata.

(1) xerophytic, clear and cloudy

(2) hydrophytic, clear and cloudy

Fig. III-XXXII.

Daily Water Content, Transpiration, Temperature and Relative Humidity.

Fig. XXXIII.

Average Daily Water Content, Transpiration, Temperature, and Relative Humid-
ity, Showing Seasonal Variation.

B = Water Content of Fagus americana.

W = Water Content of Hamamelis virginiana.
0 = Water Content of Quercus alba.
b = Transpiration of Fagus americana.
w = Transpiration of Hamamelis virginiana.
o = Transpiration of Quercus alba.
t= Temperature,
r.h. = Relative humidity.
In Fig. III-XXXII transpiration values are multiplied by 10, and increased
by 50.

In Fig. XXXIII transpiration values are multiplied by 10 and increased by 70.

132

CLARK— ON WATER CONTENT

70

40

SO

Ft. w.c. xercUa-T

Rt o ClooJv

Bt X •• •• KyJcleAY

Rt O CJouiy

a:

V

\/ Ro wc. xerclea-r

y T?g o VVC. ..Cloudy

Kg X ■• •• hyJ.c)e<i.r

Ps Q cJoody

7P«. 3 'I /A-rt. 3

(J / fiM- ^^

/OO ^ 's*

'^ A P.* '^teJn|^.X.e ?f(?t . t«mls.->fe.r,cleik-r

; ' ?t

r.clou<i|r
hyd- Clear

rhom

A,

' \

g/ \ ng — K ..hfjc ]?t — X

/\ T — \ P.9 o •■ hy«.Cl«./ \\n — o

\

7Pn. 3 " / A.in. i

F19. OJ Me^yA

II A.MIZ

J RM.

AND TRANSPIRATION OF TREES

133

6 PM

<to

Fig.ir MAyj23.

// Anv2

?5.

§.30 M^k-xj;.

3 10 II A.n <a

2 3

5PH

134

CLARK— OX WATER CONTENT

90"

^0.

FkS.^ZH JUNE iSi
\r.h- ^

II Hn a

S Pti.

<?>o.

GJ. 2IU JUNEAO.

w Kn ' /2

6Pr^

90

Flg.H. d UN E25.

i\ A.M. n

4 p.n

AND TRANSPIRATION OF TREES

135

Flg.X dUNL 5,<r.

SPfX

30_

Fi

g-XIdULY I

3 /o II t\.n 12 I

z 3

ip.fi.

§.m juuy fo^

«-ii ■ I » I > I » I » <»^

136

CLARK— ON WATER CONTENT

90^

FlgXLD JULV 10

90

F,

i§ xiE JULV m

4-4-4 .Ml I I I ■

^

-I ■_

10 II A t\. a

-J -I-

J I

7 bTh

■Vi.

Fl§.55Z AUG 15

>lll/)l|ll'H^

AND TRANSPIRATION OF TREES

137

90

Fig XXa AUG.Ifo

ti+m+'*^^^+^^''^'^^<

'^J

10 n AM n

«y.

pigyvn AUG-. 17.

3 PM

©0^

FlgXEED AUG-.2.5-

»ii II I n-iliinii'/-H

9 10 II h-M II I

h ^ 5 PM

138

3a

CLARK— ON WATER CONTENT

Fig.XIX AUG.:i<»

Vl-11 IlllHll \^\^'i~H~U^_^

5 RM.

•sy.

75

♦•5".

I§X3 SBPSb.

s^'r-^

S PM.

«fc^

Fig 223 JULY/

5 PM-

AND TRANSPIRATION OF TREES

139

FlQ.XXD JULY^Z^Ji/

5 PM;,

Flg.XXinjULYAfc.

Ill IIHtX

dULy3/-

Q " 10 II M- ft

CLARK— ON WATER CONTENT

Fl§. XX32 AUO.5.

w^^--

10 ij ^M. a )

5 PM-

3CL

Fi

I^XTH AU&.9

»^^>uo-

10 It AM /2 /

5 PM

3 /O // An /i /

1 5m

AND TRANSPIRATION OF TREES

141

U

PS.

Fig KOm SEE IT.

JO n An 12 /

5 PM

90,

io

FlQ.

g. XXX OCT 3

90

F

l§. OIX SER5,5.

5 pn

CLARK— ON WATER CONTENT

/§.XXX] OCT. 9.

9 10 //Art 12

i> PI

AND TRANSPIRATION OF TREES

143

THE DEVELOPMENT OF THE CHASMOGAMOUS

AND THE CLEISTOGAMOUS FLOWERS

OF IMPATIENS FULVA

BY

Franklin B. Carroll, M. A., Ph. D.
With Plates LV, LVI, LVII.

(Thesis presented to the Faculty of the Graduate School
in partial fulfilment of the requirements for
the Degree of Doctor of Philosophy).

I. General Description of Growth, Habits, Size of Plants,

Habitat and Coloring

The Balsaminaceae include two genera, Hydrocera and Impatiens.
Hydrocera is a monotypic genus of Asia. Impatiens includes 400 species
mostly distributed through tropical Africa and Asia. A few species
occur in the North Temperate Zone. Nine species of Impatiens occur
in North America. Impatiens fulva is stated to occur in moist ground,
from Nova Scotia to Oregon and Alaska, south to Florida and Missouri.

Impatiens fulva is usually described as a succulent herb, two to
five feet in height. These figures represent an average height, but they
do not include the whole range of variations generally met under dif-
ferent conditions. Under conditions which may be somewhat adverse,
but not sufficiently adverse to cause any diminution in the produc-
tivity of the plants, lower plants are frequently to be found. In one
locality that has been under observation for two years, the plants have
never attained a growth of more than eight inches. The locality is a
somewhat marshy piece of ground about twenty-five square yards in
extent, at the foot of a hill, in close proximity to a spring and brook,
and in rather heavy woodland shade. Even in dry weather the piece
of ground remains rather marshy. In the two years that these plants
have been under observation, they have produced only cleistogamous
flowers, but these in abundance. Although rather pale green, these
plants seemed perfectly healthy. The failure to attain greater height
apparently is related to the bog conditions, perhaps the toxicity of
bog water. To determine whether the height of these plants could be
explained ecologically, several seedlings were transplanted in June 1916
to four-inch pots of drier soil. The average height of these transplanted

FLOWERS OF IMP ATI ENS FULVA 145

specimens was one foot ten inches. The five plants which survived
produced only cleistogamous flowers. The failure to produce chas-
mogamous flowers may be due to confined rootage, as will be noted
below in discussion of further experiments. Several other similar
localities were under observation during the summer of 1916. The
average height attained by the plants was eighteen inches. They were
characterized by the same pale green color, but seemed healthy and pro-
duced abundant cleistogamous flowers, but few chasmogamous flowers.

Short-stemmed plants are produced also in situations where the
soil moisture is below the optimum for the species. At the top of the
hill, above the plants just referred to, in an open woodland one such
group of plants has been under observ^ation for two years. During
May and June the soil is quite moist, but in mid and late summer the
soil becomes decidedly dry. During June and early July in both years,
these plants were in good condition. They were of a darker green than
the bog plants noted above, and some showed some evidence of the
ruddy stems characteristic of the species. These produced abundant
cleistogamous flowers during June, and moderately abundant chas-
mogamous flowers in July. They were killed by dry weather in August.
The average height of these plants was two feet, and none were observed
over two feet seven inches. Similar groups of plants in other sunny
situations, where the soil was somewhat dry, were observed in fair
abundance, averaging two feet or less in height.

Under the most favorable conditions relatively enormous heights
are attained. The optimum conditions for vegetative growth seem to
be found along brooks in rather open woodlands, with the roots of the
plants washed by the running water. Under these conditions, plants
six feet in height were found. The tallest plant measured in the vicinity
of Philadelphia was six feet four inches in height. The stems and
branches of these plants were a dense reddish color. The plants at-
tained a circumference of over six inches near the base, and were but-
tressed by series of adventitious roots growing from the lower two or
three nodes above the ground. Dr. Macfarlane reports plants of the
species on Peak's Island, Maine, attaining a height of seven feet, and
seven feet five inches. These plants were growing upon a sloping bank
facing the sea. The abundance of percolating moisture and the open
gravelly soil probably account for the enormous growth.

The habitat given by the Manuals for this species is moist shady
places, in rich ground. The species is abundant, however, not only
in this habitat but also in more sunny and drier locations. The optimum

146 CARROLL— ON DEVELOPME^JT OF

conditions seem to be a moderate amount of shade with moist, but well-
drained ground. Plants growing in poorly drained or marshy situa-
tions behave as if under adverse conditions. The small plants already
noted as growing under such conditions showed evidence of adverse
enviroment in their height, color, foliage and flowers. The giant speci-
mens noted above, on the other hand, cannot be said to grow under
optimum conditions for the species, since the production of flowers was
decidedly meager in most cases observed. The optimum conditions
for flowering are not found when the running water is washing the roots
of the plant unless they are situated in a slight shade, or even in the
open sun. Much of the material used in investigating the histology
was collected from several open sunny localities where the plants were
never shaded, along several tributaries of Darby Creek, near Phila-
delphia. The plants were exceedingly crowded, three to four feet in
height, and densely covered with flowers. In the other localities, the
plants were close to running water on well-drained banks; all were
characterized by exceedingly dense growth of plants and great abun-
dance of flowers.

The species flourishes also in much drier situations. Plants were
noted above as growing on a hill top in open woods; these were several
hundred yards from running water or marshy ground. Along one of the
drives in Fairmount Park, Philadelphia, a quarter of a mile from the
nearest brook and in full sun, except in the later afternoon, a dense
group of plants was kept under observation during the summer of 1916.
The plants were thirty to thirty-six inches in height and produced both
cleistogamous and chasmogamous flowers in abundance. On a sunny
bank in the botanical gardens of the University of Pennsylvania num-
erous plants have maintained themselves for several years. These are
stocky plants averaging four feet eight inches in height and five inches
in circumference. From the lower nodes there are circles of adventi-
tious roots. These plants bear abundant, though rather small, chas-
mogamous flowers during August. All these plants noted grow in rich
soil, but the species is also to be found in poorer, sandy soils. It is
met with in the sands of the New Jersey Pine Barrens. Experimentally
also the plants were grown in poor soils. Seedlings transplanted to
five-inch pots of sand maintained themselves quite as well as those
transplanted to pots of rich soil, although they produced as a rule fewer
chasmogamous flowers.

The color of flowers and stems is subject to considerable variation.
The flowers are typically yellow to orange-yellow dotted with brownish

FLOWERS OF LUPA TIEXS FULVA 147

red. The yellow color is born by plastids, while the red is due to a
dissolved pigment in the cell sap. It is chiefly on the petals and on the
inside of the large spurred sepal, to a lesser extent on the anterior haired
sepal that the brownish-red coloring occurs.

The seventh edition of Gray's Manual mentions the occurrence of
spotless forms. Such forms seem to be rare. Not uncommon are forms
with only a few, five or six, small reddish spots inside the spurred sepal
and an equal number on the petals, the anterior sepal bearing only
one or two spots. From this condition, gradations may be traced to
forms with densely spotted anterior and posterior sepals, and petals
with the spots confluent in brownish-red masses. Plants with white
flowers spotted with brownish black are occasionally met with. It is
stated in Britton and Brown's Manual that flowers nearly white un-
mottled occur. The coloring of the flowers seems to be definite for a strain
of plants, plants raised from seed develop the coloring of the parent.
The ruddy coloring of the stems, however, does not seem to be so defi-
nite a genetic character, but dependent to an extent upon ecologic
conditions. Plants grown under somewhat adverse conditions are apt
to show greenish stems without the red markings. Under optimum con
ditions for vegetative growth, as in partial shade with abundance o^
running water, the stems become dark red.

Dates of Appearance and Duration of deistogamous and
chasmogamous Flowers

The dates of appearance and duration of chasmogamous and deisto-
gamous flowers vary considerably with the locality and conditions.
In the vicinity of Philadelphia, the earliest date noted for the appear-
ance of chasmogamous flowers in 1916 was June 20th. They appeared
in abundance by June 24th on a group of plants growing in the open sun
in a dry roadside situation. The dryness of the soil and consequent
higher temperature may explain their early appearance. The period
of maximum production began a week or ten days later. Plants in
moister situations were relatively late in blooming. One group of
plants under observation situated in a rather densely shaded situation at
the edge of a marsh did not produce chasmogamous flowers until early
August. The period of blooming extended to the last days of September
and in scattered instances to early October.

The deistogamous flowers appear early in June in the vicinity of
Philadelphia and last through the growing season. This statement does
not agree with several pubHshed accounts. Miss Riatt^^ states for

148 CARROLL— ON DEVELOPMENT OF

Impatiens pallida that the cleistogamous flowers appear in May and last
through June, a few lasting throughout the summer. Bennett* states
for Impatiens fulva that on the banks of a tributary of the Wey the two
kinds of flowers were "absolutely synchronous." Bennett quotes
Weddell as reporting for /. noli-tangere that the inconspicuous flowers
occur latest, and Mohl as reporting only cleistogamous in June, and
only chasmogamous flowers in September. He reports plants of /.
noli-tangere observed in the Botanic Gardens at Oxford on the last day of
September as having "abundance of cleistogenous flowers with half-
expelled corolla-cap, while scarcely any of the perfect flowers were met
with and these on different plants." The earliest date noted in the
vicinity of Philadelphia in 1916 for the appearance of cleistogamous
flowers in /. fulva was June 10 and in /. pallida June 13th. Bennett says
"I have never found the two kinds on the same branch, occasionally on
different branches of the same plant, but more often on separate plants. "
In southeastern Pennsylvania in both species the cleistogamous flowers
were found in abundance along with the chasmogamous flowers through
the entire summer until late September. In Potter County, Pa., and
in Algonquin Park, Ontario, both types of flowers were found in abun-
dance during August. Indeed one of the most noticeable facts about
the cleistogamous flowers is their universality. They are less abundant
in mid-summer on the taller and more vigorous plants, but on the ex-
tremely tall plants noted above chasmogamous flowers were also gen-
erally less abundant. Smaller plants bear the cleistogamous flowers
more abundantly, yet plants of average height growing under good
conditions were found to bear both kinds in fair abundance, the cleis-
togamous flowers appearing upon the shorter and lower branches. Even
in mid-summer careful examination of lower, short side-branches of
plants in full chasmogamous flower showed the presence of the cleis-
togamous flowers and fruit.

Pseudocleistogamous Flowers

In addition to the truly cleistogamous flower, an intermediate type,
or rather various intermediate grades between cleistogamous and chas-
mogamous, are to be found. Bennett says that he never noticed the
least indication of any intermediate condition between the two kinds
of flower (4 p. LS2). He states that Bentham and Boswell-Syme de-
scribe the two kinds of flowers as growing intermixed in the same raceme
in /. noli-tangere. I have found intermediate types in both pallida
a.nd fulva in the same inflorescences with normal chasmogamous flowers.

FLOWERS OF IMP A TIENS FULVA 149

These flowers are considerably larger than the typical cleistogamous
flowers. The cleistogamous flowers at the stage ol fertilization average
two millimeters. Their morphology is that of the chasmogamous
flowers. The four sepals are present in the same position and form.
The anterior sepals show the same double appearance; the posterior
has the well-developed spur; the two lobed petals are fully developed,
the whorl of stamens appears fused and each stamen bears four pollen
sacs like the chasmogamous, but unlike the cleistogamous flowers which
have only two pollen sacs. They could not be distinguished from chas-
mogamous buds of the same age. Apparently these flowers are devel-
oping to be chasmogamous flowers, but fertilization occurs before the
bud opens. The whole mass of calyx, corolla and stamens is then
loosened, and as the pistil expands this mass is forced off. Some of
these flowers were marked with tags at the stage when loosening of this
mass was first evident. Subsequent observations showed that the
majority of such marked flowers developed fruit and seed similar to
those of chasmogamous flowers.

These intermediate flowers belong to the type which Hansgirg
(20) has designated pseudo-cleistogamous. Morphologically the flowers
are like the chasmogamous, but physiologically they behave as cleis-
togamous flowers. Adverse conditions of light, temperature, moisture,
or drought apparently lead to self-fertilization in buds. Hansgirg
distinguishes photo-, hydro- and thermo-cleistogamous flowers. These
do not include all the causes which lead to this phenomenon. Numerous
references occur through the literature to the occurrence of the phenome-
non under varying conditions and in numerous species. Partial lists
of such species and conditions are given by Hansgirg (20) and Knuth
(26 I pp. 55-58 English Translation). In the case of I.fulva and pallida
observed in late August and early September 1916, they were probably
induced by dry weather in August. None was observed during July
of that year before the dry spell began, but they were not rare toward
the close of the dry weather and for a short time succeeding. In none
of the experiments described below, however, in which plants were
subjected to various adverse conditions, were pseudocleistogamous
flowers detected.

Still another type of flower which would fall within Hansgirg's group
of pseudocleistogamous flowers, was discovered in castration experiments
described below. Full grown buds just about to expand were cut open.
The pollen sacs were found already ruptured. The stamens were
removed, and the flower covered to prevent the entrance of insects.

150 CARROLL— ON DEVELOPMENT OF

In several of the rather small number of experiments, these flowers
produced good seed. This would indicate that self-pollination had
occurred before the removal of the stamens. How far this self-polli-
nation before the opening of the bud occurs in nature would be difficult
to determine. There is no indication of the occurrence in the behavior
of the flower. The castrated flowers were normal chasmogamous flowers
to all appearances. Apparently they would have opened their sepals
and petals within another day. In the Hght of the possibiUty of self-
pollination at this stage and the subsequent opening of the sepals and
petals, one could not say with certainty whether a particular flower
was poUinated before or after it opened.

Pollination and Fertility of cleistogamous and chasmogamous Flowers

References occur in the Hterature to a lack of agents of pollination
and to a high degree of sterility among the chasmogamous flowers of
/ fulva. Bennett (4) says "the conspicuous flowers are stated by all
observers to be usually barren." This does not agree with my own
observations. It is quite true as Gray states (18) that "many of the larger
and fully developed flowers fall away without forming fruit." In one
large group of plants under observation, not one chasmogamous flower
was observed to produce fruit. Yet the abundant production of fruit
from chasmogamous flowers was elsewhere evident. A large number
of flowers were marked with tags and observations carried through till
the collection of seed. Bennett thinks the arrangement of stamens
and pistils is such as to prevent self-pollination, yet equally adapted to
prevent cross-pollination. The filaments send out from their inner
faces broad processes which are connivent over the pistil in such manner
as to prevent any pollen from the anther reaching the pistil. The whole
whorl of anthers later falls away together, exposing the stigma. Bennett
states that he has never observed this to take place spontaneously,
but that the flower drops as a whole carrying the pistil along. This
phenomenon frequently occurs, but it seems to be an abnormal occur-
rence. Normally the whorl of stamens falls away exposing the pistil.
The petals and sepals fall several hours to a day later. The falling
away of the flowers with the pistil included, observed by Bennett on the
Wey may be associated with conditions of a foreign home. Under
adverse conditions in America the same process occurs. One group
of plants, nearly denuded of foliage by insects, showed this phenomenon
and similar instances were observed in dry weather.

FLOWERS OF IMPA TIENS FULVA 151

The agents of pollination in the vicinity of Philadelphia seem to
be chiefly bumble-bees and humming-birds. Bennett says he never
saw the flowers visited by insects. He quotes a curious old work,
"New England's Rarities discovered in Birds, Fishes, Serpents, and
Plants of that Country," by John Josselyn, Gent., London, 1675, in
which occur a drawing and a description of apparently Impatiens fulva
under the title of the Humming-Bird Tree. Dr. Macfarlane reports
stands of Impatiens fulva on Peak's Island and also on Mount Desert,
Maine, that were regularly worked over by a group of humming-birds.
Bees were not often seen to visit these flowers. In the Philadelphia
vicinity, humming-birds may frequently be seen visiting the flowers,
but the most persistent visitors seem to be bumble-bees. Smaller bees
and flies may sometimes be seen, but the bumble-bees are found visiting
flower after flower in continuous succession. The bee approaches the
flower as it hangs upside down, alights upon the broad lamina of the
petals and makes its way back to the inside of the spur, brushing its
back against the mass of stamens above. Bees emerging may be seen
to have their backs covered with pollen. The same day the flower
opens, or on the succeeding day, the mass of stamens falls away leaving
the pistil exposed in the flower. These detached whorls of stamens
are to be found lying about. A bee subsequently making its way mto
the flower must brush its back against the expanded stigmatic surfaces.
Mtiller (36) gives a similar account of pollination. Not infrequently bees
may be seen alighting upon the outside of the spur and making their
way to a point on the outside near the nectary, then on leaving to visit
another flower in the same manner. I have occasionally found small
flies working over the flowers in the same way, the flowers visited some-
times showing perforations and sometimes none in the spur. Only a
minority of the bees visited the flowers in the above manner, but the
bees following this method were not seen to enter the spur through the
open end. It was not determined whether the bees following this
method were a species different from those entering through the mouth
of the spur. Trelease (48) reports observation of this occurrence. He
determined that the bees did not make this perforation but merely took
advantage of perforations already existing.

The method of polhnation described for fulva applies also for pallida
and sultani. Somewhat different methods are reported for other species.
In /. noli-tangere, according to Knuth, (26) the pistil protrudes between
the staminal processes and so may be poUinated. In 7. parviflora
and / balsamina the same mechanism is reported (Knuth 26). In I.

152 CARROLL— ON DEVELOPMENT OF

glanduligera, Landl. ( = 1. Roylei Walp) Loew (30a) has described a dif-
ferent mechanism. A cleft between the anterior stamens forms a passage
through which the hairs on the back of the bumble-bee are thrust as
the bee makes its way below the stamens. The staminal processes aid
in the function of collecting the pollen forming a "pseudo-stigma."
According to Knuth, (26) Hildebrand and Delpino (cited by Knuth),
however, the mechanism is in this species much as in the preceding
species described.

Accounts differ as to the fertility of the various species of Impatiens
with their own pollen. In the Gardener's Chronicle for 1868 a quo-
tation from "American Gardening" states that the "Balsams, {Im-
patiens pallida and fulva) cannot be fertilized with their own pollen."
This does not agree with Darwin's account or with my own experiments.
In a personal communication from Darwin, quoted by Bennett, (4,
152), Darwin gives the following account — "Eleven such pods from
perfect flowers, spontaneously self -fertilized yielded on an average
3.45 seeds. I carefully brushed away the pollen from some of the
perfect flowers, and fertilized them with pollen from a distant plant,
but got only three pods, containing to my surprise only 2, 2 and 1 seed. "
Darwin (12, p. 366) states that Impatiens barbigera set seed abundantly
when covered with a net, and that plants of /. noli-tangere covered with
a net produced from perfect flowers eleven spontaneously self-fertilized
capsules which contained on an average 3.45 seeds (12, p. 367).

My own experiments in self-pollination included I. fulva, pallida
and sultani. Each proved fertile to its own pollen. Each was tested
by two methods, the first to determine the occurrence of spontaneous
self-pollination, the second to determine the results of the direct applica-
tion of pollen. In the first method the flowers were sealed in paper
envelopes to exclude insect visitors. In five experiments with sultani
no seed was set. In the greenhouse where were some two dozen pots
of /. sultani no seed occurred during the winter except as a result of
hand pollination. When cross-pollinated by hand they set abundant
seed. In five flowers of fulva sealed to allow spontaneous selfing, three
set seed, and in five flowers of pallida, one set seed. In the second
method to test the effects of direct application of the pollen, a loosened
whorl of stamens was removed, or one was taken that had fallen from
position but lay on the petals of the flower, care being observed to
exclude possibility of cross pollination, and the pollen was applied to
the stigma. Five experiments on each of the species resulted as follows:
in pallida four set seed; in fulva five; in sultani, five. In pallida and

FLOWERS OF IMPA TIENS FULVA 1 53

fulva some flowers were castrated just before the opening and were
sealed in paper en\'elopes. In some of these the pollen sacs had already-
ruptured. In pallida the results of five experiments were as follows:
one capsule with five seeds; one with four; one with three; and two>
failed to form fruit. Somewhat similar results were obtained with
fulva; unfortunately the records of these have been lost. The results
with these castrated flowers would indicate that spontaneous seK-
DoUination sometimes occurs before the opening of the flowers.

The number and size of the seeds in cleistogamous and chasmogamous
capsules, Miss Riatt (39 )states to be the same in I. pallida. So far as
the number of seeds in the capsule is concerned, this does not agree with
my observations on either pallida or fulva In pallida, I have found
in the vicinity of Philadelphia the number of seeds in the chasmogamous
capsules is generally four or five; in the cleistogamous capsules two or
three, or not uncommonly on the lower branches only one. This num-
ber represents also the condition in fulva. My observation of the
sizes of the two types of seeds in fulva agrees with Miss Riatt's report
of pallida. The average size in the chasmogamous seed was 5.1 X 4 m. m.
and in the cleistogamous seed 5X4 m.m. The weights varied with the
condition of drying. In seeds gathered from the bursting pod the
average weight in 48 chasmogamous seeds was 0.0208 grams and in
59 cleistogamous seeds was 0.0195 grams. The size of capsules is
quite different in the two types of flowers, but the smaller capsule of
the cleistogamous flowers is accounted for by the smaller number of
seeds.

There seems to be no general rule in the different species as to the
size and number of cleistogamous as compared with chasmogamous
seed. Shaw (45) says for Polygala polygavia that there is no difference in
the seed resulting from the different types of flowers. Helene Ritzerow
(41, p. 71) quotes Burck (11) to the effect that in Heteranthera spicata the
cleistogamous fruit is a half larger than the chasmogamous and in
Heteranthera Potamogeton, Solms, and Heteranthera Kotschyana, Fenzl,
it is more than twice as large with twice the number of seeds; in Com-
melina bengalensis (41 p. 172) the seeds of the cleistogamous flowers
are double the weight of those of the chasmogamous flower. In Van-
dellia nummular if olia (41, p. 201) the number of seeds in the cleisto-
gamous flowers is on the whole lower than in the chasmogamous. In
Houstonia caerulea (41, p. 204) the size of the seed is larger in the chas-
mogamous capsule, but the number of seeds per capsule is larger in
the cleistogamous capsules. In Specularia perfoliata (41, p. 208) the

154 CARROLL— ON DEVELOPMENT OF

number of seeds was greater in the chasmogamous form. The relative
number and size of the two types of seed are thus seen to vary considerably
among different species.

II. Experiments and Observations on the Production of chas-
mogamous AND CLEISTOGAMOUS FlOWERS

Observations in the field on the production of cleistogamous flowers
show a wide range of conditions under which they appear. In general
as was noted above, conditions.. which favor greatest vegetative growth,
lead to small production of both kinds of flowers. These conditions
seem to be rather dense shade and rootage in soil saturated with fresh
running water. Along brooks in light shade or open sunny meadows
the plants are shorter and produce chasmogamous flowers in greatest
profusion. Before the production of chasmogamous flowers these
plants are quite generally to be found bearing abundant cleistogamous
flowers and careful examination of the plants during the period of
maximum production of chasmogamous flowers will frequently show
shorter shaded branches on the lower parts, and these bear cleistogamous
flowers. In one such situation where every one of twenty-five plants
was examined the cleistogamous flowers were found in the act of pushing
off the perianth cap or these caps were found lying on leaves or other
parts of the plant. In one station by an open sunny roadside that was
under observation during the last week of July and the first week of
August, and again during September 1916, chasmogamous flowers and
seed were produced in abundance, and cleistogamous flowers which
were abundant during June and early July continued more sparingly
in the later weeks. Under adverse conditions the production of chas-
mogamous flowers is reduced, but that of cleistogamous flowers remains
constant as in the early part of the season, or increases. In a station
on a hillside in the open woods that was under observation during the
spring and summer of 1916, the plants produced abundant cleistogamous
flowers from the middle of June to the middle of September. During
July and August a few chasmogamous flowers only were produced. The
soil during June was rather moist, but during the later summer was
quite dry for a depth of two inches. In a bog in the woods at the foot
of this hill, pale green plants eight to ten inches high lasted through
the summer, producing cleistogamous flowers very sparingly and chas-
mogamous flowers not at all. Conditions here were very similar to
those in the bog noted above. Four such boggy situations in the woods
were under observation during the entire summer. The conditions

FLOWERS OF IMPA TIENS FULVA 155

were similar in all — a small, stagnant, poorly-drained spot in rather
dense shade. The plants in each were less than a foot in height, of a
pale green color, and producing rather sparingly cleistogamous, and
no chasmogamous flowers. The adverse condition here would seem tO'
be the toxicity of the bog water. Plants in the same condition o£
shade, but in well-drained soil attained heights of two and a half to
three and a half feet, with darker green, or somewhat reddish stems,
and chasmogamous flowers. Conclusions derived from observation of
field conditions are that under the optimum condition for vegetative
growth both kinds of flowers are rare; under conditions optimum for
the production of chasmogamous flowers, cleistogamous flowers are
produced in abundance in the early moister part of the season and more
sparingly throughout the season ; under adverse conditions of drought or
toxicity of bog soil only, or chiefly, cleistogamous are produced.

Goebel (16) pei formed a series of experiments upon several plants
including Impatiens noli-tangere to determine the conditions of produc-
tion of cleistogamous flowers. He planted Impatiens noli-tangere
seeds in pots of earth and sand. Some of the sand pots he watered
with nutrient solutions. The plants in earth produced both chasmo-
gamous and cleistogamous flowers; those in sand, only cleistogamous.
Those watered with nutrient solution produced more abundant cleisto-
gamous flowers, but no chasmogamous. He reported that plants of
this species, attacked by Sphaerolheca Castagnei and suffering loss of
leaves in July while in chasmogamous flower showed on August 11th
only cleistogamous flowers, while healthy plants of the species were
still in chasmogamous flower. He cites also lighting and high tem-
perature as leading to the production of cleistogamous flowers.

My own experiments on cleistogamous flowers aimed to determine
the production of the two kinds of flower under adverse conditions of
drought, nutrition, leafage and temperature. Young plants averaging
3.5 dm. in height were transplanted from the woods to pots on May
29th. Plants were marked in the woods to serve as controls. To test
the effect of drought eight pots were selected, four in rich soil and four
in white sand. These were watered sufficiently to keep the plants on
the point of withering. One plant in sand died within a week without
producing flowers; one died within two weeks after producing a few
cleistogamous flowers; two continued to live for a month producing
cleistogamous flowers sparingly. Of the plants in the soil one died
in a few days without flowers; two died before the end of the third week
after producing a few cleistogamous flowers; one lived over a month

156 CARROLL— ON DEVELOPMENT OF

with a sparse production of cleistogamous flowers. No chasmogamous
flowers were produced. The plants in the soil were somewhat healthier
probably because of the greater retention of moisture by the earth.
To determine the effect of high temperature, two pots were kept in a
■greenhouse with temperature 15° F. higher than outside. One plant
■succumbed very quickly, the other lived twenty- two days in the house
•and produced three cleistogamous flowers. To determine the effect
■of shade, two pots were placed in dense shade, where they continued in
good health for six weeks producing cleistogamous flowers continuously.
To test the effect of soil nutrition, two pots in white sand were watered
with Pfeffer's nutrient solution, two in white sand received only pure
water and two in soil served as controls. One of the pots in soil pro-
duced chasmogamous flowers; all the others produced only cleistogamous
flowers. To determine the effect of reduced leafage, plants in the
woods from which the potted plants had been taken, were stripped of
all except two or three leaves. Two died early, three continued to
produce occasional cleistogamous but no chasmogamous flowers. The
controls in the woods produced abundant cleistogamous flowers during
June and July. During July, they produced chasmogamous flowers
rather sparingly. A dry spell in August killed practically all of the
plants.

The results of experiments are in accord with the facts ascertained
by field observation. Cleistogamous flowers are generally produced
in the early moister part of the season on the majority of plants, even
those living under optimum conditions. They may continue under
even optimum conditions through the season along with chasmogamous
flowers. Adverse conditions of drought or nutrition, directly through
failure of necessary elements, or indirectly through lighting, or through
reduction of leafage by mechanical agents, or the attack of parasites,
or toxic conditions of soil, may lead to the suppression of the chas-
mogamous flowers. The cleistogamous flowers will then continue, or
will increase in numbers.

III. General Organogeny of the chasmogamous Flower

The development of the flower of Impaticns pallida has been de-
scribed by Miss Riatt (39). Her account differs somewhat from the con-
ditions in fulva. In the early stages the conditions are the same in both
cleistogamous and chasmogamous flowers. Miss Riatt states that the
development of the flower begins with the appearance of the saccate
.sepal. This is not the case with Impatiens fulva. The first member

FLOWI'RS OF IMPA TIEXS FULVA 157

to arise is the bract (Fig. 1 A, Fig. 7 A). A little later the two lateral
sepals appear (Fig. 2). Next the posterior saccate and then the anterior
sepal (Fig. 3). Figure 7 is a longitudinal section of a cleistogamous
flower, but the relation of parts here shown is exactly the same as in the
chasmogamous flower. Miss Riatt's Fig. 1 A is a longitudinal section
of a stage slightly older than the older stage of Fig. 7. The member
she marks saccate sepal looks very much like the section of the bract.
The lateral sepals would not appear in this section. The transverse
sections (Figs. 1, 2 and 3) show the relation of the bract and sepals.
Payer's account of Inipatiens Royleana agrees with the above account
oi fulva. The petals appear as two prominences inside the sepals (Fig.

4 E). These are the posterior lobes. A little later the anterior lobes
arise (Fig. 6). For a time these prominences arise separately; later
they join in pairs, and each prominence then rises to form the single
claw of a two-lobed petal. The stamens appear about the same time
(Fig. 4 F and Fig. 8 C). At first the stamens are entirely free, and
at a later stage they become joined in their anther wall and filaments.
Three papillae inside of the stamens mark the developing pistil (Fig,

5 A). The pistil is still quite open when the ovules are well advanced,
and it never entirely closes, stylar canals remaining open to receive the
pollen tubes. The stylar canals are formed here not by bieaking down
of tissue, but through failure of the pistil to entirely close over or fuse.
Figure 33 shows pollen tubes passing through the stylar canals in a
cleistogamous flower. This figure, except for the anthers and calyx
above, would serve equally well for the chasmogamous flower. The
placenta arises from the floor of the carpellary cavity, and the septa
push inward from the carpellary walls to meet the placenta (Fig. 6).
The placenta would thus seem to be axial in origin, and the ovules arising
upon it, cauline ovules.

Morphology of Petals and Sepals in the chasmogamous Flower

At one time there was considerable controversy as to the morphology
of the petals and sepals in the genus Impatiens. Payer (37) summed up
the various views up to his time. Jussieu had attributed to the flower of
Impatiens balsamina two small lateral sepals, and four petals of which
the inferior was spurred, the superior hooded and the two lateral deeply
bilobed. Richard viewed the flower as having four sepals — two small
lateral, one spurred inferior, and one superior — and four petals which
had become joined in pairs. Knuth viewed the calyx as composed of
the two small lateral sepals, of a spurred inferior sepal and a superior

158 CARROLL— ON DEVELOPMENT OF

formed of two joined, and the corolla as composed of four petals joined
in pairs, the fifth being aborted. Bernhardi viewed the two small
green sepals as two bracts. The calyx to him was composed of five
members — an inferior spurred sepal, a superior sepal composed of two
joined, and two little sepals which usually aborted and which botanists
had not mentioned. The corolla was composed of five petals — four
joined in pairs and a fifth which had united with the two joined superior
sepals. Roeper viewed the calyx as composed of five parts — two were
often rudimentary or disappeared, two were small lateral sepals,
and the fifth spurred sepal was toward the axis of the inflorescence. The
corolla was composed of five petals which alternated with the sepals —
four were joined in pairs, the fifth was the superior petal of the older
botanists. Payer accepted the view of Roeper, and agreed that the two
small prominences representing the anterior sepals were to be found
in the young stage of Impatiens balsamina and figured them in the
young stages of Impatiens Royleana. Lindley (28) followed the view of
Knuth. Henfrey (22) arguing from the morphology of double balsams
adopted Roeper's view although he failed to find the rudiments of the
two anterior sepals reported by Roeper and Payer in Impatiens bal-
samina. The view of Roeper was adopted by Warburg and Reiche
in " Pflanzenfamilien. " Gray's (18) view agreed with that of Knuth,
and this was reported in the seventh edition of Gray's Manual. Brit-
ton and Brown's Manual followed the view of Roeper. The double
appearance of the anterior member in our native species would seem
to suggest that it is formed of two members which if it be regarded
as belonging to the petaline whorl would raise the number to six mem-
bers, but if it be regarded as belonging to the sepaline whorl would
raise the number to five. The occurrence occasionally of two spurs
upon this structure would further emphasize its double nature. The
fact that supernumerary spurs occur upon the two lateral usually
greenish sepals, but not upon the two-lobed petals would tend to argue
its connection with the sepaline whorl. The occurrence in other species
of the genus of spurs on all five structures further argues for this
view. In this paper, I have therefore adopted provisionally the view
of Knuth and Gray in describing the calyx as composed of four sepals,
the anterior sepal being formed of two fused sepals and the corolla
as composed of two-lobed petals representing four petals fused in pairs.
These have been well figured by Gray (18).

Peloria is not uncommon in the genus Impatiens. It is stated that
the two lateral sepals are often spurred in double garden balsams.

FLOWERS OF IMP A TIENS FULVA 159

Masters (32) states that the balsams become peloric by the addition of
spurs. Penzig (38) takes the view that three sepals are present. The two
missing sepals sometimes reappear as small leaflets. In /. glandulifera,
all five sepals are present. Sometimes the side sepals are spurred and
sometimes all five when present. Britton and Brown (18) state in Im-
patiens fulva that "spurs are occasionally developed on the two small
exterior sepals, and spurless flowers have been observ^ed. " I have found
in I m pattens fulva not rarely spurs on the two small lateral sepals. These
spurs are generally much smaller than the posterior spur. Occasionally
the anterior sepal is found with the two spurs developed. These con-
ditions seem to be common to all flowers on the plants showing them.
These supernumerary spurs may develop in such manner as to prevent
the entrance of insects. The plants must then depend upon the cleis-
togamous flowers, or upon self-pollination in the chasmogamous flowers.

General Organogeny of the cleistogamous Flowers; Morphology of Petals

and Sepals

The general organogeny of the cleistogamous flowers is the same as
that of the chasmogamous flowers up to the stage of the differentiation
of the spur in the posterior sepal. Figures 7, 8, 9 and 33 represent the
conditions in the cleistogamous flowers. Bennett figures and describes
4, p. 148) a difference between the buds of the two kinds of flower.
"The bud of the conspicuous flower (fig. 1) has the apex of the two
exterior (lateral) sepals hooked, while in that of the inconspicuous
flower (fig. 2) it is straight, the two buds at this stage being nearly equal
in size. The removal of the two exterior sepals shows a still greater
difference (figs. 3 & 4), the spurred posterior sepal or nectary being
very easily seen in the former case (fig. 3), while in the latter (fig. 4)
the interior whorls of organs are, as described by Prof. Gray, "nearly regu-
lar but never developing beyond a very minute size. " In earlier stages,
before the differentiation of the spur, this difference could not be de-
tected; cross sections, however, through the stamens, and especially
through the pollen sacs will prove the type of flower.

In the mature cleistogamous flower all parts are present that are
present in the chasmogamous flower. Miss Riatt (39) says that the petals
are aborted in the cleistogamous flowers of Impatiens pallida, and Brit-
ton and Brown (8, p. 304) state for the family that the cleistogamous
flowers are apetalous. Knuth reports them present in the cleistogamous
flowers of /. noli-tangere (8, p. 52). They are evident in I. fulva (Fig.
9). Gray (18) figures them in his dissection of the cleistogamous flowers

160 CARROLL— ON DEVELOPMENT OF

of fulva. They show the same two lobed structure as the chasmogamous
petals. The four sepals are present in the same position as in the
chasmogamous flowers (Fig. 9). The spur sometimes appears as a
prominence, but sometimes no trace of it could be detected. Darwin,
quoted by Bennett,'* speaks of the "nectary in the cleistogamous flowers
as a mere rudiment." Five stamens are present but differ from the
chasmogamous stamens in a manner to be explained below. The ovary
is five-celled at first, but becomes one-celled at maturity through the
breaking down of the septa.

The condition in Impatiens fulva must not be regarded as typical
for cleistogamous flowers in general. In many there is reduction in
either caH'x or corolla, or in both. Helene Ritzerow^^ lists a number of
species suffering such reduction. She states that in Specularia perfol-
iata the number of sepals is reduced; in Cardamine chenopodifolia and
Helianthemum glomeratum the corolla is wanting; in Viola sp., Poly gala
sp., and AmpJiicarpaea monoica the corolla is present but rudimentary;
in quite a number of species the corolla shows lesser degrees of reduc-
tion, among which she mentions Impatiens sp. but does not state which
species. In Impatiens noli-tangere the petals appear as whitish scales. (35)

The calyx and corolla do not open, but the two whorls are cast off to-
gether as a cap, (Fig. 2)2)) which has been compared in appearance to the
calytra of a moss. This phenomenon appears also in /. noli-tangere (26)
and in /. pallida. Bennett (4) and Gray (18) both figure these caps.
Bennett suggests that the mechanism of expulsion may be in part the elas-
tic filaments. In the expelled caps the somewhat coiled position of the
filaments suggests this condition. The expansion of the ovary would
seem to play also an important part in the process. Gray's figures
suggest this explanation, and caps may frequently be found adhering
to the tips of enlarged ovaries as if forced up simply by the increase
in the size of the ovary. A similar expulsion of the united mass of sepals,
petals and stamens is described for /. noli-tangere (26).

Of some interest is the occurrence of crystals and tannin throughout
the flowers and other parts of the plant. The occurrence of raphides
and tannin is characteristic of the genus (49). Sections even of very
young parts show the presence of an internal secretion. This tested with
ferric chloride proved to be tannin. It occurs in single cells, or scat-
tered, or in definite layers, and in vessels formed by the disappearance
of end walls of elongated cells. These tannin vessels are frequent
in the cortex of stem, petioles and peduncles. In hypodermal layers
of stem and leaves and through the mesophyll both isolated cells and

FLOWERS OF IMPA TIENS FULVA 161

continuous layers of cells are found filled with the tannin secretion. It
is equally abundant in the floral parts. As soon as the bract can be
recognized the hx-podermal layer shows every cell filled with tannin.
As the sepals and petals arise they show similar hypodermal layers.
The hypodermal layer of the wall of the ovary is composed of an enlarged
zone of cells each filled with tannin (Fig. 34). As the ovules develop
their integuments, similar hypodermal layers appear, and the prominent
ridges on the seed have the cells filled with the secretion. Even in the
anther, in a circle around the connective these cells are prominent in
cross sections. Similarly, enlarged cells densely filled with needle
crystals are scattered throughout the stem, leaves and floral parts.
These appear in the sepals, in the petals, in the filaments and anther
walls, and even in the developing pollen (Fig. 11), also in the walls
of the ovary and in the integuments.

IV. The Stamens in the chasmogamous Flower

The stamens in the chasmogamous flower arise as five protuber-
ances in the manner described above. As these protuberances increase
in height, they increase in width to broad masses. Before the anther
lobes have been differentiated the upper parts of the filaments have be-
come greatly expanded. These expanded portions later come in con-
tact, and form a united whorl of stamens joined in the upper part of
the filaments but free below. The adjacent anther walls also fuse to
a greater or less extent. From the inner face of each stamen develops
a broad flap of tissue which extends inward above the pistil (Fig. 5).
These may represent ligules from the microsporophylls. These ligule-
like processes grow inward until they meet in the center. In the cen-
ter and on their edges, they become organically fused forming a roof
over the pistil. Generally no opening occurs in this roof. Occasionally,
however, the pistil may be found to have forced its way between the
ligular flaps and to project slightly above it. Gray (18) figures the parts
thus. Generally the roof seems to remain intact until the staminal
whorl drops from its position.

The pollen arises in four tracts, and the four pollen sacs remain
for some time after the formation of the endothecium. Gradually
the intervening tissue between neighboring sacs is broken down, forming
the two pollen sacs of the mature anther. Gray states that the two
pollen sacs are sometimes confluent at the apex (18). By this time inter-
nal layers of the anther wall have disappeared except for crushed tags
adhering to the inside of the endothecium. Shortly before dehiscence

162 CARROLL— ON DEVELOPMENT OF

the hypodermal layer, which is composed of greatly enlarged cells,
begins to take on the thickening characteristic of the endothecium.
Adjacent to the connective this enlargement of cells and development
of thickenings may involve two or three layers of cells. The epidermal
cells become somewhat swollen and vesicular. Toward the point of
rupture these epidermal cells disappear, and toward the connective
they increase to two layers. It is the tension set up between these two
tissues which apparently accounts for the rupture. After dehiscence,
the epidermal cells appear as shrunken remnants adherent to the outer
wall of the endothecium. If not too old these epidermal cells after
dehiscence still show a tendency to swell up. Apparently the loss of
turgor and collapse of the epidermal cell releases the tension of the
endothecium and dehiscence occurs. Loew (30a) (discussed briefly in
Engler and Prantl (49) has described the mechanism of dehiscence in
7. Roylei, Walp. About the time of origin of the endothecium, pairs
of pollen chambers of neighboring stamens become joined to form one
cavity, through breaking down of intervening walls. A similar appear-
ance occurs in I. fulva, but there is no breaking down of walls between
stamens, but rather a dehiscence of the external pollen sacs in the
lower part of the anther toward each other. In I. Roylei, according to
Loew, through a drying of the epidermal cells the anther walls press
into the pollen chamber and force the pollen upward as a column.
Dehiscence in fulva occurs longitudinally down the inner face. Usually
dehiscence takes place before the opening of the flower, so that when
the sepals and petals of the unopened flower are separated the interior
is found to contain abundant discharged pollen. Similar early dehiscence
occurs in other species (Knuth 26). Several hours to a day after the
opening of the flower the whole whorl of stamens becomes detached
together and falls from the flower. As the whorl falls it generally
carries with it much good pollen. In experiments with self-polhnation
noted above on I. fulva, pallida and sultani the pollen from the fallen
anthers was used to pollinate the stigma with entire success. Bennett*
states that he never found that the stamens fell away spontaneously,
but that the whole flower fell carrying the still concealed pistil with it.
This conduct of the plant in England seems to have been an abnormahty,
perhaps induced by a somewhat cold, moist, unfavorable climate in a
foreign home. In America the petals and sepals persist for several
hours or a day after the fall of the stamens; flowers bearing pistil but
no stamens are normal in fulva and pallida and in sultani in the green-
house.

FLOWERS OF IMPA TIENS FULVA 163

In view of the structure of the staminal whorl, the occurrence of
spontaneous self-pollination in the experiments with pallida and fulva
described above, and the occurrence of pseudocleistogamy present a
problem. The ligule-lLke staminal processes roof over the pistil in
such manner that no pollen from the anthers could lodge upon the stig-
ma. In the experiments referred to, the flowers before opening were
enclosed in paper envelopes so as to exclude insect visitors. A number
of these flowers set seed. In the pseudocleistogamous flowers fertiliza-
tion occurred before the opening of buds that morphologically were of
the chasmogamous type. Fertilization occurred at various stages
from that when the bud was five miUimeters in length to the mature
flower just about to open its sepals and petals. There are three possible
methods by which pollination might occur before opening of the flowers.
In the first place, the pistil might, as Gray figures, protrude through
the staminal ligules and pollen grains lodge upon it. The pistil is
figured protruding through the staminal ligules in /. balsamina by Engler
andPrantl (49) and in /. «o//-/awgere by Le Maout and Descaisne (27).
Knuth (26) describes the protrusion of the pistil in /. noli-tangere, parvi-
flora and balsamina. In /. latifolia, the stigmas are not covered (30a).
The detection of this protrusion in pseudocleistogamous flowers would
be a Utle difficult owing to the fact that the condition of pseudocleisto-
gamylis not likely to be recognized until the perianth cap is being pushed
off. At this stage the stamens are so far shrivelled and dried as to
place little value upon any appearance of opening through the staminal
flaps. Furthermore, the protrusion of the pistil in this manner is of
apparently rare occurrence in the chasmogamous flowers, yet the per-
centage of spontaneously self -fertilized flowers in fulva was surprisingly
high, three out of five cases.

The second possible method by which self-pollination could occur is
by the growth of pollen tubes directly through the mass of the staminal
roof to the stigmatic sufaces below. No pollen tubes taking this course
were ever detected either in the dissection of the flowers or in micro-
scopic sections. Such negative evidence is, however, of little value
because a structure as dehcate as a pollen tube would hardly remain
intact through the drying and shrivelling of the tissue that occurs before
these pseudocleistogamous flowers are recognized.

A third possibility is the detachment and dislocation of the staminal
whorl before the opening of the flower. In the smaller pseudocleisto-
gamous flowers this dislocation would be impossible because of the small
size of the flower and the compact condition of the floral parts. In the

164 CARROLL— ON DEVELOPMENT OF

older stages there is a possibility that it might occur. The failure to
detect it cannot be taken as evidence against the existence of this condi-
tion, as no definite search was made with its detection in mind. Fur-
ther investigation is necessary to explain the method of self-pollination
in these cases.

The arrangement and behavior of the stamens in other members of
the genus do not always agree with fulva. In pallida, the staminal
structure cannot be distinguished from that of fulva. In sultani the
structure is similar, but the relative sizes slightly different. In the com-
plete covering of the pistil by the Hgular flaps, in the dehiscence of the
anthers before opening of the flower and in the subsequent fall of the
whole staminal whorl, sultani resembles fulva and pallida. In noli-
tangere, (Knuth 26, II p. 236) dehiscence occurs before opening of the
flower, subsequently the anthers separate and the stigma matures.
In balsamina (Knuth 26, II p. 236) the mechanism is the same as in
noli-tangere. Knuth reports the mechanism of parviflora as being
similar to that of balsamina, but in many of these forms the whorl of
stamens is easily detached and would readily fall away owing to the
resupinate position of the flowers.

The Stamens in the cleistogamous Flowers
The early stages of staminal development in the cleistogamous
flowers are exactly like those of the chasmogamous, but the mature
condition is strikingly different. Not infrequently one o^- two of the
stamens are shorter than the others. The mature stamen is narrow and
strap shaped. There is no tendency to broaden and unite with adja-
cent stamens as in the chasmogamous. The flap-like ligular appen-
dages on the inner face are absent. The stamens meet above the pistil,
but there is no tendency toward fusion. Bennett has published good
figures of these in their mature condition (4 fig. 10 and 11). The pollen
sacs are greatly reduced in capacity and only two, instead of four as in
the chasmogamous flowers (Fig. 15). Each sac is surrounded by an
endothecium which is exactly similar to that of the chasmogamous
anthers. The great extent of the endothecium in the cleistogamous
flower is striking. Relatively it is much greater in extent than in the
chasmogamous flower, being well developed on the connective side
of the pollen sac, as well as on the outer sides. The endothecium pos-
sesses similar thickenings and opens by a stomium on its inner face
(fig. 15). The quahty of pollen is very greatly reduced, frequently
only two or three grains appearing in a microscopic section, and these

FLOWERS OF IMPATIENS FULVA 165

nearly filling the diminutive pollen sac. The anthers dehisce (Fig. ZZ)
but the pollen grains are not discharged from the pollen sacs. They
germinate within the sac and the pollen tubes pass through the stomium
directly down to the stigmatic surfaces below. Thus, although there
is a great reduction of pollen, there is also a great reduction in the
amount of waste.

This reduction of parts in the anthers is common to cleistogamous
flowers. In /. noU-tangere the pollen sacs are small, not containing
more than forty or fifty pollen grains. An endothecium is present and
dehiscence occurs, but the grains germinate in the pollen sacs and the
pollen tubes pass out to the stigma below (26 p. 53). In Oxalis acetosella
"the number of pollen grains in each loculus may not exceed two
dozen" (26). Helene Ritzerow (41) gives a summary of conditions in
various cleistogamous flowers, listing species in which the various condi-
tions occur. In not a few species a reduction in the number of stamens
occurs. Perhaps the occasional occurrence of one or two shorter stamens
in fulva is a tendency in this direction. In the majority of species
there is a tendency to reduction in the number of pollen sacs from four
to two. An endothecium is present in all cases except in Amphicarpaea.
In the majority of species, the pollen grains germinate in the anthers;
in a few they fall out and germinate on the pistil. The pollen tube may
pass directly through the wall if the endothecium is absent or does not
open, if it opens they pass through the opening.

Development of the Pollen

The development of the pollen is the same in each type of flower.
The archesporial cells were first recognized when separated by one parie-
tal layer, from the epidermis (Fig. 10). One or two cells appeared thus
in the cross sections of the young anther lobes. In longitudinal sec-
tions these appeared as a plate of cells one layer in thickness. The
primary parietal layers divide to form four or five layers. The outer-
most parietal layer, immediately under the epidermis, becomes differen-
tiated as the endothecium. The tapetum arises as a single layer from
the sporogenous cells. This primary tapetal layer later becomes a
double tapetal layer (Figs. 11 and 12). By the time the sporogenous
cells have arrived at the mother-cell stage the tapetum shows signs of
flattening as if through crushing by the sporogenous mass, or by breaking
down to furnish nourishment for the developing spores. By the time
the tetrad stage is reached the tapetum and one, two, or three parietal
layers have collapsed (Fig. 12). Before the mother-cell stage is reached

166 CARROLL— ON DEVELOPMENT OF

a remarkable degeneration of potentially sporogenous tissue is proceeding
(Fig. 11). Plates of cells ramifying through the sporogenous mass under-
go this process of degeneration. Among these plates of cells appear
cells filled with needle crystals (Fig. 11). Similar needle crystals are
quite common in the wall and other parts of the different members of
the flower. This degeneration of pollen is generally quite extensive,
involving at times fifty percent of the potentially sporogenous tissue.
The history of the pollen is an illustration of the great amount of loss
suffered by originally sporogenous cells. First, the early cutting off the
tapetum removes a large percentage of available tissue, then the exten-
sive degeneration of sporogenous cells, and lastly the loss mentioned
above of a very considerable amount of pollen with the falling away of
the whorl of stamens before the total discharge of pollen. In the cleis-
togamous anthers the loss through the organization of the tapetum
involves as great a per cent as in the chasmogamous anthers; the steriliz-
ation through degeneration is less extensive, however, and as the pollen
grains have germinated before the fall of the cap, practically none is
lost at this stage.

At maturity the pollen grains are identical in the two types of flowers.
They are oval or somewhat barrel-shaped measuring on an average
0.03 X 0.019 mm. The intine and exine are quite evident in the one
nucleate pollen grain. The exine in both types of flowers becomes
marked with flat-topped ridges dividing the surface into more or less
hexagonal areas. In the chasmogamous flowers the pollen grains are
sometimes loosely held together by threads. In the cleistogamous
flowers these threads do not appear; but if their function is to hold the
grains together, they would be functionless in the cleistogamous anthers
where the amount of pollen is very small, and it is never discharged from
the pollen sacs. There are easily distinguished four points of exit of the
pollen tubes. When germinated in sugar solution four tubes issue from
these points and attain a considerable growth. It was not discovered
whether the pollen grains on the chasmogamous stigma gave forth more
than one tube apiece, but in the germinating grains of the cleistogamous
anthers only one tube issued from a grain in the cases examined.

The germination of the microspore is quite t}q3ical. The one nucleate
stage is of relatively long duration (Fig. 13). The spore coats become
differentiated and the exine marked with the characteristic ridges.
Division gives rise to a large tube nucleus and a smaller generative
nucleus. The tube nucleus is round and takes a position in the middle

FLOWERS OF IMP A TIENS FULVA 167

of the sac. The generative nucleus takes a position alongside and
becomes much elongated (Fig. 14).

V. The Pistil

The early stages in the development of the pistil were referred to
briefly above. Three small prominences mark the origin of the pistil
(Fig. 5 A). These rise separately for a time, then the tissue rises as
a ring with more or less of a gap on one side and bearing the three promi-
nences on the top. The septa arise as five prominences pushing in from
the ovarian walls (Fig. 6). These prominences meet the rising placenta
and so divide the ovary into five chambers. Before the ovary has
closed over, one ovule generally has arisen in each chamber. These
arise laterally from the central placenta. This history suggests the
cauline nature of the ovules. Generally one ovule only arises in each
cavity, and five seeds is a common number in the mature capsule.
Before the ovule has reached the fertilization stage the septa show signs
of breaking down in their internal layers of cells (Fig. 34). This degen-
eration continues in later stages.

At pollination the ovary is five-celled. The style is practically
absent. The stigmatic surface covers the upper and inner faces of
five little teeth on the top of the ovar\\ Near the bases of these
stigmatic teeth open the stylar canals. The stylar canals originate by
the failure of the ovary to close over completely in its upward stylar
prolongation (Fig. Zi).

By the time the archesporial cell is clearly recognizable the inner
integument can be made out as a prominence on the side (Fig. 16).
Before the mother-cell has reached the synapsis stage, the inner integu-
ment has reached the top of the nucellus and the second integument is
indicated (Fig. 24). As in other anatropous ovules, the second integu-
ment is absent on the funicular side of the ovule. The second integu-
ment continues growth until it approaches the inner integument in
height. This condition is attained before the mother-cell has divided.
From this point onward the two integuments rise together. Hence
only at the tip are the two integuments to be distinguished. Below
the tip they are indistinguishable under the microscope. They are
not fused in the sense that two separate tissues have fused, but the
tissue below the tip arises as one continuous tissue. The layer of the
integument lying against the nucellus early becomes differentiated as
a zone of short broad cells apparently tapetal in function. This jacket
layer is remarkably persistent through the history of the ovule. It

168 CARROLL— ON DEVELOPMENT OF

remains after the destruction of the peripheral nucellar layer and con-
sequent increase of the size of the embryo sac after fertilization. (Fig.
28) After the embryo has filled the embryo sac, this layer appears as a
densely staining but somewhat crushed investment (Fig. 35). Miss
Riatt (39) reports a similar jacket layer in the innermost layer of the
integument of pallida. To judge from her figure 1, H, however, there
is one point of difference from fulva in the differentiation of the two
integuments. The hne of separation between the two integuments is
shown reaching well down to the basal end of the ovule.

The development of the megaspore and embryo sac presents nothing
unusual. The archesporial cell is recognizable as a hypodermal cell
soon after the ovular swelling becomes evident (Fig. 16). No parietal
cell is cut off; and so the archesporial cell becomes the megaspore mother-
cell. The megaspore mother-cell divides into two (Fig. 17), and then
four (Fig. 18), forming a hnear tetrad of megaspores. The innermost
megaspore of the tetrad rapidly enlarges and the others rapidly degen-
erate. By the time of the first division of the megaspore nucleus the
outer megaspores appear as crushed degenerating masses (Fig. 19).
Divisions into four and eight nuclei follow quickly (Figs. 20 and 21).
Three antipodal cells are cut off and rapidly degenerate (Fig. 22).
The two synergid nuclei move to positions side by side in the upper end
of the sac. A vertical wall separates the two, then horizontal walls cut
them off as separate cells. (Figs 22 and 23). They rapidly degenerate
and, at the time of entrance of the pollen tube, appear only as crushed
remnants. The two polar nuclei move to the center of the sac and lie
in contact, closely appressed, until fertlization (Figs. 22, 25, and 26).
The egg lies in position immediately above the polars. Miss Riatt (39)
states that no antipodals were found in Inipatiens pallida. She states
that the synergids lie below the egg toward the center of the sac. In
her figure at the time of fertilization, when she says that no polar nuclei
are shown, the structures which she marks synergids resemble the polars
oi fulva.

There are some interesting points in the history of the nucellus.
The archesporial cell arises as a hypodermal cell. No parietal tissue
arises, so that never more than one layer of nucellar tissue overUes the
embryo sac. By the first division of the megaspore the nucellar cells
overlying the apex may show signs of shrinkage (Fig. 17). At the stage
of fertilization the cells over the apex are disappearing, and cells along
the side of the embry^o sac, never more than one layer in thickness, show
signs of disintegration. The cells below the embryo sac, generally three

FLOWERS OF I.UPA TIE.WS FULVA 169

layers, at an early stage become elongated and tracheid-like (Fig. 17).
After fertilization the sac begins a rapid elongation, crushing the
tracheid-like cells below, and apparently absorbing their contents for its
growth. The elongation continues to the bottom of the ovule, then
turns and grows a short distance in to the funiculus. At its greatest
development, the embryo sac is thus a long narrow structure sharply
cur\'ed at its lower end (Fig. 36).

The conditions in /. pallida seem to be in some points exactly the
same as in fulva. To judge from Miss Riatt's figures 1, J and K, and
2 L, the nucellus surrounding the embryo sac is one layer in thickness.
The cells below the sac are in three rows, which with the extension down-
ward of the lateral layers above make five rows. The cells show the
same elongated tracheidal form, and like fulva extend downward and
curve into the funiculus. The degeneration of the peripheral layer
begins apparently in the megaspore stage (Riatt, fig. 1 J). In the fer-
tilization stage the nucellus is apparently entirely gone at the sides and
micropylar end, and the embr\'o sac is against the jacket layer (Riatt
Figs. 2 0 and S).

The cleistogamous pistil is similar in development and structure
except for a certain reduction. It arises in the same manner at first
as three papillae and later as a ring of tissue. The septa advance from
the side walls to meet the central placenta, dividing the ovary into five
chambers. In the pseudocleistogamous flowers the five little teeth
surmounting the very short style are much reduced or even absent,
and in the cleistogamous flowers they are generally absent entirely.
The stylar canals are present as in the chasmogamous flowers. In the
pseudocleistogamous flowers the full number of ovules may be present,
but in the cleistogamous flowers these are reduced to three, two or one.
The reduction of the number of seeds in the cleistogamous capsules is
not due to failure of fertilization or subsequent degeneration, but to a
reduction in the number of ovules which arise. The archespofial cell
divides to form four megaspores, and the chalazal spore persists. Some
sections show only three megaspores apparently, through failure of one
daughter cell to divide; often, however, the four megaspores appeared
in a linear series. The embryo sac forms eight nuclei, and the antipodals
and synergids are cut off as in the chasmogamous flowers. The mature
sac contains the egg and below it the two polar nuclei. The history
of the nucellus and integuments is the same as in the chasmogamous
ovule.

As in other cleistogamous species, the pistil of the cleistogamous
flower differs from that of the chasmogamous chiefly by reduction^of

170 CARROLL— ON DEVELOPMENT OF

parts. In /. fulva the reduction affects the short style and stigma and
the number of ovules. In / noli-tangere five short styles per-
sist. The reduction of style and stigma is common in cleistogamous
flowers. Helene Ritzerow (41) lists some ten species where it occurs.
In Aspicarpa hirtella, Aspicarpa longipes and in Specularia a reduction
in the number of carpels occurs (41, p. 212). On the whole, however,
the pistil seems to suffer less reduction than do other parts of the flower
an cleistogamy,

VI. Growth of Pollen Tubes and Fertilization

On germinating the grains in sugar solution, four tubes begin growth
from the four corners of the somewhat barrel-shaped or elliptic pollen
grain. The behavior of the nuclei with respect to these four tubes was
not discovered. In the few cases examined of pollen grains germinated
on the chasmogamous stigma, each grain gave forth apparently only one
pollen tube. In the cleistogamous flowers, the development of the
pollen tube from the grain is more easily followed. Here only one tube
issued from each grain. The tubes pass directly down from the cleis-
togamous anthers to the opening of the stylar canals (Fig. 33). Through
these open canals the tubes pass in such numbers as to fill them. Issuing
from the lower end of the short stylar canals, the tubes pass in fairly
direct course to the ovules, where they may be seen in numbers clus-
tered about the micropyle. The inner integument about the micropyle
at this stage takes a vivid stain, suggesting the presence of some sub-
stance attractive to the pollen tubes.

The history of the tubes and their nuclei in the embryo sac is the
same in both chasmogamous and cleistogamous flowers. As the pollen
tube issues from the inner end of the micropyle, it appears in two or
sometimes three branches (Figs. 25 and 26). Miss Riatt (39) reports
similar branches in Impatiens pallida, and it is recorded for other genera.
As these branches pass down through the embryo sac, one nucleus in
each branch becomes much larger (Figs. 25 and 26). This is the func-
tioning male nucleus. In addition to this nucleus, there appear one or
two smaller degenerate looking masses of nuclear material. These are
apparently the remains of the degenerating tube nuclei. Finally they
break into fragments and are lost. One branch of the tube approaches
the egg, and the other approaches the two polar nuclei. The polars are
now closely appressed, but entirely distinct. The ends of the tubes
swell out and rupture, and the male nuclei pass out one to the egg and
•one to the polars.

FLOWERS OF IMP A TIENS FULVA 1 7 1

VII. The Growth of the Embryo

After fertilization, the primary endosperm divides before the fer-
tilized egg. By the first division of the fertilized egg the endosperm
nucleus has divided to form several nuclei. The endosperm nuclei
are much smaller in diameter than the egg nucleus. They become ar-
ranged around the edge of the sac, and move upward around between
the egg and the micropyle (Fig. 27). The first division of the egg is
transverse (Fig. 27). The upper of these cells develops the suspensor,
and the lower the embryo. The upper cell divides transversely and the
lower one longitudinally. In Figure 28 three cells of a four-celled em-
br}-o appear. The two upper cells are derived by the division of the
cell which gives rise to the suspensor. The lower cell of Figure 27 has
divided in the plane of the paper, one of the daughter cells appearing
in Figure 28, and one lying immediately behind appearing in the next
section on the slide. The next division is longitudinal and at right
angles to the last, giving rise to the quadrant stage of the embryo.
Figure 29 is a section through the embryo at the stage when the derma-
togen is cut off. The section is not central, so that only one of the
central cells appears. The outer layer of dermatogen cells is evident.
The suspensor at this stage is at first two cells in length (Fig. 29). This,
however, rapidly increases to five cells in length (Fig. 30). In the mean-
time, the lower two of the suspensor cells have divided longitudinally
(Fig. 31). Further division gives rise to two layers of several cells.
Then the third cell of the suspensor divides longitudinally to form a
plate of at first four cells. Meanwhile the divisions of the cells of the
embryo have continued. The plerome and periblem are differentiated.
Toward the micropylar end of the embryo these layers do not arise
from the original embryo cell. The three lower layers of the suspensor
continue to divide and form the periblem and dermatogen in this region.
The cal>'ptrogen is later added from the outermost of these layers. The
two outermost cells of the suspensor remain undivided (Figs. 31 and 32).
These elongate, push into the micropyle, and form the suspensor proper.
Figure 32 shows the embryo at the time of origin of the cotyledons.
A peculiarity of the seeds of the genus is the existence of four side
roots already developed before germination. Heinricher-^ reports this
condition in a number of the Graminaceae and on Goebel's authority
in the genus Cucurhita. Heinricher found it in all species of Impatiens
that he investigated: — /. glanduligera, Royle, /. scabrida, D. C, /.
noli-tangere, L., /. parviflora, D. C., /. bicornuta, Wall, and /. balsaminay

172 CARROLL— ON DEVELOPMENT OF

L. These four roots form a whorl, one pair falling in the median plane
of the cotyledons and one at right angles. In cross section these appear
as the arms of a cross. On germination these appear soon after the prim-
ary roots as little hooks and soon press into the ground. Heinricher's
figures show the condition in I. glanduligera and balsamina. Heinricher
states on Klebs' authority that in Impatiens noli-tangere at germination
the four side roots surpass the main root. Similar behavior of side
roots is seen in fulva. The side roots were not found, however, already
developed in the seed.

The reserve food stored in the cells of the cotyledons has been inves-
tigated in several species of Impatiens, particularly in I. balsamina by
Heinricher. In I. balsamina, capensis and others reserve food is stored
in the form of cell- wall thickenings. These thickenings Heinricher (21)
determined were not cellulose but of amyloid nature. The thickenings
disappear on germination, and starch grains appear in the cells. Sub-
sequently the cells become green and function as assimilative cells.
The cells of the cotyledons of other species of Impatiens contain small
protein grains and a large quantity of fatty oil (21, p. 167) with no
carbohydrate. In /. fulva, the cells of the cotyledons are thin- walled,
and densely packed with large grains of reserve food. This material
proved not to be starch, but its nature was not investigated. It may
be the fatty oil characteristic of other species of the genus.

The history of the endosperm presents some interesting points.
The primary endosperm nucleus divides quickly and divisions follow
in rapid succession. The nuclei move to the periphery of the embryo
sac. Some push around above the embryo and even more into the
micropyle (Fig. 27). The nuclei in the micropylar end of the sac become
cut off by cell walls (Figs. 28 and 29). These cells soon begin to show
signs of degeneration, and in later stages appear as small crushed masses
in the micropyle and in the micropylar end of the embryo sac (Fig. 30).
The nuclei on each side of the embryo a little later are cut oflf as cells.
Thus when the cotyledons begin to appear three rather irregular layers
of endosperm cells are to be seen at the sides of the embryo (Fig. 32).
As the embryo increases in size the peripheral layer of endosperm nuclei
elsewhere in the sac become cut off into cells. At the micropylar end
of the sac the cells become much elongated, forming a row around the
hypocotylar end of the embryo as if functioning as a suspensor to force
the embryo down toward the center of the sac. There seems to be
quite a tendency for cells to force their way to the micropylar end of
the sac or even into the micropyle. The synergids when cut off become

FLOW ERS OF IMPA TIEXS FULVA 173

more or less rounded and push up into the lower end of the micropyle
and there degenerate. Then the first endosperm nuclei that are cut oflF
move around the embryo and take positions close to and in the micropyle
where one or two endosperm cells may be seen cut off (Figs. 29 and 30).
These sooner or later degenerate. Then these suspensor-like endosperm
cells arise at the micropylar end and persist for a considerable time. The
peripheral layer increases to a double layer of cells lining the embryo
sac. When the embr\'o has completely filled the sac and the outermost
layer of the integument is beginning to take on the peculiar differen-
tiation of the seed coat, one of these layers may be still made out. It
disappears, however, before the ripening of the seed.

VIII. Development of Seed Coat and Pericarp

The behavior of the seed coats after fertilization is worthy of note.
At the stage of fertilization the outer integument has risen almost to
the level of the inner integument. The two integuments are distin-
guishable only for a short distance at the micropylar end. Below this
point the tissue is perfectly continuous; no dividing line can be found
between the two. This condition results from the growth of the lower
part not as two separate integuments that afterwards fuse, but as one
tissue that rises as a whole. The inner layer of the integument lying
next to the nucellus, and, after the destruction of the nucellus, against
the embryo sac, is composed of tabular cells forming a distinct jacket
layer (Fig. 28). Very soon after fertilization, degeneration begins in
the region about three layers below the jacket cells (Fig. 36). This
degeneration spreads inward toward the jacket layer and outward
toward the exterior. The jacket layer becomes shrunken to a thin
densely-staining layer surrounding the embryo (Fig. 35). The degen-
eration of tissue continues outward until only two or three layers remain
intact. Meanwhile at the micropylar end the outer integument has
grown beyond the inner (Fig. 36). The cells of the outermost layer of
the integument become much enlarged and extend outward to form the
characteristic ridges of the mature seed coat (Figs. 34, 35 and 36). The
cells gradually become filled with the tannin that is so characteristic
of the genus.

The early history of the ovary has been considered above, the devel-
opment subsequent to fertilization remains to be considered. The
septa have begun to show signs of degeneration in their inner layers
before division of the megaspore mother-cell (Fig. 34). This degenera-
tion continues rapidly until at fertilization the septa consist each of

174 CARROLL— ON DEVELOPMENT OF

the two outer walls alone. By the time the cotyledons have been
differentiated the septa are represented only by tags adhering to the
placenta and ovarian wall (Fig. 35).

Early in the history of the ovary, very soon after its appearance
and before the appearance of the ovule, there is a striking differentiation
in the hypodermal layer as in other members of the flower (Fig. 34).
This layer is composed of large, regular, squarish cells quite distinct
in shape and size from neighboring layers. The cells very early become
filled with tannin which adds to their prominence in a section. At
five points, opposite the centers of the septa, the large squarish cells
are replaced by two, or sometimes three or four small cells also filled
with tannin (Fig. 34). Below these small hypodermal cells a line of
small cells extends transversely through the ovarian wall to the inside.
These mark the lines of rupture. Some time after the total degenera-
tion of the septa, the innermost layer of the ovarian wall begins to show
signs of degeneration and shrinkage. This degeneration extends up
the small lines of cells that mark the lines of dehiscence. As growth
continues degeneration extends backward until it reaches the vascular
bundle in the middle of the valve and until it reaches the hypodermal
layer in the lines of dehiscence (Fig. 34). The epidermal cells by this
time have become filled with tannin, and are in a state of turgescence.
No mechanical thickenings appear. The bursting of the pod is con-
trolled apparently by turgescence of the hypodermal and epidermal
layers.

Conclusions as to Cleistogamy

Many of the attempts to explain the phenomenon of cleistogamy
have been based on teleological considerations. Knuth, (26) for example,
explains the appearance of cleistogamous flowers in Drosera rotundi-
folia by the assertion that insects are more attracted by the glistening
drops on the glandular leaves and do not visit the flowers. "Owing to
the continual capture of insect-prey open flowers are useless for the
sundew, and it therefore develops cleistogamous ones." In Oxalis
acetosella, the cleistogamous flowers are developed in June and July
because insects would not visit the rather inconspicuous flowers of
Oxalis when more showy flowers are in blooom, but in spring when the
plant does not have to compete with more showy forms the chasmo-
gamous flowers appear. Knuth explains the occurrence of cleistoga-
mous flowers in the Anonaceae on the authority of Loew as an adapta-
tion for protection against ants. Darwin (13, p. 338) found the explana-

FLOWERS OF IMPA TIENS FULVA 17 5

tion in " the production of a large supply of seeds with little consumption
of nutrient matter or expenditure of vital force. " Delpino explains
cleistogamous flowers as an adaptation to secure the production of
seed under favorable climatic conditions. Shaw, (45) after giving an
historical resume of the subject, places emphasis on the cleistogamy
as an adaptation to secure the rapid development of fruit.

The explanation of cleistogamous flowers as stages of arrested
development has been ofi'ered by various authors since the phenomenon
has been an object of investigation. Gray^ (18) as long ago as 1849, re-
garded the cleistogamous flowers of Impatiens fulva as buds arrested at
an early stage. Darwin agrees in part with the arrested development
idea, but offers some objections to it which argue for his view noted
above. Since Darwin's time, the conditions which bring about the
arrest have been determined through a wide range of observation and
experiment by various investigators. Temperature, toxicity of soil
attacks of parasites, reduction of leafage, moisture and light have been
demonstrated to be associated with the production of cleistagamous
flowers. Goebel (17) finds that all of these conditions affect the produc-
tion of cleistogamous flowers through the unfavorable conditions of
nourishment. He regards the cleistogamous flowers as "Hemmungs-
bildungen" of the chasmogamous flowers to be explained solely on the
ground of nourishment. The occurrence of the arrest in all stages
adds weight to the view. From the smallest cleistogamous flower with
no resemblance to the evolved z}'gomorphic chasmogamous type,
gradations may be shown through various sizes and degrees of pseudo-
cleistogamous flowers to the type which Hansgirg (20) originally gave the
name pseudo-cleistogamous, a flower similar in every respect to the
chasmogamous flower, but failing to open before fertilization. It is
but a step more to include in the series those showy flowers which nor-
mally are self-fertilized without opening or even to the expanded self-
fertilized species.

Burck (11) added a new thought to the investigation when he ascribed
the origin of cleistogamy to mutation. It is worth noting, with mutation
in mind, the variety of families in which cleistogamy is reported. The
list of families given by Knuth (26) in which cleistogamy occurs, shows a
range from some of the simplest to the most evolved flowers. Among the
Monocotyledons it occurs in the relatively simple Alismaceae, Buto-
maceae, Graminaceae, Pontederiaceae; in the more evolved Juncaceae
and Liliaceae; in the most evolved family, the Orchidaceae. Among
Dicotyledons, it occurs in the less evolved Caryophyllaceae and Portu-

176 CARROLL— ON DEVELOPMENT OF

lacaceae; in the more evolved Rosaceae and Oxalidaceae; and finally
in the Labiatae and Scrophulariaceae. Families lying between the
extremes show abundant occurrence of cleistogamy. It would seem
that if mutation is the sole cause of cleistogamy, it must have arisen
many times in these widely separated families. This multiple origin,
however, would not necessarily be an objection to the theory, as there
is abundant evidence of other plant structures appearing in divergent
groups of plants.

The evidence at hand is on the whole in support of Goebel's explana-
tion of cleistogamous flowers as a product of reduced nourishment.
Darwin thought the existence in cleistogamous flowers of altered struc-
tures especially adapted for cleistogamous fertilization argued against
the idea of arrested development (12). He mentions in particular the
altered style in Viola and the lack of style and the open stylar canals
in many forms. Bennett believes the difference between the two types
of flowers is original, but the grounds for his opinion are not safe. He
described a difference between the sepals of cleistogamous and chas-
mogamous flowers in the earliest recognizable bud. This, however, is a
comparatively late stage. The arrest may be effected in the primordia
of the flower, or perhaps the "arrest" or alteration in vital processes
may take place at even an earlier stage, before the primordia have
appeared at all. When alteration of structure does occur in the cleis-
togamous flowers it is generally a matter of the reduction of parts. The
petals fail to develop or develop only to a reduced size, the stamens lack
parts as in Impatiens, or their number is reduced even to one stamen,
the style is lacking and stylar canals remain open, the number of ovules
is reduced. These are plainly differences due to loss or reduction of
parts. Where deviation occurs in the line of apparently added struc-
tures, these structures are evidently due to such growth as occurs from
altered primordia. The conditions which cause these deviations have
been determined; the morphology of the mature parts has been described
in a great variety of forms ; some work has been done on the histological
developments of parts. Further exact work is needed to determine
the point at which histological differentiation first occurs, and to deter-
mine the exact physiological processes which result in histological
deviation in the two types of flowers. When there is more information
along these lines, further attempts to explain the biological significance
of cleistogamy will be in order.

FLOWERS OF IMP A TIENS FULVA 177

Technique

The method of handling experimental materials was described above
in reporting the experiments. For histological work, material was
fixed in chrom-acetic, Flemming's solution, and Davis's modification
of Renner's solution. The latter solution consists of: —

90% alcohol 300 cc

distilled water 200 cc

nitric acid 10 cc

bichloride of mercury to saturation

glacial acetic 5 cc

This gives a 60% solution which for use is diluted to 40% by the addi-
tion of water. The material is left in the solution from five to seven
hours. For embr)'o-sac and other delicate structures, the shorter time
is best. For embryo-sac and earlier stages of pollen development, this
solution proved superior to chromacetic, and for the histology of the
structures quite as good as the expensive Flemming solution. The
material was stained in Haidenhain's iron-alum haematoxylin, and in
safranin and gentian violet combination. Counterstaining the iron-
alum-haematoxylin with safranin or orange G proved advantageous
in some cases.

IX. Summary

I. Plants of the species at maturity vary in height from eight
inches to seven feet. Adverse conditions of dryness or toxicity of soil
bring about dwarfing. Optimum conditions for growth are abundant
water with good drainage and light shade or sun.

II. Chasmogamous flowers appear in late June and last till early
October.

III. Cleistogamous flowers appear in early June and last through
the summer. They are most abundant before the chasmogamous flowers
appear. They are to be found throughout the summer on plants under
adverse conditions, and on the lower, short side branches of many
plants under good conditions.

IV. Pseudocleistogamous flowers are morphologically chasmogamous
buds which are self-fertilized at various stages, from the size of
four millimeters to a mature flower about to open.

V. Humming-birds and bees are the agents of pollination. The
pollen is shed upon the bee as the insect forces its way to the nectary.
Later the whorl of stamens falls away and the stigma is exposed.

178 CARROLL— ON DEVELOPMENT OF

VI. The flowers are readily self- and cross-fertilized, and seed was
produced by cross-pollination between I. fulva and I. pallida.

VII. The size of cleistogamous and chasmogamous seed is the same.
The number in chasmogamous capsules is four or five; in cleistogamous
capsules one to three.

VIII. Drought, weak light, poor soil, excessive temperatures,
toxicity of soil and reduction of leafage inhibit chasmogamous flowers,
but cleistogamous flowers may be produced under these conditions.

IX. There are two views as to the morphology of sepals and petals.
According to the first view, there are three sepals and three petals, two
of the petals being formed by the fusion of four in pairs. According to
the second view, there are four sepals and two lobed petals.

X. The cleistogamous flower shows all the members that appear in
the chasmogamous flowers. The petals and sepals are forced off by
the expanding ovary as a cap, resembling in appearance the calyptra
of a moss.

XI. The chasmogamous stamens are coherent, with broad filaments
and with ligular appendages which are coherent above the pistil; the
cleistogamous stamens are separate, strapshaped, without appendages.

XII. The development of the pollen is peculiar in the degeneration
of plates of sporogenous cells.

XIII. The cleistogamous stamens have a very small number of
pollen grains. An endothecium is formed and dehiscence occurs, but
the pollen grains remain in the pollen sacs and send down tubes to the
stigma.

XIV. The nucellus is a single layer of ephemeral cells at the sides
and micropylar end; below the embryo sac it is composed of five rows
of elongate tracheid-like cells which disappear as the embryo sac en-
larges. The mature embryo sac is a long curved structure.

XV. There are four megaspores, the inner one becoming the
embryo- sac.

XVI. Eight nuclei arise in the embryo sac; three are cut off as
antipodal cells, and two as synergids. Antipodals and synergids are
ephemeral.

XVII. The pollen tube branches on entering the embryo-sac, each
branch carrying a male nucleus.

XVIII. The first division of the fertilized egg is transverse, the
micropylar cell gives rise to the suspensor, and the lower cell to the
major part of the embryo.

FLOWERS OF IMPA TIENS FULVA 179

XIX. The original suspensor cell divides to form five cells in a
line, the lower three subdivide to form the lower part of the embryo,
the two micropylar cells elongate to form the suspensor proper.

XX. The free nucleate stage of the endosperm is followed by a
mass of ephemeral cells at the micropylar end of the sac; elsewhere the
endosperm is a double peripheral layer of cells which disappear at the
maturity of the seed.

XXL The ovule is surrounded by one mass of tissue which only at
the micropylar end is divided into two integuments.

XXII. After fertilization, degeneration begins in the middle layers
of the integument and spreads in both directions.

XXIII. The dehiscence of the capsule is brought about by the
increase in size and turgor of h\T3odermal and epidermal cells, and the
separation of the wall along a line of small cells lying originally in line
with the septa.

180

CARROLL— ON DEVELOPMENT OF

1. Bateson, A.

2. Beal, W. J.

3. Baillon, H.

4. Bennett, A. W.

5. Bennett, A. W.

6. Bentham, G. &
Hooker, J. D.

7. Bernoulli, G.

8. Britten, N. L. &
Brown, A.

9. Bninotte, C.

10. Burck, W.

n. Burck, W.

12. Darwin, C.

13. Darwin, C.

14. Eggers, E.

15. Eichler, A. W.

16. Goebel, K.

17. Goebel, K.

18. Gray, A.

19. Hansgirg, A.

20. Hansgirg, A.

21. Heinricher, E.

22. Henfrey, A.

X. LITERATURE CITED

Effects of Cross-Fertilization on inconspicuous Flowers.
Ann. Bot. 1, 1888, pp. 255-261.

Fertilization of Flowers by Humming-Birds. Amer. Nat.
XIV, 1880, pp. 126-127.
Natural History of Plants, 1888.

On the Floral Structure of Impaticns Jiilva, Nutt., with
Special Reference to the Imperfect Self-Fertilized Flowers.
Jour. Linn. Soc, Bot. XIII, 1873, pp. 147-153.
Notes on c^eistogamous Flowers, chiefly Viola, Oxalis
and Impaticns. Jour. Linn. Soc. XVII, 1880, pp. 269-280.

Genera Plantarum I, 1862, p. 277.

Zur Kenntniss dimorpher Bliiten. Bot. Ztg., 1869, Bd.
XXVII, p. 17.

Illustrated Flora of the Northern States and Canada. 1897.

Sur. I'avortement de la racine principale chez une espece

du genre Impaticns. Comptes Rendus, CXXII, 1896,

p. 897.

Ueber Kleistogamie im weitern Sinne und das Knight-

Darwinische Gesetz. Ann. Jard. Bot. Buitenzorg VIII,

1890, pp. 122-164. Ab: Just's Bot. Jahr. XVII, 1890,

pp. 467-468.

Die Mutation als Ursache der Kleistogamie. Recueil des

Travaux Botanique Neerland., 1905, Vol. II Livraisons 1-2.

Effects of Cross and Self-FertQization in the Vegetable

Kingdom, 1876.

The Different Forms of Flowers on Plants of the Same

Species, 1877.

Kleistogamie einiger westindischer Pflanzen. Bot. Cen-

tralbl. VIII, 1881, pp. 57-9.

Bliitendiagramme, 1875.

Die kleistogamen Bliiten und die Anpassungstheorien.

Biol. Centralbl. XXIV, 1904. pp. 697, 737-753, 769-786.

Bot. Jahr, 1904, 2, p. 901.

Chasmogame und kleistogame Bliiten bei Viola. Flora,

Bd. 95, 1905, p. 234.

Genera Florae Americae boreali-orientalis. II, 1849.

Phytodynamischen Untersuchungen, 1889, p. 237.

Nachtrage zu meiner .\bhandlung, "Ueber die Verbreitung

der reizbaren Staubfaden und Narben, sowie der sich peri-

odisch oder bios einmal offnenden und schliessenden Bliiten."

Bot. Centralbl. XLV, 1891, p. 70.

Zur Biologic d. Gattung Impaticns. Flora, B. 71, 1888,

pp. 163-175, 179-183.

Morphology of the Balsaminaceae, Joum. Linn. Soc. Ill

1858, pp. 159-163.

FLOWERS OF IMPATIENS FULVA

181

23. Henslow, G.

24. Hoffman, H.

25. Kerner, A.

26. Knuth, P.

27. Le Maout, E. &
Decaisne, J.

28. Lindley, J.

29. Loche, A.

30a. Loew, E.
30b. Loew, E.

31. Lubbock, J.

32. Masters, M.
2)2). Meehan, T.

34. Meehan, T.

35. von Mohl, H.

36. MiiUer, H.

37. Payer, J. B.

38. Penzig, O.

39. Riatt, A. H.

40. Richard, A.

41. Ritzerow, Helene

42. Rossler, Wilhehn.

43. Rydberg, A.

44. Shaw, C. H.

45. Shaw, C. H.

46. Schenk, A.

47. Schively, A. F.

48. Trelease, W.

49. Warburg, O. &

On the Self-fertilization of Plants. Trans. Linn See. Bot.

series 2, L 1877, pp. 317-48.

Kulturversuche iiber Variation. Bot. Ztg. XII, 1883.

The Natural Histor>' of Phints, translated by Oliver, II,

1895.

Handbook of Flower Pollination, 1906. Translation of

Handbuch der Blutenbiologie 1898.

System of Botany, English translation, 1876.

Vegetable Kingdom, 1831.

Note sur un fait anormale de fructification chez quelques

Balsaminaces. Bull. Soc. Biol, de France. Tom. XXIII,

1876, pp. 367-369.

Bliitenbau und Bestaubungseinricht. v. Impatiens Roylei

Engl. bot. Jahrb. XIV. 1892, p. 166.

Bemerkungen zu Burck, 'Die Mutation als Ursache der

Kleistogamie.' Biol. "Centralbl. XXVI, 1906 ,p. 129.

Seedlings, 1892.

Vegetable Teratology, 1869.

Flowers of Viola and Impatiens. Proc. Acad. Nat. Scs.,

Phila., 1873, p. 101.

Relations between Insects and Flowers of Impatiens fulva.

Proc. Acad. Nat. Scs. Phila. 1894, pp. 53-59.

Einige Beobachtungen iiber dimorphe Bliithen. Bot. Ztg.

XXI, 1863, pp. 309-315, 321-328.

The Fertilization of Flowers, 1883. English translation.

Traite D'Organogenie Comparee de la Fleur, 1857.

Pfianzen-Teratologie, 1890.

The Development of the Ovule of Impatiens pallida, Nutt.,

Plant World, July, 1916.

Nouveaux Elements de Botanique, 1852.

tJber Bau und Befruchtung kleistogamer Bliiten. Flora.
98, 1907, p. 163.

Beitrage zur Kleistogamie. Flora 87, 1900. p. 479.

Balsaminaceae in North American Flora XXV, 1910.

Inflorescences and Flowers of Polygama polygama. Bot.

Gaz. XXVII, 1899, p. 1121.

Comparative Structure of Flowers in Poly gala polygama

and Polygama pancijlora, Contributions from the Botanical

Laboratory of the Univ. of Pa., II, pp. 122-148, 1904.

Handbuch der Botanik.

Concributions to the Life History of Amphlcarpaea monoica.

Contributions from the Botanical Laboratory of the Univ.

of Penna., I, 1892, p. 270.

Action of Bees toward Impatiens fulva. Bull. Torrey Club.

VII. 1880, pp. 20-21.

182 CARROLL— ON DEVELOPMENT OF

Reiche, K. Balsamlnaceae, 1895, in Engler and Prantl, Die Natiir-

lichen Pflanzenfamilien.

50. Robertson, C. Flowers and Insects. Ill Bot. Gaz. XIV, 1889, p. 300.

51. Van Ingen, G. A note: Bees Mutilating Flowers. Bot. Gaz. XII, 1887,

p. 229.

Plate LV
Fig. 1. Cross section of bud of chasmogamous flower with only bract developed.

xiio.

A — bract.
Fig. 2. Cross section of bud of chasmogamous flower with lateral and posterior

sepals. X125.

A — bract.

B — lateral sepal.

C — posterior sepal.
Fig. 3. Cross section of bud of chasmogamous flower with four sepals. X125.

A — bract.

B — lateral sepals.

C — posterior sepals.

D — anterior sepal.
Fig. 4. Cross section of bud of chasmogamous flower with stamens arising. X 100.

E — anterior lobe of petal.

F — stamen.
Fig. 5. Longitudinal section of bud of chasmogamous flower. X30.

A — carpels.

B — ligular appendage of stamen.
Fig. 6. Cross section of bud of chasmogamous flower showing origin of septa of

ovary. X40.
Fig. 7. Longitudinal section of buds of cleistogamous flower. X50.

A — bract.

B — posterior sepal.
Fig. 8. Longitudinal section of bud of cleistogamous flower showing origin of

stamens X95.

C — stamen.
Fig. 9. Cross section of mature cleistogamous flower. X40.

A — bract.

B — lateral sepal.

C — posterior sepal.

D— anterior sepal.

E — anterior lobe of petal.

F-^posterior lobe of petal.
Fig. 10. Cross section of anther lobe showing pollen archesporium. XIOOO.
Fig. 11. Cross section of anther lobe showing degenerating plates of pollen mother

cells. X125.
Fig. 12. Cross section of anther lobe in pollen tetrad stage. X 100.
Fig. 13. One nucleate pollen grain. X 1000.
Fig. 14. Pollen grain with tube and generative nuclei. XIOOO.
Fig. 15. Cross section of cleistogamous anther near maturity. X200.

FLOWERS OF IMP A TIENS FULVA 183

Plate LVI

Fig. 16. Ovule showing single archesporial cell and beginning of inner integument.
X300.
Fig. 17. Megaspore daughter cells.

Fig. 18, Four megaspores, X850.

Fig. 19. Embryo-sac two nuclei. Two degenerating megaspores. XIOOO.

Fig. 20. Embr>'o-sac with four nuclei. X 1000.

Fig. 21. Embrj^o-sac with egg and two polar nuclei; one synergid cut ofif above and
three antipodals below. XIOOO.

Fig. 22. Embry-o-sac with six nuclei. XIOOO.

Fig. 23. Embryo-sac with two polar and egg nuclei; two synergids cut off above.
X 1000.

Fig. 24. Origin of the second integument.

Fig. 25. Embryo-sac showing two polar nuclei; branched pollen tube with male
nuclei and remains of tube nucleus.

Fig. 26. Embryo-sac with egg and polars; branched pollen tube with male nuclei
and remains of tube nucleus.

Fig. 27. First division of fertilized egg; endosperm nuclei. X290.

Fig. 28. Three cells of a four-celled embryo; beginning of segmentation of endosperm ;
the jacket layer of the integument. X400.

Fig. 29. Embryo showing two-celled suspensor, and dermatogen, and the first cells
of the endosperm. X300.

Fig. 30. Embryo with five-celled suspensor; the first cells of the endosperm degen-
erating. X300.

Fig. 31. Embryo showing division of the basal suspensor cells. X450.

Fig. 32. Embryo with the cotyledons arising, mature suspensor and layers of en-
dosperm cells. X75.

Plate LVII

Fig. 33. Longitudinal section of a cleistogamous flower showing the forcing off of
the perianth cap, the pollen tubes and the stylar canals. X30.

Fig. 34. Cross section of ovary of chasmogamous flower showing the enlarged cells
of hypodermal layer, the smaller cells opposite the septa marking the
lines of dehiscence and the beginning of degeneration of the septa. X 124.

FiG. 35. Cross section of ovary of chasmogamous flower with seed near maturity
showing degeneration of inner layers of ovary wall, the enlargement of
epidermal and hypodermal layers of ovary wall, and development of
seed coats. X50.

F.G. 36. Longitudinal section of ovule with embryo showing degeneration of middle
layers of integument.

Bot. Contrih. Univ. Penn.

VoK IV, Plate LV

CARROLL ON IMPATIENS

Bot. Coutrib. Univ. Pcnn.

Vol IV, Plate LVI

18

22

26

23\^ 28

CARROLL ON IMPATIENS

Bot. Contrih. Univ. Pcnn.

Vol. IV, Plate LVII

Fig. 35

Fig. 36

Fig. 34

CARROLL ON IMPATIEN§

Fig. ii

THE COMPARATIVE HISTOLOGY AND IRRITABILITY

OF SENSITIVE PLANTS

BY

D. Walter Steckbeck, M. A., Ph. D.
With plates LVIII— LXV.

Thesis presented to the Faculty of the Graduate School

in partial fulfilment of the requirements for

the Degree of Doctor of Philosophy

Introduction

The term Sensitive Plants was applied by the earlier botanists to
certain forms that exhibit the remarkable phenomenon of "closing up
on the approach of man, as if they felt or in some way became aware
of his approach. " The application of the term should not be restricted
to comparatively few plants, for all plants— all living organisms— are
sensitive in that they respond to and are influenced in a variety of ways
by environal factors. The marked difference between "sensitives'"
and "non-sensitives" lies in the fact that the former have certain
mechanisms which enable them to respond relatively quickly to stimuli.
Such response usually results in a change of position of the plant or of
certain parts of it, and such parts are said to be irrito-contractile. The
degree of response, and the duration of the change of position, depend
on the character of the stimulus and the time during which it acts.

This investigation is restricted to a study of some of the sensitive
plants— those that have the special mechanisms— (pulvini) referred to
above. The leaves of these plants are of especial interest from the
irrito-contractile relation. A study of these, as well as the stems, was
taken up with a view of ascertaining whether any structural details
might aid in explaining the sensitivity of these plants, and whether
there is a correlation between structural peculiarities and relative sen-
sitivity.

The movements of leaves in response to various stimuli is observed
in many plants. This phenomenon has been studied since the time of
Pliny. The most sensitive movements, and those earliest observed are
the day and night movements. Sensitive leaves at night occupy posi-
tions other than those which they occupy by day, exhibiting a closing
and an opening movement. Such changes of position are known as

185

186 STECKBECK— ON COMPARATIVE HISTOLOGY

Nyctitropic movements; also called night sleep of plants, since the
rising and sinking of leaves and leaflets coincides with the nocturnal
sleep of animals and the phenomenon was interpreted earlier in this
sense.

Leaf sensitivity as usually understood is almost wholly confined to
dicotyledons, although there are several monocotyledons, g>^mnosperms
and pteridophytes that exhibit this peculiarity.

Darwin describes sleep movement of the leaves of one of the water
ferns, Marsilea quadrifoliata. Each leaf bears four leaflets, each of
which is provided with a well-developed pulvinus. During the late
afternoon, the leaflets move upward and fold upon each other in such
a manner as to form a vertical packet. "When the leaves sleep, the
two terminal leaflets rise up, twist half around and come into contact
with one another and are afterwards embraced by the two lower leaflets. "

Among the monocotyledons, probably the best known example of
night turning of leaves is shown by various species of Maranta. During
the day the rather large blades of the leaves are in an approximately
horizontal position, while at night they stand vertically.

Amongst dicotyledons, two families stand out prominently as exhib-
iting the most specialized types of leaf sensitivity, these the Leguminosae
and the Oxalidaceae, two families that in sensitive relations, in struc-
tural details, and in distribution show many similarities. Both contain
forms that are practically non-sensitive with all transitions from these
to the very highly sensitive types. Both have compound leaves in
nearly all members, with well-defined primary pulvini at the bases of
the petiole, secondary pulvini in some and tertiary pulvini in those with
bi-pinnately compound leaves. Both families are essentially tropical
in distribution of the sensitive types.

This investigation is confined to the Leguminosae and the Oxali-
daceae, for practically all sensitive plants, as the term is usually applied,
belong to these two families. Other sensitives, such as Drosera and
Dionaea, are omitted. These have been studied very extensively by
Darwin, Macfarlane and others.

The work was suggested by Professor John M. Macfarlane, of the
University of Pennsylvania, to whom I want to express my deep grati-
tude for his advice, aid and criticisms.

Historical Review
The phenomenon of leaf sensitivity and leaf movement has proved
of botanical interest since the days of the early Greeks. From the time
of Pliny, who made studies of such movements, the historical records

AND IRRITABILITY OF SENSITIVE PLANTS 187"

of leaf sensitivity indicate no definite accounts until the eighteenth;
century. During this century a number of observations on sensitive-
movements were published. Among the early observers were Mairan^
Bonnet, Linnaeus, Acosta, Alpinus, Ray, Hill, Du Hamel and others.
Most of the studies were made on the sensitive plant, Mimosa pudica,
which had then been introduced into Europe and aroused much interest
on account of its leaf movements.

Miller (35) recognized five species of Mimosa, all more or less sensi-
tive. "The first sort (M. pudica) is commonly known by the name of
Sensitive Plant, to distinguish it from the others, which are generally
called Humble Plants, because upon being touched, the pedicel of their
leaves falls downward, whereas the leaves of the other sort are only con-
tracted upon the touch. "

Hill (20) carefully observed the sequence of motion in Mimosa pudica
and pointed out that the effect of total darkness on the plant is greater
than the rudest touch. He also found that the contact stimulus must
be of a sufficient intensity, and that the degree of the subsequent motion
depended upon the potency of the stimulus. Hill further observed that
the movements of Mimosa are less well marked at a lower temperature
than that in which the plants have been reared. His explanation for
this more sluggish movement is — "This is probably due to the juices
stagnating in the clusters of fibres, and to the contraction of the bark
by cold." Hill's explanation of the response to the contact stimulus
is interesting because it is an illustration of the view current at the time
that such motion was due to the fibres which acted like those of muscle.
Hill also made observations on the movements of the leaves of Abrus
and Tamarindus, and found that "In these and all others, the degree
of elevation or expansion in the lobes, is exactly proportional to the
quality of the light, and is solely dependent upon it." Hill's work was
severely criticised by his contemporaries.

Lindsay (24) studied especially the fall of the leaf of Mimosa when
stimulated, and discovered that the force which raises the petiole exists
in the lower part of the intumescence (pulvinus), and that which
depresses it in the upper. He seems to have considered that the tem-
porary excess of force in either part is produced by an impulsion of the
sap from the vessels of the yielding portion unto those of the opposite
portion.

Dutrochet (16) concluded that the sleep movements were due tO'
opposite changes of the energy of expansion in the antagonistic halves
of the pulvinus. This writer was the first to show that stimuli are

188 STECKBECK— ON COMPARATIVE HISTOLOGY

conducted through the vascular bundles of Mimosa pudica, and con-
cluded that the transmission was due to a pulsation of water. Du-
trochet, as well as some of the succeeding observers, attempted to com-
pare propagation in vegetable tissue to that by nerves in animals.

Burnett and Mayo (8) described the structure of the leaves of Mimo-
sa pudica and their movements, when stimulated or when the plant
enters the condition of night sleep. These observers noted the flush
changes which pass over the primary and secondary pulvini when the
leaves are stimulated. They believed that the under half of the pri-
mary pulvinus of Mimosa is more irritable than the upper.

De Candolle (13, p. 647) showed that in many cases even the coty-
ledons of sensitive plants are irritable, as illustrated in Mimosa pudica.

Meyen (33) agreed with Dutrochet that the propagation of stimuli
in sensitive plants is effected by the water movements in the elements
of the xylem. This same view was later emphasized by Sachs, Pfeffer
and others.

The structural details of pulvini were probably first described and
figured by Sachs (40, p. 793).

Briicke (7) studied the changes in the turgor relation of the pulvini
during stimulation, and recognized that the curvature of the pulvinus
of Mimosa pudica was connected with the flacidity of the responsive
half of the pulvinus produced by the escape of water.

Millardet (34) and Bert (2) concluded that the changes of expansion
in the antagonistic halves of the pulvinus were alike in both halves, but
differed quantitatively, and also in progress of time.

Bert studied the rate of propagation of stimuli in sensitive plants.

Pfeffer has contributed a great deal to our knowledge of irritability.
His "Die periodischen Bewegungen der Pflanzen" is one of the most
important of his works (37).

Darwin (11) made very extensive studies on the sensitivity of plants.
He was the first to apply the term " Nyctitropism " to the sleep move-
ments in plants. The irritable movements of the cotyledons of various
sensitives were emphasized by Darwin.

Gardiner (17) believed that in all irritable organs the movements
are brought about in consequence of a definite contraction of the pro-
toplasm of the irritable cells and that during such contraction some
of the cell sap escapes to the exterior.

Haberlandt (18) put forth the view that the stimuli are carried by
special cells in the phloem of the vascular bundles. Hanstein before
Haberlandt came to the conclusion that the sieve tubes are structures

AND IRRITABILITY OF SENSITIVE PLANTS 189

similar to nerves. Haberlandt considered the carriers of stimuli to be
special cells that are found only in sensitive plants. He described these
cells as being 6 to 4.2 mm. long and of an average width of .18 mm.
The walls of these cells are thin and are dotted; the dividing end walls
are finely porous and penetrated by plasma threads. The cell contents
consist of a thin plasmatic peripheral layer with a very large, round, or
elongated nucleus. The contents of each cell contain mucilage, glu-
coside and resin. These cells, according to Haberlandt, are interwoven
with the vascular bundles of stem, leaf, petiole and pulvini. The whole
conducting system acts as one in which intercommunication exists, and
one that is fused and through which hydrostatic waves pass and propa-
gate stimuli.

Borzi (4) emphasized certain structural details of various sensitives —
species of Mimosa, Aeschynomene, Neptunia — as aiding in the explana-
tion of sensitive relations. He described (1) the structure of the peri-
pheral regions — epidermis, hairs, etc. — as the parts that perceive sti-
muli directly; (2) The deeper conducting regions; (3) The pulvini as
the irrito-contractile centers.

Mac Dougal (27), in his studies on Mimosa pudica, showed that
Haberlandt's explanation of transmission of stimuli by the special cells
of the phloem was not tenable. Stems from which the phloem region
was removed were still able to transmit stimuli. "Excluding the
hydrostatic theory of Haberlandt, at present it seems necessary to
assume transmission by the tissues of the entire cross section."

Mac Dougal (28, 29) also studied the propagation of stimuli in
Biophytum sensitiviim. In this work he came to the conclusion that
the path of the transmission of the stimuli is the parenchyma of the
fibrovascular bundles and that the impulse is conducted plasmatically.
He says "it seems quite possible that the protoplasm action plays a
part in carrying the impulse from the point of reception to the motor
organ, and that while hydrostatic disturbance does not constitute an
impulse it may play a minor part in the transmission."

Nemec (36) concluded that stimuli are propagated by the plasma
membranes of the cells, and by the intercellular connecting fibrils which
come in contact with the plasma membranes.

Macfarlane (30) observed the flush changes in the tertiary pulvini
of various sensitives, such as Mimosa pudica, M. lupulina, M. sensitiva,
Schrankia angustata. "After stimulation of a leaflet and toward the
close of the latent period a sudden flush travels centrifugally across the
surface of the pulvinus. Immediately thereafter the leaflet contracts

190 STECKBECK— ON COMPARATIVE HISTOLOGY

and the pulvinus previously of a whitish hue assumes a dull greenish
aspect. This color change is due to migration of liquid into the upper
region of the pulvinus."

Cunningham (10) fully described the pulvini of various sensitives —
mainly Mimosa pudica — as motor organs. He states "The idea is
erroneous, that in the opposing masses of tissue in pulvini, we have to
deal with differences depending solely on invisible molecular structure,
and not on the presence of any visible difference in organization.

Haberlandt (19) considered that the epidermal hairs found on sen-
sitive plants especially those on the pulvini of Mimosa and the leaflets
of Biophytum are receptors of stimuli. Experiments are given in support
of this view. Shortly before Haberlandt's paper, Renner (39) had
emphasized the presence of various kinds of hairs on sensitives.

The different works of Bose (5) that have appeared during recent
years have added much to our knowledge of Plant Irritability. By
means of very ingenious devices he has recorded accurately the various
movements of some of the sensitive plants, especially Mimosa pudica
and Desmodium gyrans.

Recently Linsbauer (25) made observations on propagation of stimuli
in Mimosa pudica. He writes that propagation of wound stimuli
(traumatic) can take place for considerable distances in the stem, even
if all of the tissue outside of the wood is removed. He denies the pre-
sence of special conducting elements in the phloem.

The above brief historical review includes only some of the more
important references on irritability. Other references will be quoted
in the body of this work.

Distribution and Relative Sensitivity

Leguminosae

This family is very widely distributed in both hemispheres, from
the cold, almost frigid, regions to the tropics, where a large number of
genera grow in great profusion.

In the colder temperate regions, only a few members are found that
are sensitive to any degree, and such sensitiveness is shown only in
response to light stimulation, that is to nyctitropic stimuli and to the
effects of rather intense sunlight. There is no response, or a very slight
response, to other kinds of stimuli, such as mechanical, chemical, ther-
mic, electric and other forms. Species of Lathyrus, Vicia and Amorpha
extend northward to the Arctic Circle. The northern limit of some

AND IRRITABILITY OF SENSITIVE PLANTS 191

species of Trifolmm, Lotus, Astragalus, Phaseolus, Strophostyles and
Petalostemum, and also additional members of the three genera previously-
mentioned, is at a somewhat lower latitude. The leaves of these forms
show the day and night novements, some of them, like Trifolium, to a
very marked degree. Darwin (p. 349) examined eleven species of
Trifolium and found that nearly every one showed nyctitropic move-
ment in its leaves and also in the cot}dedons in some of them.

The paraheliotropic response is well illustrated in some of these
more northern representatives of the family. In Strophostyles helvola,
the trailing wild bean, of eastern North America from Quebec south-
ward, the leaflets of each of the tri-foliate leaves turn their edges toward
the sun when the light is intense and the temperature approaches 32° C.

In the warm temperate regions an increasing number of Leguminosae
show sensitivity. In a climate such as ours, with its warm, or almost
sub-tropical summers, with temperatures not unusually during the day
of 32 to 40° C, or even higher, and generally with quite a sudden drop
of temperature after sunset, a marked effect on the vegetation is shown.

When there is a drop in temperature, it means that some plants
must provide against the too rapid radiation of heat. This provision
the leguminous plants show in their capacity to close the leaflets at
night. Practically all the plants of this family in our region show this
phenomenon. Of those which are native, or introduced here, are species
of Cassia, Gleditschia, Gyvinocladus, Lupinus, Phaseolus, Medicago,
Melilotus, Amorpha, Tephrosia, Robinia, Astragalus, Aeschynomene,
Desmodium, Clitoria, Apios, Strophostyles, Pueraria, Amphicarpaea, etc.

All of these are sensitive to nyctitropism, and nearly all to para-
heliotropism in that the leaflets turn their edges toward the sun when
the light intensity gets above a certain optimum which varies with the
different plants. Many of these plants also are sensitive to mechanical
and other forms of stimuli.

Nyctitropic response is very well seen in the genus Cassia (Plate
LVIII, Fig. 1-2), that is tropical in its distribution and is represented here
by three outlying species. Night sleep of leaves is beautifully shown
by the arborescent members mentioned above, — Gleditschia, Gymnoda-
dus and Robinia. The last of these, the black locust, shows the pheno-
menon to good advantage. Shortly before sunset the leaflets, which
during the day, if the temperature does not exceed 32° C, are in a
horizontal position, begin to sink until in an hour or more they are back
to back, having described an angle of 90°. The day position of leaflets,
when the temperature rises to 32°-35°, or higher, is such that the edges

192 STECKBECK— ON COMPARATIVE HISTOLOGY

are turned toward the sun, so that the incident rays of light fall
parallel to the surface of the leaflets. The paraheliotropic movement
therefore is above the plane of the horizon instead of below as seen in
the nyctitropic response. The sensitive movements of the leaves of
Melilotus alba are described by Wilson and Greenman (47), and those
of other genera by Macfarlane (31).

In the sub-tropical regions the members of this family are abundant,
with many species of Cassia, Neptunia, Aeschynomene, Desmodium,
Melilotus, Acacia, Calliandra, Prosopis, Schrankia, Mimosa, Desmanthus,
Caesalpinia, Crotalaria, Indigofera, Sesbania, Arachis, Phaseolus, etc.
Practically all of these plants, as far as studied, are sensitive. In some,
the sensitivity has been developed to a high degree. The genera that
include some of the most sensitive plants of this family are here repre-
sented. A number of species of Mimosa, of Schrankia, of Prosopis and
of Calliandra appear for the first time. In the tropics numerous species
of these genera are found. In the sub-tropics many of the genera com-
mon in the temperature zones are also represented here. Some of these
extend over 80 degrees of latitude and are found in both hemispheres,
as for example, species of Desmodium which extend from south central
Canada on the north to sub-tropical Australia and New Zealand on
the south.

In the tropics the Leguminosae reach their climax of development,
comprising a large number of genera and species. In the western
hemisphere, Brazil is the center of distribution of the family. From
this region there have been obtained species of Mimosa, Schrankia,
Cassia, Inga, Pithecolobium, Calliandra, Bauhinia, Sweetia, Zollernia,
Exostyles, Caesalpinia, Peltogyne, Hymenaea, Parkia, Prosopis, Neptunia,
Aeschynomene, Abrus, Vigna, Dolichos and a number of smaller genera.
Some genera include a large number of species, — e.g.. Mimosa, more
than 300 species; Cassia, over 300 species; Calliandra, nearly 100 species.

The abundance of leguminous plants in the tropical vegetation, and
the very pronounced types of sensitivity exhibited by these, have been
noted by many botanists and travelers to the Amazon Valley and other
parts of the tropics. De Mello (15, p. 253) writes:— "The order of
Leguminosae is here (Campinas, Brazil) a most extensive one, repre-
sented by very many genera and innumerable species, from small herbs
to the tallest trees. "

Richard Spruce (44) in "Notes of a Botanist on the Amazon and
Andes" refers repeatedly to the abundance and great variety of dif-
ferent sensitive plants. He says (p. 257): "But of all orders by far the

AND IRRITABILITY OF SENSITIVE PLANTS 193

most abundant constituent of the flora of the Amazon is Leguminosae. "
(p. 258) "Sensitive plants, here called Sleepy Plants (Dormideiras),
are here so common that almost e\-erv' day I scratch my fingers or my
shins against some thorny member of the group. "

In reference to Mimosa, Wallace (46, p. 262) says: "Among the
more humble forms of vegetation that attract the traveler's notice none
are more interesting than the sensitive species of Mimosa. Where a
large surface of ground is covered with these plants the effect of walking
over it is most peculiar. At each step the plants for some distance
round suddenly droop, as if struck by paralysis, and a broad track of
prostrate herbage, several feet wide, is distinctly marked out by the
different colour of the closed leaflets. The true sensitive plants are all
low-growing herbs or shrubs with deUcate foliage, which might possibly
be liable to destruction by herbivorous animals, a fate which they may
perhaps escape by their singular power of suddenly collapsing before
the jaws opened to devour them."

In the Old World tropics many Leguminosae are found, but not so
abundant as in the New World. Some genera, like Mimosa, Cassia,
Desmoditim, Acacia, Prosopis, Parkia, Cynometra, Copaiba, Bauhinia,
Sophora, Crotalaria, Indigofera, Astragalus, Aeschynomene, Vigna and
Dolichos have representatives in the tropics and sub-tropics of both
hemispheres. Comparatively few genera are restricted to the Old
W'orld.

From the general survey on the distribution of the Leguminosae
it will be noted that a few members extend to the Arctic regions, e.g.,
Lathyrus maritimus; several species of Desmodium reach their southern-
most limit in Patagonia. More representatives of the family are found
in the temperate zones, with a gradual increase in the number of genera
and species toward the tropics where the leguminous plants flourish
in great abundance.

The increase of temperature is correlated with an increase of the
number of plants of this family and with the gradual advance in the
sensitivity of the different members that show this phenomenon. Nearly
the entire family shows night sleep of leaves. A considerable number
move paraheliotropically. Other sensitive responses have been evolved
gradually. A considerable number of leguminous plants of the tem-
perate regions respond to mechanical, electrical, and chemical stimuli,
but the movement is sluggish and not propagated for any great distance
from the center of the stimulation. The plants that show the best

194 STECKBECK— ON COMPARATIVE HISTOLOGY

response to the three methods of stimulation given are found in the
tropics. The forms with the most highly specialized sensitiveness to
the greatest number of stimuli are natives of the tropics, especially
in the New World.

Mimosa pudica, the best known of the sensitive plants, is native
to the Amazon Valley, but has been introduced into Asia and Africa,
so that sensitive plants now abound in many parts of the eastern and
western tropics. Mimosa Spegazzini, which is more sensitive than
M. pudica, is also indigenous to the South American tropics. These
two forms represent the acme of leaf sensitivity among the forms that
are in cultivation and that have been studied. There may be others
in the tropics that are equally sensitive, or even more sensitive, but which
have not been studied.

Why the climax in the number of species of sensitives and in the
high degree of irrito-contractile movements in the tropics? This can
probably be attributed to the climatic conditions peculiar to the tropics.
During the day, the temperature averages approximately 35° C. The
atmosphere is rather humid with rainfall every afternoon during the
wet season.

Bates (1) gives a very interesting account of the equatorial climate.
He refers to the shortness of the twilight and the consequent rapid
transition from day to night and from night to day. The deposition
of dew on the foliage at night is very heavy. The nights are cool. In
describing the daily rains he writes: "The heat increased hourly, and
toward two o'clock reached 93° F. The leaves which were so moist
and fresh in early morning, now became lax and drooping. On most
days in June and July a heavy shower would fall some time in the
afternoon. The cool sea breezes died away. The heat and the electric
tension of the atmosphere would then become almost insupportable.
The whole eastern horizon would become almost suddenly black. Then
the rush of a mighty wind is heard through the forest, swaying the tree
tops; a vivid flash of lightning bursts forth, then a crash of thunder,
and down streams the deluging rain. After the rain heaps of flower
petals and fallen leaves are seen under the trees."

The high temperatures by day, the sudden drop of temperature after
sunset, the rapid transition from day to night, the heavy dew, the
heavy daily rains with the accompanying winds, and dark skies (and
the presence of browsing animals, as suggested by Wallace), all these
factors aid in explaining the advantage resulting from closure of the

AND IRRITABILITY OF SENSITIVE PLANTS 195

leaves of the sensitives, and these further explain why so many of the
Leguminosae have acquired this phenomenon of leaf movement.

IVIany genera probably once sensitive, now, owing to colder condi-
tions as the plants were farther and farther removed from the tropics,
have non-sensitive, or slightly sensitive species. Other genera that are
distributed through colder regions and that are sensitive to day and
night changes have not acquired sensitivity to ordinary stimuli, or may
have lost sensitivity.

Oxalidaceae
This family in its distribution and sensitive relations, as well as in
structural details, shows many similarities to the Leguminosae. The
family is much smaller than the one previously described, comprising
seven genera with about 250 species. The Oxalids are found in both
hemispheres, with the largest number in the Western Hemisphere,
extending from Northern Canada to the southern extremity of South
America. In the Old World the range is from northern Europe and
Siberia to New Zealand in the Southern Hemisphere.

Oxalis Acetosella is the most northern representative of the genus
and of the family, being found in northern and middle Europe, eastward
through Siberia and south to the Himalaya Mountains. In the New
World, this plant is distributed throughout north central Canada, in
eastern Canada, south in the mountains to North Carolina. It has long
been a favorite plant for study, since it shows very striking nyctitropic
movements of its leaflets. It also shows slight paraheliotropic response,
and response to other stimuli.

In our region, seven species of Oxalis are found; two of these — 0.
corniculata and 0. strida — are very common. The leaflets of the dif-
ferent species move nyctitropically and paraheliotropically, but exhibit
very sluggish movements when stimulated by mechanical, chemical,
or other means of stimulation.

In the sub-tropics there is a marked increase in the number of species
of Oxalis, and other genersi— Bio phy turn and Averrhoa—a.re also repre-
sented. Biophytum dendr aides is found from central Mexico south-
ward. Averrhoa is distinctly a tropical genus, but both of the species
extend into some of the sub-tropical regions of Asia.

By far the largest number of species of Oxalidaceae is found in the
tropics of both hemispheres, with the majority in the New World. In
South America the genus Oxalis is represented by more than 200 species,

196 STECKBECK— ON COMPARATIVE HISTOLOGY

including the most sensitive members of this genus. Of those in cultiva-
tion such sensitive forms as 0. bupleiirifolia, O. scandens and 0. Ortgiesii
are native to the New World tropics. Comparatively few species of
Oxalis are found in the Eastern Hemisphere. The genus Biophytum
is essentially tropical in distribution in America, Asia, and Africa. B.
sensitivum, the most sensitive species of this genus that has been studied,
is found in both the Old and the New World. B. dendroides, scarcely
less sensitive than the above species, is indigenous to the American
Tropics. Averrhoa, a small genus, but including several species, e.g.,
A. carambola and A. bilimbi, that have been studied quite extensively,
is found wild or cultivated in the tropics of both hemispheres.

Practically all the tropical species that have been studied exhibit
irrito-contractility in their leaves. Not only are they sensitive to day
and night conditions, but when stimulated mechanically, chemically,
electrically or otherwise the leaflets close, dropping back to back, below
the plane of the horizon. The sensitivity varies for the different species,
and, as compared with the most sensitive members of the Leguminosae,
is less marked.

In the Oxalidaceae, as in the Leguminosae, the range of distribution
extends from the Arctic Circle on the north to the colder temperate
regions on the south, with by far the largest number of representatives
in the tropics.

The relative sensitivity shows a gradual increase from the species
found in the colder regions to those types that exhibit the highest degree
of sensitiveness and which are distributed throughout the tropics.

The probable reasons why the tropical forms of this family have
acquired sensitivity that is in advance of those in the cooler regions are
the same as those brought out for the Leguminosae. The two families
suggest many similarities. As indicated, the range of distribution is
practically identical with the climax in the number of species in the
tropics for both groups. In both we find compound leaves (a few excep-
tions in the Leguminosae), that are provided with primary pulvini at
the bases of the petioles, a secondary pulvinus at the base of each leaflet,
and if the leaf is bi-pinnately compound, as in Mimosa pudica, a tertiary
pulvinus at the base of each secondary leaflet. Both families show
similar response to stimuli.

In the Leguminosae and in the Oxalidaceae, the phases of stimula-
tion are alike. In both: (1) there must be an optimum environment in
order that response may be shown. Such environal factors as suitable

AND IRRITABILITY OF SENSITIVE PLANTS 197

temperature, sufficient light, a certain amount of oxygen in the sur-
rounding atmosphere, and also the health and vigor of the plant used
in the experiment, all have a very important bearing on the response
to be obtained. (2) given that the environment of the plant is favorable,
if a specific stimulus, which acts as so much energ>-, is applied to irrito-
contractile parts of the plant, a response follows. (3) The response
of the irrito-contractile parts is not immediate, but a time interval,
the latent period, elapses before there is a change of position of any of
the organs of the plant. During this period, the stimulus is presented
to and perceived by the plant parts that are irritable. Owing to vary-
ing degrees of irritability in different organs of the same plant, or of
different plants, the duration of the latent period varies within quite
wide limits— from 1/4 second, or even less under optimum conditions
in Mimosa Spegazzini, to an hour or more in the feebly sensitive members
of the two families under discussion. (4) Following a certain time
interval after the stimulus has been applied response in the irrito-con-
tractile organs follows. This may be slow and steady or rapid and
extensive in the plant. There can only be response if the stimulus,
which has been applied to the plant, was of a certain definite intensity,
and if it acted for a sufficiently long period of time. A weak stimulus,
presented for a short time, may stimulate to a certain degree, but there
may be no response in the irrito-contractile organs. In such case there
has been an insufficient intensity of excitation to bring about response.
Response follows on applying the weak stimulus for a longer time, or
applying it a number of times to the irritable plant parts.

(5) Hence here a summation of stimuli is necessary to produce a
response. In many species of Leguminosae and of Oxalidaceae that are
very slightly sensitive a considerable number of stimuli are required to
bring about closure of leaflets. (6) The stimulus is propagated from
the center of stimulation to other parts of the plant; the distance
through which it is carried being determined by the intensity of the
stimulus, the time during which it acts, and the relative irritability of
the plant used. (7) After contraction of the irrito-contractile parts has
ceased, there follows the neutral period, during which there is no evi-
dent movement of the parts of the plant involved in the stimulation act.
(8) After the neutral period the contracted parts re-expand and assume
the original position; this phase is called the re-expansion period.

The eight phases just given comprise a complete stimulation act,
and are identical for the Leguminosae and the Oxalidaceae.

198 STECKBECK— ON COMPARATIVE HISTOLOGY

In illustration of the above eight phases, the following illustrative
example might be cited in a stimulation act for Mimosa pudica: — When
a mechanical stimulus in the form of a forceps pinch is applied to, say,
the tenth secondary leaflet from the apex on a left-handed primary
leaflet, an extremely short latent period ensues, followed by rise of the
stimulated leaflet and almost instantly thereafter by rise of the neigh-
boring leaflet. Thus irrito-contraction of the pair initiates movement
in the leaf. A propagation of the stimulus then travels in two direc-
tions; one more rapidly in an upward direction causing successive closure
of all of the leaflets to the tip; another more slowly in downward direc-
tion causing similar closure. Such propagation is effected under opti-
mum summer conditions in eight seconds to the terminal pair of second-
ary leaflets, and ten seconds to basal pair. The distances through which
the stimulus is propagated are equal. This, therefore, indicates that
the rate of propagation of stimuli is slightly higher in a centrifugal than
in a centripetal direction. When the stimulus reaches the base of this
primary leaflet a flush change passes over the pulvinus, followed by a
shght converging of the secondary petioles. Burnett and Mayo (8,
p. 81), Paul Bert (2, p. 12) and successive observers noted the flush
changes of the secondary and primary pulvini. The tertiary pulvini
also show these changes just before the pairs of secondary leaflets close.
Propagation of the stimulus is rapidly effected down the petiole to the
primary pulvinus, and a drop of the entire leaf follows in two to three
seconds after the change in the position of the secondary petioles had
taken place.

The stimulus next travels into the leaflet adjoining and almost at
the same time into the leaflet opposite. A short time thereafter the
stimulus reaches the remaining leaflet of the four. In all three leaflets
the secondary leaflets close, the stimulus moving toward the apex of
each leaflet. Thus in twenty-five to thirty seconds after the stimulus
had been applied all the secondary leaflets are closed.

The neutral period which now ensues may last on an average of three
to five minutes. The leaflets then re-expand and the whole leaf rises
to its original position. The re-expansion period is ten or eleven min-
utes. Thus are illustrated successively a stimulation act, latent period,
response of leaflets, propagation of stimuli to successively removed
centers, neutral period and re-expansion period.

Having noted the similarity in sensitive movement and in distri-
bution in the Leguminosae and the Oxalidaceae, the histological study

AND IRRITABILITY OF SENSITIVE PLANTS 199

of the leaves of certain members of these families will be taken up in the
succeeding chapters. The plants studied include forms that are feebly
sensitive, others that are sensitive to a marked degree, and others that
are very highly sensitive.

A very suggestive feature of histological interest in the Leguminosae
and the Oxalidaceae is the presence of crystals. These are found in
practically all members of these families. In the individual plant they
are present in leaves and stems.

The writer made a careful study of the crystal distribution in the
different sensitives, from forms that are scarcely sensitive to those that
exhibit the most specialized t>T3es of sensitivity.

Plant crystals have long been known and studied. Malpighi (32)
was the first to observe "crystal clusters" in plants.

Von Leeuwenhock (23) noted several forms of crystals and described
these briefly.

Scheele (41) was the first to make a chemical investigation of plant
crystals, and indicated that they are composed, for the most part, of
calcium oxalate.

These earlier investigations of crystals were made on plants not
included in the families here studied.

Meyen {33) was probably the first one to study rather carefully the
crystals found in the cortex of the stem of Robinia pseiidacacia. The
crystals were studied from the chemical standpoint. He agreed with
Scheele that the crystals are composed of calcium oxalate. Meyen
observ^ed that the crystals are always contained in cells, and not in
intercellular spaces, as the preceding workers beheved. In this work
no reference is made to the distribution of crystals, but their forms and
shapes are described.

Holzner (21) takes up a general study of crystals, with brief reference
to those of Robinia pseudacacia. This work is rather important in
that it includes a summary of all the literature on crystals up to 1864.
Poulsen (38) briefly describes the crystals found in various papiliona-
ceous plants, such as Apios, Phaseolus and Dolichos. In these, accord-
ing to Poulsen, crystals are especially abundant in the characteris-
tically swollen bases of the petioles (pulvini) and in the thickened
flower stalks. They are not found in the epidermis but in the parenchy-
ma tissue beneath, in the parenchyma cells of the fibro- vascular bundles
and in the central pith-like tissues. Those in Phaseolus, he says, are
large, single crystals which have no definite relation to the axis of the
organs. They are elongated, prismatic, rhombic in cross section, and

200 STECKBECK— ON COMPARATIVE HISTOLOGY

are invested by a cellulose mantle which is usually quite thick. When
the crystals are too small to touch the cell walls, the cellulose envelope
at the ends of the crystals is prolonged into cellulose beams which are in
contact with the cell membranes.

Coester (9), in his study of Mimosas, states that calcium oxalate
crystals are deposited in all plant parts of these sensitives. Two forms
of crystals occur— as individual twin crystals and as crystal clusters.
The former kind is found in all the forms examined, occurring as mono-
cHnic plates, or styloid-like crystals. Coester further says that the
presence of the crystals is generally associated with the vascular bundles.
Here they are enclosed in the parenchyma cells around the bundles.
The formation of crystals is so abundant that over the larger veins of
the leaves every cell contains a crystal. When viewed from the surface,
the vascular bundle appears to be covered by crystals, arranged in
regular series. The crystal system is usually found in connection
with even the smallest bundles. Crystals are found in the stems as
well as in the leaves.

Solereder (43, p. 266) refers to the wide distribution of crystals in
the Leguminosae, and says two types are present — crystal clusters,
which are very rare, and soHtary crystals, which have either the ordinary
rhombohedral shape, or that of small rods or styloids. "The ordinary
rhombohedra occur especially in the tissue accompanying the vascular
bundles of the veins, in the chambered fibers of the bast and wood, and
also in the primary cortex. Those crystals which are rod-shaped, or
resemble styloids are especially peculiar to the mesophyll, but occasion-
ally occur in the epidermis of the leaf, in the crystal-containing cham-
bered fibers of the bast and also accompanying the sclerenchyma of the
veins. They are, strictly speaking, not solitary, but hemitropic crystals,
consisting of two or more individual crystals, arranged with their longi-
tudinal axes in approximately the same direction."

Solereder (43, p. 171) also refers to the crystals of the Oxalidaceae,
as occurring in the form of raphides, clustered crystals and solitary
crystals, but says nothing of the distribution of the crystals in the
different members of the family.

Leguminosae
Piieraria Thunbergiana
This plant is native to the tropical and sub-tropical regions of Asia,
but grows very well in the temperate regions, where it has been intro-

AND IRRITABILITY OF SENSITIVE PLANTS 201

duced as a hardy, perennial twiner. It bears large trifoliate leaves
that show paraheliotropic and nyctitropic responses to a marked degree.
When the temperature rises to 32° C. or higher, the leaflets turn their
edges toward the sun. On account of the huge size of the leaflets, often
7 or 8 inches long, and 4 or 5 inches wide, the paraheliotropic effect is
ver\' striking. The night sleep of the leaflets is equally prominent. The
leaflets drop in the nyctitropic movement.

Leaflets. The cr\'stals in the leaflets are of the styloid type (Plate
LIX Fig. 3) and are mainly distributed along the veins, with a few
scattered through the mesophyll tissue. The cr\'stals along the veins are
arranged in broken lines, with rather wide gaps between the crystals.
Each crystal is rod-shaped with short projections at each end, at right
angles to the long axis of the cr\-stal. The average dimensions of the
cr}'stals are 11.4 microns x 5.65 microns.

Pulvini. Crystals are found in the primary pulvinus and in the
secondary pulvini of each leaf. The type of crystal is similar to that
found in the leaf (Plate LIX, Fig. 4). The greater number of crystals
in the pulvini are present just around the vascular bundle cylinder, with
scattered crystals throughout the cortex from the bundles to the epi-
dermis. Here and there a cr>^stal is found with a faint line across its
middle portion, at right angles to the long axis. This is the beginning
of the twin-cr\stal t>-pe so characteristic of the more sensitive plants.

Petiole. The rod-shaped, or slyloid-like cr\'stals are also present
in the petioles, but in smaller numbers relatively than in the pulvini.
In addition to these crystals there occur in the petiole a second type, a
rhombohedral form. The latter kind of crystal occurs along the bun-
dles in the endodermal region. Here the crystals are arranged in
discontinuous lines. The individual crystals are considerably smaller
than the cells in which they occur, each crystal occupying about one-
half of the lumen of the cell in which it is contained. The longer axis
of these cr>^stals is parallel to the axis of the vascular bundles.

Bauhinia diphylla

The genus Bauhinia belongs to the sub-family Caesalpinioideae of
the Leguminosae, comprising about 150 species, distributed throughout
the tropics and the sub-tropics of both hemispheres. B. diphylla is a
native of India, growing at rather high elevations. This plant, which
is often grown in green houses, bears simple, deeply two-lobed leaves.
The single leaf probably represents the terminal leaflet of an originally

202 STECKBECK— ON COMPARATIVE HISTOLOGY

compound leaf. At the base of the petiole there is a primary pulvinus,
and at the base of the lamina a secondary pulvinus, both of these only
slightly thicker than the petiole. The leaves are very feebly responsive
to mechanical stimuh. Only after numerous stimuli (as many as 15
or 20) do the two lobes of each leaf rise through a small angle. Marked
nyctitropic response is shown, the lobes at night moving upward
through an angle of 80 to 90 degrees. There is also a slight drop of
the entire leaf from the day position.

Leaf. Scattered throughout the mesophyll of the leaves are a
number of conglomerate crystals (Plate LIX, Fig. 5). In addition to
this type there are prismatic crystals arranged in discontinuous single
rows along the veins. This latter type shows considerable variation
in the shape of the individual crystals. Some are irregularly rounded
and in general outline resemble the scattered conglomerates; others are
regular prisms, usually four or six angled. The hexagonal form is
rather rare. The distinct cross-line, so well seen in crystals of the more
sensitive plants, is indicated very faintly on a few of the six-sided
crystals. The crystals are small, rarely more than 9 microns long.
The breaks in the hnes of crystals are shown (Plate LIX, Fig. 5).

Pulvini. In both the secondary and the primary pulvini conglomer-
ate crystals are scattered through the cortex, but none along the vascu-
lar bundles.

Petiole. In the petiole lines of crystals occur regularly in the endo-
dermal cells surrounding the vascular bundles. In some cases only a
few crystals occur, or there may be twenty or even more crystals in a
line, then a break, then another line and so on. The crystals are rather
small as compared with the size of the cells in which they are found.
A thin protoplasmic sheath surrounds each crystal.

Gleditschia triacanthos

The genus Gleditschia is essentially tropical and sub-tropical in its
geographical distribution, being represented in tropical Africa, tropical
and sub-tropical Asia, sub-tropical Argentina, and in eastern and
southern North America.

G. triacanthos, the Honey Locust is indigenous over a wide region
from southern Canada to Texas. The pinnately compound leaves
show slight nyctitropic movement. In no case has the writer observed
a drop of the leaflets of more than 30 degrees. Rather feeble parahelio-
tropic movements have been observed, and no response to other forms

AND IRRITABILITY OF SENSITIVE PLANTS 20S

of stimulation. It is quite probable that this outlying species of a
genus, that is tropical and sub-tropical in its distribution, has lost its
sensitivity. The protection afforded by a closure of the leaflets is not
necessary in this plant. The small leaflets are sufficiently protected
against a too rapid radiation of heat by the well developed cuticle.
Leaflets. Broken lines of cr>'stals abound along the veins of the leaflet.
One to six or eight lines of cr>'stals are present, according to the size
of the vein. In shape the crystals are quite regular, mostly four-angled,
and relatively few six-angled. The size of the crystals is extremely
various, from the smallest ones less than 3 microns long, to the largest
ones, 15 microns long. The cr>'stals show no definite polarity of
arrangement with respect to the direction of the vascular bundles. In
many cells the longer axis of the crystal is at right angles to the greatest
length of the cell containing the cr\'stal.

Pulvini. In the secondary pulvini, as well as in the primary pulvi-
nus of each leaf, are found the same types of crystals as in the leaflets,
but these are more scattered in their arrangement. There are few
conglomerate cr>'stals present in the cortex of the pulvini.

Petiole. The crystals of the leaflets, the pulvini, the midrib and the
petiole of the honey locust are very similar throughout in shape and
arrangement, but with a greater number of crystals in the leaflets and
in the petiole than in the pulvini. The cr>-stals are imbedded in a
proto-plasmic sac or envelope which is not as dense as the envelopes in
the more sensitive plants are, as is indicated by the rather hght stam
of the enveloping sheaths when protoplasmic stains are applied.

Gymnocladus dioica

Gymnocladus is a small genus, including two species, G. chinensis
of eastern Asia, and G. dioica (the Kentucky Coffee Tree) of eastern
North America, from southern Canada to Oklahoma. The bi-pinnately
compound leaves of the latter species show night sleep very strikingly.
When observed at night the leaflets are in a position that is almost
vertical, having moved downward. The movement and the position
of the leaflets at night are very similar to those of Robinia. The dis-
tribution of the two plants is almost exactly identical with a slightly
wider range for the Kentucky Coffee Tree. The paraheliotropic response
is not as marked as it is for the leaflets of Robinia. In the latter the
upward movement begins when the temperature reaches 24° to 27° C.
and when there is bright sunlight, but Gymnocladus requires a higher

204 STECKBECK— ON COMPARATIVE HISTOLOGY

temperature to initiate the movement. The maximum angle through
which the secondary leaflets of this tree move is less than that of the
black locust. The writer was unable to get any visible response in the
leaflets to mechanical stimuli, but noted a slight drop of the secondary
leaflets when 50% sulphuric acid was applied to the tertiary pulvini.
The primary pulvinus of the leaf is not well defined, appearing as a
slightly swollen, clasping leaf base, and shows no irrito-contractility
when stimuli are applied to it.

Leaf. The crystals in the secondary leaflets are found along the
veins and are arranged in discontinuous lines. A single line is present
along the smaller veins. With an increase in the size of the veins
there are introduced additional ranks of crystals until in the midrib
there are present 7 to 9 broken rows of crystals. The crystals vary
considerably in shape and size, with the four angled prismatic type in
greatest abundance, and among these relatively few hexagonal prisms.
The latter in some cases indicate a very faint line across the middle of
each crystal. The hexagonal forms are rather small, with an average
length of 12.3 microns and a width of 6.83 microns. Conglomerate
crystals -are present in the bundle of the midrib, but none in the meso-
phyU of the leaflets.

Pulvini. Two kinds of crystals are present in the tertiary puKini —
the quadrangular type noted in the leaflet bundles, and an abundance
of conglomerates The regular prismatic forms are found in the endo-
dermal region, the lines being continuous from the midrib bundles. The
conglomerate crystals are found in the cortex of the pulvini, where they
occur in greater number on the dorsal side of the pulvini, and are also
present in the vascular bundles.

In secondary pulvini the crystal relation is similar to that described
for the tertiary pulvini (Plate LX, Fig. 6), and the same is true for the
primary pulvinus, but here very few crystals are present. Of the regular
prismatic type, none indicated the hexagonal form.

Petiole. In the secondary petioles we again note the regular pris-
matic' crystals arranged in fairly continuous lines in the endodermis.
As in the leaflets, the shape of the crystals varies, with here and there a
hexagonal crystal present. Conglomerate crystals are also present in
the inner cortex and in the vascular bundles. These, like the prisms,
show variation in shape, with transitions toward the prismatic forms.
The crystal distribution in the primary petiole of each leaf is similar to
that in the secondary petioles, with fewer crystals relatively and more
breaks in the lines of crystals than in the secondary petioles.

AND IRRITABILITY OF SENSITIVE PLANTS 205

Robinia Pseudacacia

The black locust is native to the eastern United States, from Penn-
sylvania southward to Georgia, westward to Oklahoma, but has been
naturalized as far north as southern Canada. The imparipinnate leaves
of this tree show pronounced sensitivity to nyctitropic and parahelio-
tropic stimuli. At night the leaflets drop so as to be placed in an almost
vertical position. The movement under the effects of intense sunlight
is upward. In midsummer in bright sunlight when the temperature
reaches 27° C. or higher, the leaflets of the black locust are directed
toward the light at an angle to the vertical of 35 to 40 degrees. Later
in the day with a decrease in intensity of the sunlight and a gradual drop
in temperature the leaflets sink, until by 5 o'clock, as observed in a
number of cases, the leaflets were nearly or quite flat. Then there
followed a gradual sinking of the leaflets below the plane of the horizon,
by 7:30 or 8 o'clock they will have attained the night position, making
a very slight angle to the vertical. To mechanical or chemical stimuli,
Robinia is feebly responsive. If when the leaflets are expanded shock
stimuli be apphed to them, a slight drop will follow after a number of
stimuli have been given. Here then a summation of stimuli is necessary
to bring about response.

Leaflets. Along the larger bundles of the leaflets two or three quite
regular rows of crystals are found. Along the small veins there is a
single line of crystals that is broken here and there. The crystals are
mainly of two kinds; one that is small, quadrangular in surface view,
seldom more than 7.6 microns long and 6.5 microns wide. These
crystals lie near the middle of cells of considerable size, each crystal
fining less than one third of the cavity of the cell in which it is contained.
The other type of crystal is hexagonal is outline, 15.2 microns long and
7.6 microns wide. Occasionally a crystal is noted in the chains along
the bundles that shows a line across its middle. The two types of
crystals are found in the same line with rather more of the hexagonal
prisms. The individual crystals of both kinds show quite wide varia-
tions in shape and size. The crystals are imbedded in protoplasmic
sheaths or envelopes.

Pulvini. The secondary pulvinus of each leaflet shows rather a
wide cortical region in which crystals are distributed — relatively few
toward the epidermis and a considerable number just around the bundle,
in the endodermal region. The crystals are nearly all of the styloid or

206 STECKBECK— ON COMPARATIVE HISTOLOGY

rod shaped type with pronounced transversely projecting processes at
both ends. These processes project from both sides of the ends and
seem to be of a different composition from the elongated portion of the
crystal, for, usually they take a slight stain when protoplasmic stains
are applied to the crystal containing cells, while the rest of the crystal
remains unstained. Some of the crystals show the partition lines acrsos
the middle, as was indicated for some of the larger six angled crystals
in the leaflets. The styloid crystals, on an average are of about the same in
length as the hexagonal type but quite narrow, from 1.8 ot 5 microns
side. In addition to the elongated form few quadrangular crystals
occur in the pulvini. In the endodermal region the crystals show a
certain polarity of arrangement, in that the longer axes of most of the
crystals are in the same direction as that of the bundles.

In the primary pulvinus of the leaf the crystal distribution differs
from that of the secondary pulvini in the smaller number of crystals
present in the former.

Petiole. Both the elongated and the quadrangular types of crystal
are found in the petiole, with a larger number of the latter kind. Both
are arranged in discontinuous lines in the endodermal zone. The
styloid type is most abundant in proximity to the pulvini. (Plate LX,
Fig. 7).

Desmodium rotimdifolium

The large genus Desmodium, of more than 150 species, is widely
distributed throughout the tropics of both hemispheres. Representa-
tives of the genus are also found in the sub-tropical and temperate
regions. In North America, Desmodiums are found from southern
Canada southward. Twenty species are included in the range covered
by Gray's Manual. All of these species, as far as studied, show sensi-
tivity in their leaflets. The degree of sensitivity varies considerably
with the different species.

D. rotund i folium, of the four species studied, both from the crystal
relation and the relative sensitivity, can well be considered as the least
specialized. This species is found in the eastern United States from the
Canadian border, southward to Louisiana. It is usually found in rather
dry woods, where it creeps along the ground. The prostrate habit is
given by Macfarlane (30, p. 202) as a reason for the sluggish irrito-
contractiHty of this species as compared with other species of the
genus. Amphicqrpaea, Desmodium canescens and Desmodium panicu-

AND IRRITABILITY OF SENSITIVE PLANTS 207

latutn are all upright growers, and are therefore exposed in their leaflets
to the full effects of night cold and heat radiation from the tissues, and
I believe that this may largely explain why in evolutionary development
they have become much superior to Desmodium rotundifolium, whose
long, sucker-like shoots run along the ground, and give off leaflets that
nestle among the surrounding herbage."

D. rotundifolium shows night sleep in its leaflets and also parahelio-
tropic movement. The leaflets are slightly sensitive to mechanical
and thermal stimuli. Both of these forms of stimulation cause a drop
of the leaflets through a relatively small angle but only after a long
latent period.

Leaflets. Crystals are present in the endodermis of the orbicular
leaflets. The crystals, which are arranged in lines with occasionally
a small gap in the crystal cell continuity, are small prisms, mostly of
the six-angled t>'pe. The variation in shape and size of crystals is most
marked. The average size for a number of the cr}^stals is 9.5 microns
long and 5.7 microns wide; the average dimensions of the crystal cells
are 19.5 microns long and 14.2 microns wide.

Pulvini. In both primary and secondary pulvini irregular crystals
are scattered throughout the cortex. These vary in shape from forms
that approach the conglomerate type to others that are quadrangular,
and relatively few that are hexagonal, but more elongated and not as
wide as the hexagonal forms in the leaflets.

Petiole. The crystals are of the same kind and show similar distri-
bution to those in the leaflets.

Desmodium Dillenii

This species is very common in the eastern half of the United States.
From the few experiments made, the writer considers this plant to be
close in its sensitive action and relation to D. rotundifolium, but slightly
more sensitive, in that when the leaflets are stimulated mechanically, they
drop after a shorter latent period than in the species previously described.
The angle through which the leaflets drop is also greater than in the
former species.

Leaflets. The crystal distribution in the leaflets of this species
resembles rather closely that of Desmodium rotundifolium, but the crystal
lines along the veins show fewer gaps. The crystals are larger rela-
tively to the size of the cells in which they are contained. The hexa-

208 STECKBECK— ON COMPARATIVE HISTOLOGY

gonal prismatic form is more in evidence than in the former species,
and in the pulvini are fewer crystals than in D. rotundifolium.

Desmodium paniculatum

In distribution, in relative sensitivity, and in histological details
this species resembles very closely the two Desmodiums just described.
In the crystal sheaths of the leaflet bundles there is a decided increase
in the twin rhombohedral crystals. The line across the middle of these
six angled crystals is more evident in a greater number of crystals than
for any plant studied so far. The crystals on an average are consider-
ably larger than in D. Dillenii and D. rotundifolium.

Desmodium canescens

Of the native species of Desmodium studied, this is the most sensitive
in its leaf movements. Both the nyctitropic and the paraheliotropic
responses are most marked. When mechanical stimuU are applied
the leaflets drop after a rather long latent period. In one experiment
the median leaflet was stimulated by giving it a slight blow. After a
latent period of 80 seconds, it dropped through an angle of about 20
degrees; 90 seconds after the first stimulus a second was given, followed
by a further drop of the leaflet. A third and a fourth stimulus were
applied with a 90 second interval between them. In all as a result of
the four stimuli the leaflet dropped about 70 degrees. Dr. Macfarlane
(30) states that the leaflets respond to chemical stimuU. A 6% ether
solution applied to 8 median leaflets caused contraction in aU of them.
A drop resulted equally whether the ether was applied to the secondary
pulvini or to the general leaf surface. A latent period of 1 to VA min-
utes followed the application of the ether.

Leaflets. The crystals along the leaflet bundles are arranged in
fairly continuous lines, with few breaks in the continuity of the crystal
cells. The more uniform regularity in the shapes of the crystals and
their larger size, as compared with those in the other species of Des-
modium is very marked (Plate LX, Fig. 8). The majority of the crystals
are of the six angled prismatic type, with the partition line across the
middle of most of the crystals evident. Polarity of arrangement of
the crystals is more marked than in the other species of Desmodium.

Pulvini. Elongated, prismatic crystals are scattered throughout
the cortex in both primary and secondary pulvini. In the endodermal
region of the pulvini a greater abundance of crystals is present than

AND IRRITABILITY OF SENSITIVE PLANTS 209

toward the surface, and here in addition to the elongated type, the kind
so typical for the leaflets is noted.

Desmodium gyrans

Next to Mimosa pudica there is probably no sensitive plant that has
been studied more extensively than Desmodium gyrans. The autono-
mous movements of the lateral leaflets have been known and studied
for several centuries. These oscillating movements are often so rapid
that they may be readily followed with the naked eye. The larger
terminal leaflet performs well marked nyctitropic, as well as less notice-
able oscillating movements. The irregular twitching movements of
the lateral leaflets have given the name Telegraph "Plant to this species.

Leaflets. The crystals in the bundles of the lateral and the terminal
leaflets are very abundant, forming continuous lines. The number
of lines varies with the size of the veins with which they are associated.
The crystals are practically all of the very regular six-sided, prismatic
type. There is a greater abundance of crystals in the lateral than in
the terminal leaflets.

Pulvini. In both primary and secondary pulvini, the kinds of crys-
tals and their distribution in the cortex are similar to that described
for D. canescens. The elongated rhombohedral type of crystal is here
the dominant one.

Petiole. A great abundance of regular prismatic crystals is present
along the vascular bundles of the petioles. So great is the number of
these crystals that they form a broad continuous sheath over the bun-
dles. In nyctitropic response the petiole moves upward from the day
position.

Amphicar paea monoica

This common leguminous plant, the Hog Peanut, abounds in eastern
North America from Canada to Mexico. The thin tri-foliate leaves
are sensitive to Hght and to mechanical stimuli. The day and night
positions of the leaflets are fully described by Schively (42, p. 308).
Usually 4 to 5 mechanical stimuli are required to cause the leaflets to
drop through an angle of 70 to 75 degrees. Leaflets that have been so
stimulated will recover the original position in 13 to 16 minutes. The
latent period, the summation required, the response period, the neutral
period and the re-expansion period, are all about the same, or slightly
shorter, as for Desmodium canescens.

210 STECKBECK— ON COMPARATIVE HISTOLOGY

The crystal distribution in Amphicarpaea monoica is emphasized by
Schively (42, p. 358), who writes: — "In early development numerous
crystals appear in the cells forming the inner row of the cortex of the
stem. These ultimately constitute a distinct crystal sheath. The
crystals are somewhat prismatic in form, possessing an apparent parti-
tion across the middle. These twin structures occur also round the
vascular areas of the leaves, and also in the cortex of the pulvini. "

The writer compared very carefully the crystals in the first formed,
simple, opposite leaves, and the later formed pinnately trifoliate leaves,
and found the crystal system better developed in the compound leaves.
The crystals are more abundant, larger, and the crystal Unes are more
continuous than they are in the simple leaves.

From experiments made on the relative sensitivity of the two kind
of leaves, the writer found the simple leaves more sluggish in their move-
ments. These experiments were made May 14, when the simple leaves
were fully expanded, while the compound leaves were not fully devel-
oped. The younger condition of the compound leaves may explain
their apparent greater sensitivity, for younger leaves always show quicker
response than do those that are older.

Cassia chamaecrista

This species of Cassia and two others are northern outlying repre-
sentatives of a large tropical and sub-tropical genus, which includes
more than 300 species. C. chamaecrista, the Partridge Pea, is found in
sandy soil throughout the eastern United States from New England
southward and westward to Texas. This form is quite abundant in
New Jersey, where it is often found growing with the next species,
Cassia nictitans, from which it can be distinguished by its larger flowers.
The Cassias are remarkably sensitive in their leaves, representing the
most highly developed types of sensitivity among our native sensitives.
The leaves of C. chamaecrista are evenly pinnate with 9 to 15 pairs of
leaflets on each leaf. A short almost sessile gland, 1 mm. high, is present
on the petiole, close to the primary pulvinus. The top of the gland is
elliptical, or oval, in outline, slightly concave, and about 1 mm. long.
The primary pulvinus is well developed, as are the secondary pulvini
at the bases of the leaflets. The gland, as well as the leaflets and the
pulvini, are irritable. In the superficial cells of the gland the appearance
is as in the pulvini, to be described later — there being rather dense masses
in each cell (Plate LXV, Fig. 26).

AND IRRITABILITY OF SENSITIVE PLANTS 2 1 1

The night sleep of the leaves of the Partridge Pea is very pronounced,
the leaflets folding upward and the whole leaf drops through an angle
of 65 to 75 degrees. The paraheliotropic response is equally marked.
At a temperature of 21 to 23° C. the leaflets arc fully expanded; when the
temperature rises to 26 or 27° C. the leaflets close ^ 2; at 31 to 32° C.
the leaflets are ^ closed, and at 33 to 35° C. complete closure is observed,
with a drop of the entire leaf.

Not only are the leaves of C. chamaecrista sensitive to hght stimuli
but also to other forms of stimulation. WTien leaflets are stimulated
mechanically they begin to close after a latent period of 3 to 4 seconds,
at a temperature of 27° C. after the first stimulus the leaflets close
yi and the leaf drops through an angle of about 25 degrees, the move-
ment continuing during a period of 1 minute. A second stimulus is
then applied and after a latent period as before the leaflets close %
that is they will have moved through an angle of about 60 degrees;
the leaf drops through an angle of 15 degrees. On the application of a
third stimulus, a latent period of 2 to 3 seconds ensues, the leaflets
move through an additional angle of 20 degrees and the leaf drops 8
to 10 degrees further. The leaflets are now almost closed and additional
stimuli cause no change in their position. Here, as for Desmodium
canescens, a summation of stimuli is required, but fewer stimuli are needed
than in the latter plant to cause closure of the leaflets. The leaf re-
expands in 14 to 18 minutes. The leaflets are equally sensitive to
chemical stimuli.

That the gland on the petiole is an irritable center was shown in
experiments with certain chemicals by Professor Macfarlane (Lecture
Notes). When he placed a drop of turpentine or alcohol or 60 p.c. sul-
phuric acid on the basal gland of a leaf with 13 pairs of leaflets, the
upper 8 pairs closed rapidly, the lower 5 passed slowly through an angle
of 40 degrees.

Leaflets. Continuous lines of crystals are present along the veins
of the leaflets — two to five rows according to the size of the vein. The
crystals are large, regular, hexagonal rhombohedra, and are so placed
in the cells as to have the longer axes of most of them parallel to the
direction of the vascular bundle with which they are associated. The
average size of the crystals is 15.8 microns long and 9.4 microns wide.
Each crystal cell contains one crystal, and is one-third larger than the
crystal itself. The protoplasmic sheath in which each crystal is imbed-
ded stains deeply with protoplasmic stains, such as eosin or aniline blue.

212 STECKBECK— ON COMPARATIVE HISTOLOGY

Pulvini. There are relatively few crystals present in the pulvini.
These are all of the regular type found in the leaflets, and are located
in the endodermal region, immediately around the bundles.

Petiole. The crystals in the petiole and midrib are so abundant
as to form a crystal sheath that covers the vascular bundles, the main
crystal development being found on the ventral side of the petiole. The
crystals are of the regular type noted in the leaflets. Branches., of
crystal cells extend into the bases of the petiolar glands and can be
traced to a distance representing i^ the height of each gland. The
crystal cells do not have the same regular continuity that they have
in the crystal lines of the petiole, but are more broken.

Cassia nictitans

This species, the Wild Sensitive-Plant of the eastern United States,
shows close similarities, in its structural and histological details and in
its irritability, to Cassia chamaecrista. Each pinnately compound leaf
bears 8 to 20 pairs of leaflets. Britton (6, p. 529) states the number
to be 12-44. The larger number probably applies to plants growing
further south. On the upper surface of the petiole, near the basal
pair of leaflets, a slightly-stalked gland is borne. The gland in this
species, as in the last, is an irritable center.

The movements of the leaves under nyctitropic and paraheliotropic
stimulation are like those described for Cassia chamaecrista, except that
this species is slightly more sensitive, not only to light stimulation, but
also to mechanical and chemical stimuh.

The illustration (Plate LVIII, Fig. 1) shows the expanded leaves of
young plants grown in the green houses. In night sleep (Plate LVIII,
Fig. 2) the leaflets are folded upward, back to back, and the leaves have
dropped from the day position.

The upward movement of the leaflets in paraheliotropic stimulation
begins at a slightly lower temperature than for C. chamaecrista. Here
the movement begins at 24° to 26° C. in bright sunlight. As the tem-
perature rises, the movement continues until a maximum closure is
reached at about 32° C.

Under optimum environal conditions two mechanical stimuh cause
a maximum closure of the leaflets and a drop of the entire leaf. After
one stimulus has been applied there follows a latent period of 3^ seconds,
and then a closure of the leaflets through an angle of 40 degrees. During
the upward and slightly forward movement of the leaflets, the midrib
of the leaf stimulated drops 35° to 40°. If a second stimulus is then

AND IRRITABILITY OF SENSITIVE PLANTS 213

applied, the leaflets close further through an angle of 30° and the mid-
rib drops 10 to 15 additional degrees. A third stimulus brings very
slight additional movement. The leaf so closed re-expands in 9 to 12
minutes.

Macfarlane (30, p. 201) proved that when carbonate of ammonia
is applied to the petiolar gland, closure of the leaflets on that same leaf
follows.

Leaflets. The regular lines of cr^'stal cells along the veins stand out
very clearly in this species. The crystals are uniform in size and are
of the six-angled rhombohedral type. The line across the middle of
the crystals is evident. The polarity of arrangement of the crystals
with reference to the ^ascular bundles is very marked. The crystals
on an average are larger than for Cassia chamaecrista, some attaining
a size of 18 microns long and 13.5 microns wide (Plate LX, Fig. 9).
The crystals are continued into the apex of the leaflets.

Pulvini. The regular six-sided prismatic crystals of the leaflets
are also present in the pulvini, and are located in the endodermal region.
There are relatively few of these crystals.

Petiole. Throughout the endodermis of the bundles an abundance
of crystals is present, forming a continuous crystal sheath, with branches
passing into the gland that is present on the dorsal side of the petiole
(Plate LXV, Fig. 27). The crystals are very abundant in the lower
half of the gland, being massed around the vascular bundle which enters
the gland. No crystals are found in the terminal region of the gland.
This upper portion of the gland resembles pulvinar structure. In the
stem, the crystal distribution is like that in the petioles. The crystals
are always present in the endodermis, and are found in all parts of the
stem. Branches of crystal cells extend into the base of the cotyledons,
but no crystals were observed in the cotyledons themselves. The crystal
lines cease at the point in the main axis where the tap root begins. The
writer was unable to find crystals in the roots of this plant.

The crystals of Cassia nictitans show the enveloping sacs better
than any plant described thus far. The intercellular connecting threads
are also very well brought out. By staining crystal cells in eosin for
48 hours, that had previously been fixed in chrom-acetic, and then de-
staining rather sharply, the writer was able to demonstrate very clearly
the presence of intercellular connections. Good results were also
obtained by following Gardiner's method to bring out intercellular
connections. By this method the sections are put in iodine solution
for 20 minutes, then in strong sulphuric acid for a few seconds; then

214 STECKBECK— ON COMPARATIVE HISTOLOGY

stained deeply in aniline blue. After this process the crystals stand out
clearly as white prisms in sharp contrast to the deep blue of the proto-
plasmic sacs. Between the adjoining crystal cells delicate stria tions
are observed.

Mimosa albida

The large genus Mimosa was referred to before as including species
that represent the climax of sensitivity in the Leguminosae. Mimosa
albida {M. sensitiva) is native to the tropics of the New World like the
great majority of the members of the genus. In the seedling stage,
the first leaves bear 3 pairs of leaflets. Professor Macfarlane (Lecture
Notes) demonstrated that the basal pair is less sensitive than the two
terminal pairs. After 4 shock stimuli had been applied, the two ter-
minal pairs closed, while the basal pair was closed to about half the
extent of the terminals. The leaves that are developed later are bi-
pinnately compound, with two leaflets to each leaf, and each leaflet
bearing two pairs of secondary leaflets. The leaflets of the terminal
pair are of equal size, and are obovate in shape. Of the basal pair,
the outer is of the same shape, but slightly smaller than those of the
terminal pair; the inner is very small and rudimentary. In the older
leaves the small secondary leaflet disappears entirely. There is a well-
defined primary pulvinus at the base of the petiole; secondary pulvini
are present at the bases of the leaf midribs, and a tertiary pulvinus
at the base of each secondary leaflet.

In the nyctitropic movement the terminal pair of secondary leaflets
move upward back to back. The large basal secondary leaflet folds
over the outer terminal pair; the small basal secondary leaflet also
moves upward; the midribs converge and drop so as to make an angle
of about 75 or 80 degrees with the petiole which also drops 40 or 45
degrees. This fully closed condition is observed only in the younger
leaves. In the older leaves, the leaflets are closed ^ of their full extent.

The paraheliotropic movement of the leaves is in the same direction,
and the same changes in position are noted as in the nyctitropic response.
The movement begins in bright sunlight when the temperature reaches
29° C.

The leaves of M. albida are sensitive to mechanical, chemical and
thermal stimuli. When a terminal secondary leaflet is stimulated,
after a latent period of 1 to 13^ seconds, it rises, followed very soon by
its partner; in 28 seconds the basal pair moved upward, and in 3 seconds
longer the midribs converged; in 8 seconds after that the basal pair of

AND IRRITABILITY OF SENSITIVE PLANTS 215

the other leaflet moved, followed by a partial closure of the terminal
pair in 20 seconds longer. To cause complete closure of the leaflets
2 to 3 stimuli are necessary. The leaf drops 30 to 40 degrees. The
re-expansion time of the leaves is 12 to 16 minutes. The small basal
secondary leaflets show the same sensitivity as the larger ones, for the
same sequence follows when one of them is stimulated, as when its large
partner is stimulated. The propagation of stimuli through the stem
of this plant is rather limited. If a leaf be stimulated, as described
above, the stimulus may be carried to the next leaf above, or occasionally
to two leaves in a young shoot. Chemical stimuli in the form of acids,
alkalies and alcohol cause a closure of the leaflets. Closure is effected
also when heat is applied, or when ice is placed on the leaflets.

Leaflets. The crystals along the veins are mostly of the six-angled
prismatic type, but there are also quadrangular prisms present. The
Hnes are fairly regular, but occasionally gaps are noted in the continuity
of the crystal cells. Usually a crystal fills }4 to H oi the lumen of the
cell in which it is contained. A fairly diffuse protoplasmic sac envelops
each crystal.

Pulvini. No crystals were observed in any of the pulvini. This
is the first of the sensitives described so far in which no crystals, neither
prisms nor conglomerates, are present. The structural details of the
pulvini resemble very closely those for Mimosa pudica.

Petiole and Stem. Continuous crystal lines extend through petiole
and stem in the endodermis. The crystals are mostly the six-angled
rhombohedral type, or the quadrangular type, very much the same rela-
tion as in the leaflets (Plate LX, Fig. 10). Occasionally an elongated,
rather narrow, styloid-like crystal is observed in the lines of crystals.
This form is of the type already noted in Robinia and Pueraria, and
as noted for those plants, has projecting processes at the ends of the
crystals. It is noteworthy that all three forms of crystal, in some of
the individual crystals, show the partition across the middle. This
would indicate that they are in afl probability twin crystals. The
occasional presence of two crystals in the same cell, as noted before,
further points to the same conclusion.

Mimosa pudica

There is probably no sensitive plant that has been observed for so

long a time and has been studied more extensively than Mimosa pudica—

the Sensitive Plant. It is of good size, with quite large bi-pinnately

compound leaves that show irrito-contractihty to a high degree. On

216 STECKBECK— ON COMPARATIVE HISTOLOGY

account of the very striking sensitive movement of the leaves the plant
has long been a fa\orite object for the study of irritable responses.
The relative ease of growing this plant in greenhouses in different
climates has been another helpful factor in making it a world wide
favorite for laboratory purposes. The leaves are highly sensitive
to nyctitropic, paraheliotropic, mechanical, chemical and other forms
of stimulation. The direction of movement and the position after
contraction is the same for the different kinds of stimuli just given.
That is in all cases of stimulation the secondary leaflets move upward
and forward, the midribs converge and the whole leaf drops. Such
complete closure, to be sure, takes place only under optimum environ-
ment.

Cotyledons. The cotyledons of Mimosa pudica show sensitive
movement when they are quite young — 3 to 8 days old. When older
they become yellowish and lose all sensitivity. Darwin (11, p. 127)
considers the cotyledons of this plant to be very feebly sensitive. Mac-
farlane (31, p. 203) states that the cotyledons are markedly sensitive.
"The cotyledons of the Sensitive Plant are most active during the
period that their activity is of greatest benefit, viz. in the very young
state (seedlings 2 to 10 days old), since their great function is to protect
the first leaves and the growing bud. When the latter have pushed
out above the cotyledonary tips, the protective function has ceased,
and their irritable movements are greatly lost. " Two to three stimuli
are required to cause the cotyledons to rise through an angle of 80
degrees. The movement is rather slow, resembling the movement of
the leaves of less sensitive plants, like some of the Desmodiums.

Large conglomerate crystals are scattered through the mesophyll
of the cotyledons. No prismatic crystals are present. The pulvini
of the cotyledons are devoid of crystals.

Leaflets. The crystals are mostly of the hexagonal type, with
relatively few of the quadrangular forms. The crystal cells are con-
tinuous along the veins, forming close sheaths on the ventral side of
the vascular bundles. The crystals are large, usually filling % to ^
of the cell cavities. Very often the crystals touch the dividing cell
walls.

Pulvini. No crystals are present in the pulvini.

Petiole, Stem and Root. A continuous system of crystal cells, which
form sheaths round the bundles, is present in the petioles and in the
stem of Mimosa pudica, with breaks only in the continuity in the pul-
vini. The regular crystals are also present in the sensitive lanceolate

AND IRRITABILITY OF SENSITIVE PLANTS 217

stipules. In the stem the crystal cells become more diffuse near the
beginning of the primary root, and are entirely wanting in the root
endodermis.

The regularity in the shape of the individual crystals, the extensive
distribution of crystals throughout the plant, the distinct protoplasmic
sheath around each crystal, intercellular connections between the
crystal cells, contact of some of the crystals with the dividing cell walls,
all these show a decided advance in cr>'stal relation over the plant pre-
viously described.

Mimosa Spegazzini
Of the species of Mimosa that have been studied, M. Spegazzini
shows the most highly specialized irritabiUty. It is slightly more
sensitive than Mimosa pudica in that the different phases of a stimula-
tion act are slightly shorter than for the latter species. Twenty to
thirty pairs of secondary leaflets are borne by each of the two leaflets.
Primary, secondary and tertiary puh'ini are well developed. Rather
strong hairs are developed on the entire plant.

The nyctitropic movement of the leaves is very similar to that of
Mimosa pudica. The secondary leaflets move upward and forward,,
the midribs converge and the whole leaf drops about 90 degrees (Plate
LXI, Fig. 12 and 13).

When a mechanical stimulus, such as a forceps pinch, is applied to
one of the terminal secondary leaflets after a latent period of less than
\i second the leaflet stimulated rises and its partner almost at the
same time. The stimulus is then carried down the midrib, the pairs
of secondary leaflets closing in order; in 9 seconds all the secondary
leaflets have closed, the midribs converge followed in 3 seconds by a
drop of the entire leaf. The stimulus moves up the other leaflet with
the result that the secondary leaflets close in order. In 20 seconds after
the stimulus had been appUed all the secondary leaflets are closed.
The stimulus is propagated through the stem to other leaves. The
writer in one case observed that five leaves above the one stimulated
and three below dropped. The extent of propagation of stimuli through
the stem is a decided advance from the irrito-contractile view point
over that of Mimosa pudica.

In crystal distribution the relation in M. Spegazzini is almost exactly
parallel with that of M. pudica. The crystals are of the same type,
of the same relative size, are present in leaflets, stipules, petioles and
stem, but not in the pulvini. The chief difference between the two

218 STECKBECK— ON COMPARATIVE HISTOLOGY

species, in regard to the crystals, is that in M. Spegazzini 5 to 7 lines
of crystals pass into the bases of the strong, stiff hairs that are scattered
over both surfaces of the leaflets. Are these hairs irritable? The
writer was unable to prove conclusively that the hairs are irritable
centers. Haberlandt (19) states that the hairs of Biophytum are sensi-
tive.

The distinct intercellular connections of the crystal cells can be
clearly demonstrated in M. Spegazzini. By staining such cells very
deeply, followed by destaining, the threads connecting the adjacent
cells stand out quite clearly (Plate LX, Fig. 11).

OXALIDACEAE

As in the case of Leguminosae, the writer studied the histological
details of certain species of this family, including forms that are native
and others that have been introduced from the tropics, and are grown
in greenhouses.

Oxalis stricta

This common native species exhibits very interesting sensitive
movements of its leaflets. The nyctitropic movements were carefully
studied by Ulrich (45, p. 226). The same author also noted the para-
heliotropic response of the leaflets. Macfarlane (30, p. 189) describes
the behavior of the leaflets when mechanical stimuli are applied.
^' After a sharp but delicate mechanical stimulus applied with a pencil
or other instrument to a terminal leaflet, a latent period of 33^8 seconds
elapses, followed by a period of slow but gradually accelerating con-
traction during the next 4 seconds. From the 7 to the 20th second
the motion is rapid, but thereafter slows down gradually to the 30th
second and then becomes increasingly slow till the 45 th second when
the contraction ceases. After 15-18 minutes expansion begins, and
a very slow rise can be traced till the leaf regains its expanded state
in 45-50 minutes."

Leaflets. In the leaflets small crystals are distributed through
the mesophyll tissue. These vary in shape from some that are somewhat
irregular, approaching a conglomerate form, to others that are prismatic
— four to eight-angled usually.

Pulvini. In the secondary pulvini no crystals are present. In
the primary pulvinus few large conglomerates are found scattered in
the cortical zone.

Petiole. Large crystals, arranged in irregular lines, are present in
the cortex just around the vascular bundles. Some of the crystals

AND IRRITABILITY OF SENSITIVE PLANTS 219-

are 40 microns in diameter, and lie in very large cells. In shape the
crystals might be described as being intermediate between conglomer-
ates and quadrangular prisms. The same type is represented for
0. Ortgiesi (Plate LXIV, Fig. 21).

Oxalis violacea

This specimen was compared with Oxalis stricta in its nyctitropic
movements and the crystal relation. The movement of the leaflets
in night sleep is exactly like that of 0. stricta.

The crystals scattered throughout the leaflet tissue are more nu-
merous than in O. stricta and are more varied in shape. Some of the
crystals are long and narrow, 38 microns long and 8.5 microns wide.

Oxalis corniculata

In this, another very common native species, the crystals in the
leaflets are more numerous and more regular in shape than in either
of the two species described. Most of them are prisms, or tetrahedra
fewer are of the conglomerate type. The crystal relation approaches
very closely that of the next species.

Oxalis floribunda

This introduced species is native to the mountainous regions of
Chili. The three obcordate leaflets of each leaf show nyctitropic
movement (Plate LXII, Fig. 14-15). The leaflets are very feebly sensi-
tive to mechanical stimuli.

The crystals are numerous, vary in shape but are mostly prismatic,
and are scattered through the mesophyll tissue (Plate LXIV, Fig. 23).
A few of. the crystals are elongated with a partition line across the
middle, and in general structure resemble the styloid-like crystals
described for the Leguminosae.

Oxalis Ortgiesi

This is another South American species, being indigenous to Peru.
The leaflets exhibit night sleep.

The leaflets also respond to mechanical stimuli, but show slow,,
rather sluggish movement.

The type of crystal present in the leaflets is similar to that of O.
floribunda. In the petioles and stem large crystals are found, which
in shape approach a prismatic form. These are scattered through the-
cortex but more abundantly near the bundles (Plate LXIV, Fig 21)_

220 STECKBECK— ON COMPARATIVE HISTOLOGY

Oxalis bupleurifolia

Of the different species of Oxalis examined, this is the most sensitive
in its leaflets. Each leaf consists of a rather long, flattened petiole
a quarter of an inch wide, that bears three small delicate obovate leaflets.
In nyctitropic movements, the petiole, as well as the leaflets, occupy
different positions from those of the day positions. Darwin (11, p. 328)
notes the movement of the petiole. "The foliaceous petiole rose during
the day and early part of the night, and fell during the remainder
of the night and early morning. " He also observed the movement
of the leaflets. Ulrich (45, p. 220) gives results obtained on stimulating
the leaflets with mechanical and with electrical stimuli. "When
stimulated mechanically the leaflet dropped in 0.75 seconds at a temper-
ature of 30° C. Eleven subsequent stimulations were given of equal
intensity a minute apart; each caused the leaf to fall slightly lower
than the preceding one. The leaflet had entirely recovered in thirty
minutes from the begining of stimulation."

Leaflets. Numerous huge, somewhat irregular crystals are scattered
through the mesophyll of the leaflets, with more of these toward the
lower surfaces. The crystals resemble very closely those found in
0. Ortgiesi, but exhibit greater regularity and uniformity in their shapes.
(Plate LXIV, Fig-22). Each crystal is 32-36 microns in diameter.

Pulvini. The type of crystal found in the leaflet is continued into
the small secondary pulvini, but here relatively few crystals are found.
In the primary pulvinus a greater number of the large crystals are
present scattered through the cortex.

Petiole. The foliaceous petiole of this plant shows a very interest-
ing crystal development, with one to three continuous lines of regular
crystals along the ventral side of each of the numerous bundle strands
^extending through the petiole. In shape the crystals resemble the
six-angled rhombohedral type so characteristic of the more sensitive
members of the Leguminosae, but the angles are not as sharp as those
of legumes, and the individual crystal is somewhat barrel-shaped.
The line across the widest part of the crystal is very evident in most
of the crystals. Each crystal occupies 3^-% of the crystal containing
cell, and is en\ eloped in a rather diffuse protoplasmic sac.

Of the different species of Oxalis that were examined the relation
in the petiole of O. bupleurifolia represents the highest development of
crystals, that is not including the species of Biophytum which by some
writers are included in the genus Oxalis.

AND IRRITABILITY OF SENSITIVE PLANTS 221

Biophytum dendroides
The genus Biophytum includes some 20 species scattered through
the tropics of both hemispheres. B. dcndroides, indigenous to Brazil
and Peru, is readily grown in greenhouses where it flowers and fruits
very freely. Fifteen to twenty-five pairs of oblong leaflets are borne on
each of the pinnately compound leaves. A short, rather thick secondary
pulvinus is present at the base of each leaflet, and a well defined primary
pulvinus at the base of the petiole. A similar pulvinus is found at
the base of each of the long flower stalks.

The night sleep of this plant is very striking. The leaflets drop
downward through an angle of 85-90 degrees, and the whole leaf falls
through 15-20 degrees. The upright flower stalks drop 70-80 degrees
(Plate LXIII, Fig. 16-17). The paraheliotropic response is equally pro-
nounced when the temperature, in bright light, reaches 29-30° C.

The leaflets are likewise quite highly sensitive to mechanical,
chemical, thermal and electrical stimuli. Macfarlane (30 p. 194-195)
describes fully the effects of the first three of these forms of stimulation.
When a terminal leaflet is stimulated mechanically it and its neighbor fall
through an angle of 40-45 degrees, after a latent period of % second to
2 seconds, varying with the relative age of the leaflets. The stimulus
is propagated through the midrib, the pairs closing in regular succession
toward the base, with a time interval of 2}/^ seconds between each pair.
If a second stimulus is applied, the leaflets will fall through 20-25 de-
grees, and on the application of a third stimulus they fall through a small
angle. A fourth and even a fifth stimulus may cause a slight additional
drop. Summation action is here necessary to cause closure of the leaf-
lets."

Macfarlane (30 p. 196) proved very conclusively that in Oxalis
dendroides, as in other sensitives, a stimulus is propagated more rapidly
in a centrifugal, than in a centripetal direction. "A particle of ice
placed on the pulvini of the middle pair, i. e., the tenth if there 19
pairs, will excite all the pairs above it within 15-17 seconds, but 21-23
seconds will elapse before the lowest pair in such a leaf as the tenth
from the apex-bud closes."

Leaflets. The crystals here are confined only to the endodermal re-
gion of the vascular bundles, being arranged in fairly continuous lines.
In shape, the crystals resemble those of the Mimosas, but are less uni-
form in size and more irregular in shape than in the Mimosas. Many
of the crystals are six-angled and indicate a line across the middle.
(Plate LXIV, Fig. 20).

222 STECKBECK— ON COMPARATIVE HISTOLOGY

In general, the long axes of the crystals are parallel to the
direction of the vascular bundles, but occasionally the crystals lie
cross-wise in the crystal cells.

Pulvini. In both primary pulvinus, and secondary pulvini, a few-
large scattered crystals occur in the cortex. These are quite irregular
and do not resemble the type of crystal present in the leaf.

Petiole. The lines of crystal cells in the petiole are of the same
nature as those along the veins of the leaflets. Each crystal is im-
bedded in a protoplasmic sac.

Biophytum sensitivum

This species of Biophytum is another fairly common green house plant,
that is native to the Tropics of both the Old and the New World. In
structural details and in leaf sensitivity it resembles rather closely
B. dendroides, but is slightly more sensitive than the latter.

The morphological characters of the plant, and the day and night
positions of the leaflets, are indicated in the illustration (Plate LXIII,
Fig. 18-19).

Leaflets. As in the preceding species, continuous lines of crystal
cells are associated with the vascular bundles of the leaflets. The
majority of crystals are of the regular, six-angled, rhombohedral type.
Here and there a barrel-shaped crystal is noted in the lines of crystals.
In the fully developed leaflets, each crystal fills % of the lumen of
the cell in which it is contained (Plate LXIV, Fig. 24-25). While the
leaflets are small, immature and are non-chlorophylloid, or are very
pale green, the crystals are relatively small. It is noteworthy that at
this stage the leaflets are non-sensitive or very feebly sensitive. Cunning-
ham (10) showed that in Mimosa pudica the leaflets do not attain
their maximum sensitivity until they have developed the character-
istic deep green color.

Pulvini. The lines of crystal cells end at the beginning of the pulvini.

Petiole. The crystals are of the same type as in the eaflets, and form
crystal sheaths along the vascular bundles. Each crystal is enveloped
in a protoplasmic sac, with fine connecting intercellular threads between
the adjacent crystal cells.

Nature of the Crystals

The crystals present in the two families studied — Leguminosae

and Oxalidaceae — consist of calcium oxalate, as was proved by applying

various chemical tests. Very thin fresh sections were used in the tests.

It is important to use thin sections so as to have the crystals exposed

AND IRRITABILITY OF SENSITIVE PLANTS 223

as much as possible to the chemical reagents that are applied. To
such preparations chemicals were added and the results observed carefully
under the microscope.

When concentrated hydrochloric acid is added to sections con-
taining crystals, it is noted that the crystals dissolve gradually. Con-
glomerate crystals dissolve more readily than the prismatic forms.

In the solution process the conglomerates are broken up into the
small prisms of which each crystal is composed, and these particles
are then dissolved. Each of the rhombohedral crystals in the solution
action of the acid is divided into two segments, the division taking
place along the partition line across the middle of the crystal. This
seems to indicate that this line is different in resistance to the rest of
the crystal, and may even consist of a different substance. Complete
solution of the crystals usually takes place in 1-3 minutes. Crystals
separated from the mucilaginous sheaths in which they are enveloped
are dissolved in a shorter time, the sheaths retarding the solution action.

Concentrated nitric acid acts like concentrated hydrochloric but
dissolves the crystals in a slightly shorter time. The crystals are like-
wise soluble in concentrated sulphuric acid, concentrated caustic potash,
but insoluble in acetic acid.

Aggregation

Darwin (12, p. 38) in his extensive studies on Drosera rotundifolia
observed that a very interesting change takes place in the appearance
of the contents of tentacular cells when stimuH are applied to the irrito-
contractile tentacles. "If the tentacle of a young, yet fully matured,
leaf that has never been excited or become inflected be examined, the
cells forming the pedicels are seen to be filled with homogeneous, purple
fluid. The walls are lined by a layer of colourless circulating protoplasm.
If a tentacle is examined some hours after the gland has been excited
by repeated touches, or by an inorganic or organic particle placed on
it, or by the absorption of certain fluids, it presents a wholly changed
appearance. The cells, instead of being filled with homogeneous purple
fluid, now contain variously shaped masses of purple matter, suspended
in a colourless or almost colourless fluid." Darwin described the
phenomenon as an aggregation process, and the masses formed are
"aggregation bodies." The shapes of these masses are fully described
in Darwin's "Insectivorous Plants".

De Vries (14, p. 6) made additional researches on the aggregation
phenomena in Drosera rotundifolia, and distinguished three phases in
agregation— (1) Retarded and much more clearly differentiated circula-

224 STECKBECK— ON COMPARATIVE HISTOLOGY

tion of the peripheral protoplasm. (2) Division of the vacuole into
numerous small vacuoles which all remain surrounded by a part of the
wall of the original \acuole. (3) A very significant diminution of the
volumes of the vacuoles.

Bokorny (3, p. 427-473) who studied aggregation in a number of
plants, not including any of the sensitives, considered the aggregation
mass to be a proteid which separates from the protoplasm. As to the
biological significance of aggregation, this author ventures no explana-
tion.

Loew and Bokorny (26, p. 614) refer to aggregation as an "echte
Lebensreaction ".

Klemm (22, p. 395-420) studied the nature of the substances that
appear in the aggregation masses. Among the constituents are proteins,
fats and occasionally tannin.

The irrito-contractile centers, especially the pulvini, of most of the
sensitives taken up in this imestigation, were examined for aggregation.
Each puh inus in its structure, as fully described by Sachs (40, p. 793)
shows a central vascular bundle cyHnder, around which a wide cortex
is present, the width of the cortex varying in the different sensitives.
It is noteworthy that the structure of the pulvini is very similar in both
Leguminosae and Oxalidaceae.

If the pulvini are put into a strong fixing solution of iron or gold or
platinum chloride while the leaflets are expanded, the cortical cells of
the pulvini are granular in their contents, but indicate no definite masses.
It should be observed that it is difficult to fix pulvinar tissue without
getting aggregation in that the cutting of the leaflets to put into the fixing
solution acts as a stimulation and causes aggregation.

When sections, taken from pulvini of such plants as Mimosa pudica
or Biophytiim sensUiviim, are examined under the microscope they
indicate rather clear shining globular masses in most of the cells of the
cortex. These masses occur throughout the entire cortex, but relatively
more abundantly on the ventral side of the pulvinus, that is they are
more abundant on the side away from the light. Each contracted
aggregation mass occupies %-3^ of the cavity of the cell.

The aggregation masses can be studied to better advantage if certain
chemicals are applied. For this purpose the writer has found the
following to be helpful— gold chloride, platinic chloride, ferric chloride,
ammonium carbonate, caffeine, silver chloride and silver nitrate. Gold
chloride when applied in a 5% solution stains the masses a purplish
brown.

AND IRRITABILITY OF SENSITIVE PLANTS 225

Aggregation bodies in the puh'ini of the various sensitives become

relati\ely more abundant, are larger and more uniform in size as one

passes from the less sensitive to the highly sensitive plants. In such

plants as Oxalis corniculata, O. floribunda and Amphkarpaea monoica

aggregation is not as pronounced as in more specialized plants like the

Biophvtums and the Mimosas. Cassia chamaecrista represents a good

intermediate t\pe. In the Cassias aggregation is observed in the cells

of the terminal portion of the petiolar gland as well as in the cells of

the puhini. In feebly sensiti\e plants like Gleditschia or species of

Oxalis when sections of the puhini are treated with, e.g., gold chloride,

or ferric chloride, it is observed that the aggregation bodies form slowly,

being small at first, and are often irregular, such as Darwin noted in

Drosera. Very often a number of small masses may be present in the

same cell. These smaller masses may coalesce and form a large body.

In the highly sensitive plants the aggregation masses form almost

instantly when treated with the \arious chemicals mentioned above.

Nature of the Aggregation Bodies, various chemical tests were used as

to the to determine if possible the nature of the aggregation substance.

These tests indicate that the masses are proteinaceous in nature. The

xanthoproteic test gives a positive protein test. Osmic acid stains the

masses a grayish-black, changing to a deep black. This suggests that

a fat or lipoid body may be present in the aggregation masses, but

other characteristic tests for fats, such as staining with alkannin gave

negative results.

The aggregation masses are readily soluble in 95% alcohol, con-
centrated caustic potash, and glacial acetic acid, but are resistant to
concentrated nitric and concentrated sulphuric acids.

Summary of Result

1. In geographical distribution it is shown that sensitive plants are
very rare in cold temperate regions, become more abundant in warm
temperate and sub-tropical regions, and attain their climax of develop-
ment alike in number of species and high degree of sensitivity, in the
tropics.

2. The majority of the highly sensitive species are natives of sub-
tropical and tropical America, from the Southern States, through
Mexico, Central America, tropical and sub-tropical South America.

3. The most wide spread irritable response shown by these is the
nyctitropic, or so-called sleep movement, which has in all probability
been induced by rapid changes in temperature, transpiration and
radiation effect in transition from day to night.

226 STECKBECK— ON COMPARATIVE HISTOLOGY

4. The paraheliotropic response probably next succeeded, v/hile
response to light stimuli, and to mechanical stimuli next became pro-
nounced.

5. Under experimentation, equally marked responses to chemical,
thermal and electrical stimuli are observed.

6. In relation to previous studies and discussions on propagation
of stimuli, the writer advances the view that this phenomenon is centered
in the endodermis. The cells of this layer contain a greater or less
number of crystals of oxalate of lime, which in number and in progressive
perfection of development show transition from less sensitive to the
most sensitive types.

7. Amongst the less sensitive plants, the crystals are often con-
glomerates, irregularly scattered outside the endodermis, or in discon-
tinuous lines along the endodermis and surrounding the bundle.

8. With increasing sensitivity and advance toward tropical conditions
the crystals show, alike in the Leguminosae and the Oxalidaceae, similar
progressive advances in abundance of crystals, regularity of shape in
these, and restriction of them to the endodermis.

9. The cHmax in this process is seen in Mimosa pudica and M.
Spegazzini amongst the Leguminosae, and in Biophytum sensitivum
and B. dendroides in the Oxalidaceae. These represent also the most
sensitive plants of the two families.

10. In histological distribution the crystals and crystal cells of the
endodermis become in the most sensitive species continuous from the
leaf margins and even in M. Spegazzini from the bases of the hairs,
along all of the veins of the leaflets, the petiole and the stem. But in
the irri to-contractile pulvinus regions, that receive and respond to stimuli,
the crystals are absent. In less sensitive pulvini, however, conglomerate
and transition crystals toward the most perfect type are present.

11. Each crystal cell in the higher types contains a rhombohedral
crystal that is traversed across its middle by a cleavage plane, that is
first affected by solvents such as concentrated hydrochloric and nitric
acids.

12. Surrounding each crystal is a protoplasmic sac that stains
deeply with protoplasmic stains, while from the sac intercellular uniting
threads seem to pass through the common membranes between the
adjacent cells, so as to form continuous protoplasmic connections
throughout the endodermal tissue.

13. The writer would view the crystals and the continuous proto-
plasmic investments and connections between these as the special
conducting lines for the passage of stimuli.

AND IRRITABILITY OF SENSITIVE PLANTS 227

14. With transition from the less sensitive to the more sensitive
species the cells of the pulvini contain, in increasing amount and com-
plexity, aggregation bodies resembling those previously described by
Darwin and other investigators, as associated with irrito-contractile

centers.

IS.These masses show contraction and aggregation changes under

stimulation.

16. From varied chemical reactions (tests for various cell contents)
the writer would view the aggregation bodies as being proteinaceous in

nature.

17. All irrito-contractile changes, alike during contraction and
expansion, seem to be due to changes primarily in the protoplasmic sac
surrounding each aggregation body, next in the aggregation body itself,
finally in the amount of liquid these may absorb or give off.

18. In fixation, demonstration and identification of the aggregation
masses the most helpful chemical agents were found to be gold chloride,
platinic chloride, ferric chloride, ammonium carbonate, silver chloride
and silver nitrate, the first two giving a brownish purple to purple
color to the aggregation bodies.

19. The writer agrees with Haberlandt in considering that the
complex hairs, often found over the irrito-contractile pulvini, or even
over the general leaflet surface, seem to act as delicate receptors of
environal stimuli.

20. Among the varied types examined the two most specialized
hairs are those distributed over the leaflets of Desmodium canescens and
Mimosa Spegazzini, in the former of which special bundle diverticula
pass into the bases of the hairs; in the latter prolongations, from the
zone of crystal cells rise up into the bases of the hairs.

21. All present evidence seems to point to the conclusion that, alike
in seedling and adult axes, crystal cells and aggregation masses gradually
become fewer in the hypocotyl and are entirely absent in the root.

22. For all of the above sensitive plants now studied, as for Drosera
recorded by Darwin, and Dionaea recorded by Macfarlane, all forms of
energy— thermal, luminous, chemical, mechanical, electrical and molar
stimuU, can act as stimulants to produce irrito-contractile movements.

Fig.
Fig.

1.

2.

Fig.

3.

Fig.

4.

Fig.

5.

228 STECKBECK— ON COMPARATIVE HISTOLOGY

Explanation of Plates
Plate LVIII
Cassia nictitans showing the expanded day position of the leaves.
Cassia nictitans. Leaves in nyctitropic position. The leaves have dropped
and the leaflets are folded upward.

Plate LIX
Section of leaflet of Pucraria Thunbergiana. Styloid type of crystal along

veins and in mesophyll.
T. S. secondary pulvinus of leaflet of Pneraria Thtinhergiana crystals around

vascular bundle cylinder and in cortex.
Section of leaf of Bauhinia diphylla. Scattered conglomerate cr>'stals and

discontinuous lines of prismatic crystals along the veins.

Plate LX

Fig. 6. L. S. secondary pulvinus of Gymnocladus leaflet of Cymnodadus disica.
Conglomerate crystals scattered through the cortex and a broken line
of prismatic quadrangular crj'stals along the vascular bundle.

Fig. 7. L. S. petiole of leaf of Robinia Pscudacacia showing elongated styloid and
prismatic types of crystals.

Fig. 8. Surface section leaflet of Desmodium canescens. Lines of crystals along the
veins.

Fig. 9. Macerated leaflet of Cassia nictitans. Lines of crystals along the veins.

Fig. 10. L. S. petiole of leaf of Cassia nictitans. Crystals are mostly of the six-
angled rhombohedral type; others are quadrangular and a few styloid.

Fig. 11. Section of leaflet of Mimosa Spegazzini. Lines of crj-stal cells. Crystals
are regular six-angled and enveloped in distinct protoplasmic sheaths.

Plate LXI

Fig. 12. Mimosa Spegazzini. Expanded day position of the leaves.
Fig. 13. Mimosa Spegazzini. Nyctitropic position of the leaves.

Plate LXII

Fig. 14. Oxalis florihunda. Day position of the leaves.

Fig. 15. Oxalis florihunda. Nyctitropic position of the leaves.

Plate LXIII

Fig. 16. Biophylum dendroides. Day position of the leaves.

Fig. 17. Biophytum dendroides. Nyctitropic position of the leaves.

Fig. 18. Biophyium sensitivum. Day position of leaves.

Fig. 19. Biophytum sensitivum. Nyctitropic position of leaves.

Plate LXIV
Fig. 20. Section, leaflet of Biophytum defidroidcs. Many of the crystals are six-
angled and indicate a line across the middle.
Fig. 21. L. S. stem Oxalis Ortgicsi. Crystals scattered in cortex are intermediate

between conglomerates and quadrangular prisms.
Fig. 22. Section leaflet Oxalis bupleurifolia. Crystals in mesophyll and along veins-
Fig. 23. Section oi leaQet Oxalis floribimda. Small crystals scattered through meso.
phyll tissue.

AND IRRITABILITY OF SENSITIVE PLANTS

229

Fig. 24. Section leaflet Biophylum saisitivum. Lines of large hexagonal prismatic

cr>stals.
Fig. 25. Surface view younger leaflet Biophytum sensitivum. Crystals relatively

small.

Pl.vte LXV

Fig. 26. T. S. petiole and gland from leaf of Cassia chamaecrista. Aggregation
material in the gland.

Fig. 27. L. S. petiole and gland of Cassia nicliians. Lines of crystals in petiole and
in basal two-thirds of gland.

Fig. 28. T. S. secondary puhinus of leaflet Amphicarpaea monoica. Relatively few
aggregation bodies.

Fig. 29. T. S. secondary pulvinus, Mimosa Spegazzini. Aggregation bodies num-
erous.

1.

Bates

2.

Bert

3.

Bokorny

4.

Borzi

5.

Bose

6.

Britton

7.

Briicke

8.

Burnett and]Mayo

9.

Coester

10.

Cunningham

n.

Darwin

12.

Darwin

13.

De CandoUe

14.

De Vries

15.

De Mello

16.

Dutrochet

17.

Gardiner

18.

Haberlandt

19.

Haberlandt

20.

Hill

21.

Holzner

22.

, Klemm

23.

. Leeuwenhoek, von

24,

. Lindsay

BIBLIOGR.\PHY

A Naturalist on the .Amazon.

Mem. d. la. Soc. de. Science physiques et naturelles de

Bordeau-x. 1870.
Jahr. f. Wiss. Bot. 1889. vol. 20.

L'apparato di mote delle Sensitive. Revista di Science
Biologiche. 1899.

Researches on Irritability of Hants. Calcutta 1913.

Manual of the Flora of the Northern States and Canada.
1905.

Archiv fvr Physiologic. 1848, p. 443.

The Quart. Journ. of Science, Literature and Art. 1827,
p. 76. 1828, p. 434.

Ueber die anatomischen charaktere der Mimossen.
Munchen 1894

Annals of the Royal Botanic Garden of Calcutta, vol. 6.

The Power of Movement in Plants.

Insectivorous Plants.

Pflanzenphysiologie. Tubingen 1832-35.

Bot. Zeitung. 1886, p. 6.

Journ. Linn. Society. Vol. VI, p. 253.

Rech. anatom. et physiol. s. la structure intime d. ani-
maux et d. vegetaux. Paris 1824.

Annals of Botany. Vol. I, p. 336.

Physiologische Pflanzenanatomie. 1904, p. 551-556.

Flora, vol. 99, p. 280-283.

The Sleep of Plants and Causes of Motion in the Sen-
sitive Plant, Explained. 1757.

Flora. 1867, p. 497.

Flora. 1892, vol. 75, p. 395.

Epistolae physiologicae. 1719. Epistola 44, p. 417.

MSS. in the Library of the Royal Society. London 1790.

230

SPECKBECK— ON COMPARATIVE HISTOLOGY

25. Linsbauer

26. Loew and Bokorny

27. Mac Dougal

28. Mac Dougal

29. Mac Dougal

30. Macfarlane

31. Macfarlane

32. Malpighi

33. Meyen

34. Millardet

35. MUler

36. Nemec

37. Pfeffer

38. Poulsen

39. Renner

40. Sachs

41. Scheele

42. Schivelv

43. Solereder

44. Spruce

45. Ulrich

46. Wallace

47. Wilson and Greenman

Ber. d. deutsch Botan. Gesellschaft. Bd. XXXII.
1914, p. 609.

Botan. Centralblatt. 1889, p. 614.

Bot. Gazette. Vol. 20, p. 411.

Bot. Gazette. Vol. 22.

Botan. Centralblatt. Bd. 77, p. 297.
Lectures at Marine Biological Laboratory at Woods
Hole. 1893.

Botan. Centralblatt. Bd. 61, 1895, p. 183.

Opera Omnia. 1687, p. 52.

Anatomisch physiologische Untersuchungen iiber den
Inhalt der Pflanzenzelle. Berlin, 1828, p. 59.

Nouvelles recherches sur la periodicite de la tension.
1869, p. 31.

The Gardeners Dictionary, 1733.

Die reizleitung und die reizleitenden strukturen bei den
Pflanzen. Jena, 1901.

Die periodischen Bewegungen der Pflanzen. 1875.

Flora. 1877, p. 45.

Flora. 1908, p. 87.

Botanische Zeitung. 1857, vol. 15.

Chemische Annalen, 1785. Bd. 1, p. 19.

The History of Amphicarpaea nionoica. Contributions
from the Botanical Laboratory, Univ. of Pa. Vol.
1, No. 1.

Systematic Anatomy of the Dicotyledons. Vol. 2, 1908.

"Notes of a Botanist on the Amazon." Vols. I and II.

Leaf Movements in the Family Oxalidaceae. Con-
tributions from the Botanical Laboratory of the
University of Pennsylvania. Vol. 2, No. 3.

Natural Selection Tropical Nature. 1895.

Preliminary Observations on the Movements of the
Leaves of Melilotus alba and Other Plants. Con-
tributions from the Botanical Laboratory, Univ.
of Pa. Vol. 1, No. 1.

Bot. Contrib. Uiih. Penn.

Vol. IV, Plate LVin

Fig. 1

Fig. 2
STECKBECK ON SENSITIVE PLANTS

Bot. Contrib. Univ. Pcnn.

Vol. IV, Plate LIX

Fig. 3

Fig. 4

N.

•*{ '

♦ » ' . • ' f

. if

-.* .• '-/^ .■ ' * %

..' ^

■ :'-\\

I . • •

' - I > >

Fig. 5
STECKBECK ON SENSITIVE PLANTS

Bot. Contrih. Univ. Penii.

Vol. IV, Plate LX

Fig. 6

Fig. 7

Fig. 8

Fig. 9

Fig. 10 Fig. 11

STECKBECK ON SENSITIVE PLANTS

Bot. Contrib. Univ. Pcnn.

Vol. IV, Plate LXI

Fig. 12

Fig. 13
STECKBECK ON SENSITIVE PLANTS

Bot. Contrib. Univ. Pciiii.

Vol. rV, Plate LXII

Fig. 14

Fig. 15
STECKBECK ON SENSITIVE PLANTS

Bat. Contrib. Univ. Penn.

Vol. IV, Plate LXIII

FiG. 16

Fig. 17

Fig. is Fig. 19

SIECKBECK ON SENSITIVE PLANTS

Bol. Conlrib. Univ. Pcini.

Vol. IV, Plato LX IV

^rr2 >

^

Fu;. 20

\-

^

Fig. 21

" i-

^ . f

Fig. 22

Fig. 23

s''^*^".•

-^

^.^^^^y^Pt^^

•

ir^ ■ ■

-

■\

t

Fig. 24 Fig. 25

STECKBECK ON SENSITIVE PLANTS

Bot. Contrib. Uuh. Pcun.

Vol. I\\ Plate LXV

Xt-J»

1.

Fig. 26

Fig. 27

Fig. 28 Fig. 29

STECKBECK ON SENSITIVE PLANTS

OBSERVATIONS OX THE BEACH PLUM
A STUDY IN PLANT VARIATION

BY

John Young Penn>-packer, A.M., Ph.D.
With Plates LXVI-LXX

(Thesis presented to the Faculty of the Graduate School

of the University of Pennsylvania, May 1915,

in partial fulfilment of the requirement for

the Degree of Doctor of Philosophy).

CONTENTS

I. iNTRODrCTlON AND HiSTORY 232

1. Introduction 232

2. Previous literature 232

?,. Material and methods : 235

4. Distribution 235

(1) Along the coast 235

(2) Inland 237

5. Habitat and Environment -'-'o

II. Structure of Prunus Maritima 239

1. Stem 239

2. Leak 240

3. Flower 243

4. Fruit "" • 244

III. Classification of Varieties 250

1. Prunus maritima var. caeruleo magna 251

2. Prunus maritima var. caeruleo parva 253

3. Prunus maritima var. praecox - • 254

^. Prunus maritima var. purpurea magna 255

5. Prunus maritima var. purpurea parva 257

6. Prunus maritima var. rubra magna ■ 258

7. Prunus maritima var. rubra parva 260

8. Prunus maritima var. lutea magna 261

9. Prunus maritima var. lutea parva 263

rV. Economic Value 264

V. Summary •"• • 266

VI. Bibliography 268

VII. Explanation of Plates ■ 269

232

PENNYP ACKER— ON THE BEACH PLUM

I. Introduction and History

1. Introduction
This subject was first suggested to me by Dr. J. M. Macfarlane, he
having pointed out to me the marked variations which he had observed
in the fruits of the plants which are produced at various places along the
Atlantic coast from Cape Cod, Mass., to Cape May Point, N. J., and
upon which he later pubhshed a paper entitled, "The Beach Plum,
Viewed from Botanical and Economic Aspects" (18, p. 216). In that
paper, he dwelt entirely on the variations as exhibited by the fruits, and
also suggested various evolutionary lines by which these varieties may
have evolved. He suggested to me, in view of the facts just set forth, the
possibility of finding variations in other aspects of the plants, which I
might correlate with the fruit variations. It was with this idea in mind
that I undertook the present work. After extensive research, I have
found certain varietal differences, which seem of sufficient importance
to warrant the division of Prunus maritima (Marsh) iato the following
varieties: —

1. Prunus maritima YQX. caerulea-magna.

2. " " var. caerulea parva.

3. " " var. praecox.

4. " " var. pu -purea magna.

5. " " var. purpurea parva.

6. " " var. rubra magna.

7. " " var. rubra parva.

8. " " var. lutea magna.

9. " " var. lutea parva.

2. Review of Literature

The history of the work which has been done upon this plant deals
almost entirely with the taxonomic or systematic phase of the species.
At first it was described by difi'erent botanists under numerous synonyms,
but the similarity of these being later discovered it is now known as
Prunus maritima.

It was first described by Marshall (19, p. 112) in 1785. He called
it Prunus maritima, "The Seaside Plumb," and the description, though
brief, was probably sufficient for that day. By way of general interest
and comparison, it is now quoted: "This grows naturally toward the
sea coast, rising to the height of eight to ten feet, often leaning and
spreading into many branches. The leaves are oblong, rather smaller

A STUDY IN PLANT VARIATION 233

and not so pointed as those of the common plumb; smooth and of a
shining green on the upper surface, but something Hghter underneath,,
and sHghtly sawed on the edges. It is generally well filled with flowers^
a few of which are succeeded by small, roundish fruit. "

The authorit>' usually given as the first describer is Wangenheim
(31, p. 103) in 1781. A copy of this work from the Academy of Natural
Sciences of Philadelphia was carefully searched but no record of Prunus
maritima was found. In another work {il, p. 103) published in 1787,
Wangenheim describes this species. Through some error in the dates
of these publications, which has been passed on without investigation,
it seems that Marshall (19, p. 112) has never been given credit for what
he truly deserves, namely — the original describer of P. maritima. On
the title page of Wangenheim (31) I found that he served as a captain
in the Hessian Army during the Revolutionary War from 1777-1780.
According to Stone (27, p. 491) this specimen which Wangenheim des-
scribed came from Long Island, N. Y. In view of these facts, and since
there is no mention of the name of the collector given by Stone (27),
at the above named place, it seems probable that he saw this plant,
while in the army service, and upon his subsequent return to Germany
published what has been considered the original description.^

Loudon (17, p. 691) described it as a native of North America; intro-
duced into Europe in 1818, and, while it produces blossoms there in
great abundance, yet he is not aware of any producing ripened fruits.
In America, according to Pursh (22, p. 332), they are succeeded by fruit,
of the size of a pigeon's egg, dark purple and very good to eat.

Until 1825, Prunus marilima was described under various synonyms.
Beck (2, p. 95) points out the similarity of descriptions among certain
species, yet suggests that they may be distinct. Torrey (29, p. 194-5)
and Torrey and Gray (30, p. 408) confirm Beck's suggestion and give
a list of some of the synonyms used.

Prunus cerasi^era
Prunus acuminata
Prunus sphaerocarpa
Prunus pygmaea
Prunus pubescens
Prunus littoralis

All these are now grouped, however, into the one species — now under
consideration. '

1 Since the above was written, Wight {Si, page 55) in a recent publication,
dated April 2, 1915, arrived at the same conclusion.

1789

Ehrhart

(12, p. 17)

1803

Michaux

(20, p. 284)

1803

Michaux

(20, p. 284)

1809

Willdenow

(34, p. 519)

1814

Pursh

(22, p. 332)

1824

Bigelow

( 3, p. 193)

234 PENNYP ACKER— ON THE BEACH PLUM

Until 1904 all the work was taxonomic and consisted of descriptions
more or less complete. At this time, Macfarlane (18, p 216), pointed
out marked variations in the fruits of the species under the following
heads: —

(1) color

(2) weight

(3) size and shape

(4) consistence

(5) taste

(6) time of maturation

(7) comparison of the stones

all of which the writer has confirmed by independent observation and
research eleven years later.

Bigelow (3, p. 193) had, however, previously (1824) called attention
to three varieties, namely: —
Variety a. Fruit an inch in diameter, purple, with a glaucous bloom.

b. Fruit similar, but smaller.

c. Fruit crimson shining.

He further states, "This is our Common Beach Plum, much prized
for its agreeable fruit, and deserving attempts at cultivation. I do
not find it described by any author, unless possibly by Michaux under
the name of P. sphaerocarpa, a name previously appropriated by Swartz
to a West Indian species. From P. maritima of Pursh it appears widely
different in its inflorescence, acumination, and fruit."

Torrey and Gray (30, p. 408) fourteen years later (1838) described
two varieties.

Var. 1. Leaves softly pubescent or pubescent beneath; fruit large,
pleasant. Corresponds to P. sphaerocarpa, Michaux; and P. liUoralis,
Bigelow.

Var. 2. Leaves when old mostly glabrous on both sides; fruit smaller,
red or purplish. Corresponds to P. pygmaea, Willdenow and P. acu-
minata, Michaux. They state that these two forms may be traced
into each other with great certainty; and Bigelow seems to have in-
cluded both under his P. liUoralis.

Stone (27) gives the best description of the distribution of the species
in the state of New Jersey. On page 250 the writer gives a detailed
technical description of P. maritima.

A STUDY IN PLANT VARIATION 235

3. Material and Methods

The material for this study was first collected by the writer in
early autumn of 1912, during the period of maturation of the fruit.
Those bushes were tagged which showed marked variations in the
fruits, and material was taken from them from time to time during
the summer of 1913, in order to trace variations in the stem, bud, blossom,
leaf, young fruit and mature fruit. As a check to my work, during
the late summer and autumn of 1913, 1 tagged other bushes and collected
material from them, as indicated above, during the summer of 1914.
In this way, I worked in duplicate the second summer, and checked it
with my work of the previous year. This should serve to strengthen
my conclusions, as my observations were not taken from any one bush
that I might arbitrarily select for my type, but rather from two to five
tagged bushes. (By so doing, there was reasonable safety that the tags
would not be torn off the bushes by the wind, or the rain.) Great diffi-
culty was experienced in obtaining fully matured fruits from the bushes,
since the natives gathered them just as they began to color and used
them for making preserves. Alcohol of 30% strength was used as a
kiUing fluid, and 70% was used as a preserving fluid. The material
from Long Island and from Nantucket was collected by Dr. Harshberger
during August 1913. As this material included only the leaf, stem and
immature fruit, no attempt was made to study it as was done in the
case of the New Jersey material.

4. Distribution

(1) Along the Coast— North and South
While this plant is a native of North America, and found only along
the Atlantic coast, yet authors do not seem to agree upon its wide dis-
tribution. In Gray's New Manual of Botany (16, p. 498) the range
is given as extending from Maine to Virginia. Britton (4, p. 524-5)
states that it has been found growing on the sandy coasts of New Bruns-
wick. Hutchinson (7) in an article in the Botanical Magazine (Decem-
ber 1909) confirms Britton's statement. As to its southern extension,
Chapman (6, p. 131) records it from Alabama; the basis of this record is,
however, a very imperfect specimen collected by Buckley in the Alleghany
Mountains, and Sargent in his "Silva of North America" is probably
justified in concluding that as no other trace of Prunus maritima has
been met with in this now well explored region, the specimen obtained by
Buckley merely represents a form of P. Alleghaniensis (Porter).

236 PENNYPACKER— ON THE BEACH PLUM

Sargent states (Silva of North America, 4(1892), p. 28): — "There is
preserved in the Herbarium of Columbia College a specimen of a Prunus
collected in Alabama many years ago by Mr. S. B. Buckley, and referred
by Torreyand Gray (30 v. 1.408) to their variety 2. of Primus maritima,
and in the same collection a specimen of what is described as 'a small
tree 10-15 feet high; fruit oval, small, blue, glaucous, very austere to
the taste,' and which was seen many years ago in Lincoln County, N.C.
by Mr. M. A. Curtis, who mentions it in his report of the trees of that
state (Rep. Geol. Sur. N. C. 1860, 3, 56). It is possible, as Professor
Britton is inclined to believe, that these specimens represent a southern
form of Prunus Alleghaniensis; but they are without flowers, and hardly
suffice to justify the extension of the range of the species of which no
other trace has been found in the now well explored region of the southern
Alleghany Mountains."

In a catalogue of the plants of Louisiana pubHshed in 1849, Prunus
maritima is given as growing there, but Professor R. S. Cocks, of Tulane
University, who has made a study of the vegetation of that state, and,
together with Professor C. S. Sargent, of the Arnold Arboretum, has
collected plums in all parts of it, writes that they have not come across
any specimen of Prunus maritima. While it may occur there, yet he
has no good evidence that it does.

My interest in writing to Professor Cocks came from a specimen
I saw preserved in the Herbarium of the Academy of Natural Sc"ences
of Philadelphia, and collected by Dr. Lincecum about 50 years ago,
along the sandy shores of Texas. He called it the Post Oak Plum. On
the same sheet was a brief description from which I quote: — "Small
shrub 1-2 feet high, two varieties, found in patches on the poorest sandy
post ridges. " I presume he means on the sand dunes which are covered
with scattered growths of the Post Oak. As that habitat is very much
like the one in which I collected my specimens, and the few characters
which he gives seem to correspond with Prunus maritima of New Jersey,
it leads me to believe that it was a closely related species of Prunus,
modified by the different climatic conditions under which it grew. The
specimen is in poor condition containing very few leaves or flowers.
He further states: — "It blooms before the leaves and so numerous are
they that the patches of the shrub are perfectly white and may be seen
a great way. Fruit small and sour. Purple in color. Blooms in
February. "

The specimens and material, which I have seen and examined and
from which the results of this paper were derived, range from Cape

A STUDY IX PLAXT VARIATION 237

Cod, Mass. to Lincoln County, N. C. The collections include the
following places: — Falmouth and Nantucket, Mass., Oyster Bay, Long
Island, and in New Jersey, at Island Heights Junction, Sea Side Park,
Barnegat, Manasquan, Manahawkin, Brigantine Beach, Pleasantville,
Ocean City, North Wildwood, Manumuskin, Maurice River and Cape
May Point. The latter place was the most favorable for collection
and study, as all the different varieties grow there, thus enabling me to
compare them all in a single day.

Just across the bay from Cape May in Delaware, Pnuius maritima
var. pygmaea is reported; specimens of w^hich can be seen at the Academy
of Natural Sciences, Philadelphia, and which were collected at Rehoboth
and Lewes, Delaware. It is characterized by smaller and more narrow

leaves.

Pruniis pygtnaea, a specimen of which can also be seen at the Acad-
emy, and which resembles Prumis maritima var. pygmaea from Delaware,
has been collected in Lincoln County, N. C. The specimen at the
Academy is very old and shows only a few leaves.

(2) Inland

About forty miles inland and running almost parallel with the present
shore line of New Jersey, beginning at a point near Salem, and extending
in a northeasterly direction to near Brown's Mills, occur local isolated
patches of Prunus maritima. Along this line, I have collected material
at Hainesport, Medford, Clementon and Atco. While I have not col-
lected any material, nor on my numerous excursions to the shore have
I seen it growing in the Pine Barren region, yet it is reported by others,
e.g. Stone (27), as growdng there. This distribution is quite unusual
as it is the only region, as far as I have been able to learn, where it is
found growing so far from the sea. The question naturally arises why
this distribution.

It has been suggested that birds or Indians may have carried them
there. But the distribution of plants is now generally regarded as
being connected with climatic or ecologic factors, it occurred to me there-
fore that a geologic explanation or suggestion at least was possible.
The soil at Clementon and at Atco, where these plants grow, is composed
of a light colored loose sand resembling in every respect our present
beach sand. The rounded surfaces of the grains indicate that it is water
worn. Occasionally the soil may be firm where strata of clay approach
and form the surface. Most geologists regard this formation as Miocene.
A few miles to the west is found the green sand marl or glauconite of
the Upper Cretaceous. The marked character in the color of the sand

238 PENNYP ACKER— ON THE BEACH PLUM

easily separates the two formations. At a period in the early Cenozoic,
when the waters of the Atlantic washed the land at a point near Clemen-
ton, Prunus maritima was evidently growing along the then sandy
shore line, but as the ocean slowly retreated eastward and time passed,
isolated patches of it seem to have become stranded. Further evidence
of this is illustrated by Stone (27) in his distribution of Prunus maritima
over the Pine Barren region to the east, which is geologically younger.
With the on-coming of the ice, and the approach of the glacial period,
we find that although this area was not glaciated yet the effects of the
period upon the plant Hfe could not have been more than to vary the
type, if it were permitted to live at all, and that is just exactly what I
have found, a variety with characters markedly different from any
found along the coast. Only one variety, the small blue, is found there,
which I have named — P. maritima var. praecox.

5. Habitat and Environment
Prunus maritima is found growing most abundantly on the drifting
sand dune area along the Atlantic coast. At Cape May Point, where
I have made extensive studies of the plant, I have found it growing with-
in twenty-five yards of the limit of high tide. The thicket formation
there is rather narrow, as back of the limited sand dune area is an exten-
ded marsh. To the south west, however, it extends for a distance of
several miles, covering an area of many acres. At other places along
the coast, the areas are much smaller, consisting in some places of only
a few hundred bushes, while at Hainesport and at Atco we find isolated
colonies of 50 to 100 bushes. The plants of this habitat consist oiQuercus
falcata, Juniperus virginiana, Myrica carolinensis, Ilex opaca, Pinus
rigida Rhus radicans, Sassafras officinale, Solidago sempervirens, Ammo-
phila arenaria, Lathyrus maritimus, Hudsonia tomentosa, etc. Regarding
the inland distribution of the species it is quite interesting to note that
the blossoms appear about two weeks earlier than those of the shore
varieties, and the fruits are fully matured before the ones at the shore
have begun to ripen. This is undoubtedly due to the environmental
conditions under which they are growing, as it is true of many other
species of shore plants, that they are often retarded in their development
and growth in the spring by cold breezes from the ocean. Again I have
noted that the blossoms on the bushes nearest the ocean are the last
to open. Likewise, the smallest bushes bloom the earliest, and the lowest
blossoms open the earliest, i.e., the blossoms at the tip of the branches
are the last to open. The explanation obviously is the radiation of
heat from the sun-heated sand.

A STUDY IN PLANT VARIATION 239

II. Structure

1. Stem

(a) External Structure: — The Beach Plum is a bush which usually
grows to a height of two to six feet and may then be from ten to twenty-
five years old. Occasionally plants from twelve to fifteen feet high
are found which indicate an age of eighty to one hundred years; the
main stem or trunk then measuring three to three and one-half inches
in diameter, at the level of the ground. The bark of the old stems is
rough, fissured, dark brown to gray, and usually covered with a per-
sistent growth of lichens. The younger branches are dark purpHsh
gray with numerous irregular, flattened lenticels. On the youngest
twigs the lenticels are more circular in outline and elevated. These
twigs are usually of a light silvery gray color, due to an epidermis which
is shed rather irregularly in the different varieties. After the shedding
of the epidermis, which is usually completed by the beginning of the
second year, the twigs are of a reddish brown color, the light gray len-
ticels being quite conspicuous. The epidermis, during the first year's
growth, is slightly puberulous. The leaf scars are small and crescent
shaped, and contain three small fibro-vascular bundle scars, which may
be seen only by aid of a hand lens.

Branching: — The method of branching is quite marked and varies
more with the situation in which the bush may be growing than with
the variety. As far as I have been able to ascertain, the branching
should not be considered a constant factor, although it may apparently
seem to be from some of the plates incorporated in this thesis. The bushes
exposed to the cold sea breezes, and the sand laden winds are the most
stunted in their growth, the yearly additions being rarely more than
two or three inches. As a result, the stems are more sturdy and more
branched, like that shown on Plate LXVI, Fig. 1. This is quite
typical of the large blue variety^ and was taken from a bush bearing
that color of fruit. The yellow variety Plate LXVI, Fig. 4 illustrates the
other extreme of the series of photographs taken. This branch was taken
from a bush growing five-hundred yards from the shore where the more
sheltered conditions favored an elongated less branched type. However,
when the yellow variety grows nearer the ocean it more closely resembles
one of the red variety of which Plate LXVI, Fig. 3, is very typical.
The purple variety, a specimen of which is shown on Plate LXVI, Fig.
2, shows spine like lateral branches, which although they are only

240 PENNYP ACKER— ON THE BEACH PLUM

■5hown as belonging to this variety, should not be considered as wholly
characteristic of it. They may occur on the reds and blues, even though
they are not shown on the photographs, but I have never found any on
the yellows. This purple type grew under intermediate conditions
as the branching would seem to indicate. The short lateral branches
— which usually fruit well — shown on Plate LXVI, Fig. 1 will later become
sub-spinescent and resemble the spine like branches of Plate LXVI,
Fig. 2.

(b) Internal Structure, (microscopic): — The internal structure of
the stem which is shown on Plate LXVIII in cross section conforms to
previously described woody stems of the Rosaceae, of which the following
parts may be summarized briefly: — Epidermis rather thin, dark reddish
brown, which by the end of the first year is beginning to be shed. By
the end of the third or fourth year a thin bark has formed which sloughs
off in thin persistent scales or plates. Beneath is a layer of cork from
fifteen to thirty cells thick, which gradually pushes outward to form
the bark as new cork cells are cut off. These protective coverings sur-
round a rather thin cortex composed of ten to fifteen rows of parenchyma
cells; included in which are numerous clustered crystals of calcium
oxalate and numerous irregular masses of sclerenchymatous tissue or
stone cells. The latter are distributed around the stem in the inner
portion of the cortex. Beneath is a ring of rather loosely arranged
phloem tissue which extends outward from the wood in curious finger like
projections. Numerous air spaces are thus formed between them and the
rows of pith cells as they extend outward to join the cells of the medullary
rays with those of the cortex. The wood, which is composed of narrow
annular growths, is hard and compact. It contains numerous water
conducting vessels which vary in relative number in the different varie-
ties. It is also characterized by many broad medullary rays, widened-
out extensions of which separate the phloem into the finger like projec-
tions mentioned before as they pass outward to join the cortex cells.
The pith, which comprises the central portion of the stem, is solid and
compact, while the cells of it store starch during the resting period.
Clustered crystals of calcium oxalate are frequently but irregularly dis-
tributed throughout it.

2. Leaf

(a) External Structure: — The leaves opening with or after the blos-
soms, rarely before, vary in shape from oval to ovate to obovate. They
are rather firm in texture, greenish above paler beneath, and more pubes-

A STUDY IN PLANT VARIATION

241

cent when young. This pubescence is due to hyaline, unicellular,
epidermal hairs. The stipules are Unear, lobed, pubescent, glandular-
serrate and very early deciduous. The leaves are more or less rounded at
the base, acute at the apex, finely and sharply serrate with glandular
tipped teeth when young. At the junction of the lamina and petiole
are two rather large conspicuous glands. In regard to leaf variation,
in lamina and petiole considerable diversity exists, and the main types
that are associated with the varying colors of fruits are set forth in the
subjoined table.

VARIETY

SIZE OF LEAF

LENGTH OF PETIOLE

Large Blue

3.5 X 6.8 cms.
3.0 X 5.7 "
2.5 X 5.5 "
3.5 X 6.8 "
2.0 X 4.0 "
3.2 X 6.8 "
2.0x4.5 "
3.5 X 6.8 "
2.5 X 5.7 "

1.0 cms.

Small Blue (1)

1.0 "

(2)

.7 "

Larere Purole

1.2 "

Small Purole

1.0 "

Large Red

1.1 "

Small Red

.9 "

Large Yellow

.8 "

Small Yellow

1.0 "

The above measurements are the average of ten selected leaves.
Small Blue (1) is var. coerulea parva from Cape May Point; (2) var.
praecox from Hainesport, N. J. The leaves are puberulous on the up-
per surface along the midrib in all varieties, while in the blues and in
the yellows the pubescence seems to exist along the main veins as
well. In the reds and in the purples the upper surface of the blade
appears glabrous except along the midrib. On the lower epidermis the
pubescence is located along the midrib and in the axils of the main
veins. The small inland blue var. praecox appears as pubescent on the
lower surface as on the upper, while the small yellow var. lutea parva is
pubescent along both veins and cross veins, but not as pubescent on the
lower as on the upper surface. These details are set forth in compara-
tive order in the table, on the following page.

Small Blue (1) var. coerulea parva; (2) var. praecox.
The pubescence of the small yellow is more dense on the upper epi-
dermis, and appears to be very characteristic of it. Stomata occur only
on the lower surface in island-like formations. Sometimes these islands
are irregular and indistinct, i.e., broken up into a number of smaller
ones, which makes it rather difficult to get an accurate account of the
number of stomata in an island area. The midrib, veins and nerves

242

PENNYPACKER— ON THE BEACH PLUM

VARIETY

Large Blue

Small Blue (1).

(2).

Large Purple...

Small Purple

Large Red

Small Red

Large Yellow...
Small Yellow...

UPPER EPIDERMIS

Pubescent along mid rib and lead-
ing veins
Pubescent only along midrib

Pubescent along midrib, veins and

cross veins

Pubescent only along midrib

Pubescent only along midrib

Pubescent only along midrib

Pubescent only along midrib

Pubescent along midrib slightly
along the leading veins
Pubescent over whole surface,
midrib, veins and cross veins

LOWER EPIDERMIS

Pubescent only along
midrib

Pubescent only along
midrib

Very pubescent as des-
cribed for up. epid.
Pubescent only along
midrib

Pubescent only along
midrib

Pubescent only along
midrib

Pubescent only along
midrib

Pubescent only along
midrib

Pubescent along mid-
rib, veins and cross
veins

are sharply prominent beneath, where 6 to 8 veins usually occur on each
side of the midrib.

(b) Internal Structure (microscopic). The leaf is characterized by
a thick epidermis that prevents too rapid transpiration of water. The
cells of the upper epidermis are 2 to 2 3^ times larger and more dense
than those of the lower. Frequently mucilage is deposited in the cells
of the upper, less often in the cells of the lower. Two rows of palisade
cells can usually be distinguished, but these are sometimes small and
hard to differentiate. The mesophyll is composed of an undifferentiated
mass of minute cells, which is very characteristic of all varieties, the
air spaces are numerous, but small, while along the midrib and main
veins are numerous clustered crystals of calcium oxalate. These vary
in size and relative number in the different varieties. In cross section,
the crystals are seen to be scattered throughout the mesophyll, some-
times in the palisade layers near the upper epidermis, at other times
in the undifferentiated mesophyll near the lower epidermis, but never
so far as I have been able to discover, are they located in the epidermis.
Frequently, these crystals vary in size in the same leaf and may be classi-
fied as large, medium, and small. The midrib contains a deeply crescen-
tic vascular bundle that occupies one-half of its area and is surrounded
by a loose cortex of mesophyll cells, which, like the leaf mesophyll, con-

A STUDY IN PLANT VARIATION 243

tains crystals. The crescentic shaped patch of xylem consists of a
mass of woody or lignified tissue. Its elements are arranged as a radiat-
ing series of cells. These are surrounded on the lower side by a cres-
centic patch of phloem in the outer zone of which are numerous mucilage
cells and crystals of calcium oxalate.

3. Flower

The flowers of P. maritima are white, with a pinkish tinge in the
blue and in the purple varieties. This is especially noticeable in the
sepals and stamens after the petals have fallen, although some of the
purple variety have decidedly pinkish petals. They are borne in umbels
with 2 to 5 flowers (usually 3) in a cluster near the ends of the youngest
twigs, or on the short lateral spinescent branches of the older wood. The
period of blooming lasts from ten days to two weeks. It ranges from
April 19th, for the small inland blue (observation made at Hainesport,
N. J.) to May 15th for the late varieties at Cape May Point. In most
of the varieties the blossoms appear before the leaves, yet in some of the
more protected bushes the leaves are well developed before the blossoms
fall. Evidence of this may be seen on some of the accompanying plates.
The writer has no accurate date as to the periods in other places north
and south, but their relative periods of blooming can easily be approxi-
mated from the time stated for the New Jersey varieties. The pedicels
are very pubescent; the hairs on the reds and purples being the longest,
while those on the blues and yellows are relatively short. The sepals
are green with pinkish margins as noted above, concave, pubescent,
uniting to form a campanulate receptacular tube 2 to 3 mm. deep. The
deepest cups are found on the varieties bearing the larger sized fruits.
The sepals vary in length from 2 to 4 mm. and from 1 to 2 mm. in width.
The outer surface of the receptacular tube is very pubescent, while the
inner concave surface is completely overspread by an orange yellow
glandular tissue which secretes the abundant nectar. The petals are
white with a slightly pinkish tinge in the blue and purple varieties,
rounded at the apex and contracted at the base into claws. They vary
in width from 4 to 7 mm. and in length from 6 to 10 mm. The corolla
spreads out to form a relatively flat surface of which the diameter is
15 to 20 mm. in the large and 10 to 15 mm. in the small varieties. The
stamens are variable in number, but usually about 30 of which some are
frequently abortive; 25 being the average number in the small inland
blue, 35 in the large blue black. These vary in length from 1-2 mm.
in the small abortive ones nearest the center, to 5-10 mm. in the outer-
most ones, and are inserted on the edge of the receptacular tube. The

244 PENNYP ACKER— ON THE BEACH PLUM

anthers are two lobed and bear numerous triangular-shaped pollen grains.
The flowers are apparently proterogynous, since the stigma matures
before the flower is fuUy open and may easily be seen protruding between
the petals of a partly opened flower bud, but homogamy is probably
not uncommon. The style projects in the middle of the flower beyond
the obUquely diverging stamens, and is 10 to 20 mm. long. The ovary
is one celled and characterized by two pendulous ovules. Frequently
the flowers are andro-dioecious which would account for the profusely
abundant blooms matured on every bush during April and May, and
also for the total lack of fruits on some bushes, the rarity on others, and
the extreme abundance on still others. The last noted would naturally
be those that developed stamens and a well formed pistil with swollen
ovary and elongated style. At Hainesport, N. J. the writer has observed
bushes for three years, during which time they have never set fruit, yet
have constantly bloomed in great profusion. The same condition pre-
vails in other locahties. The flowers are insect pollinated as is indicated
by the vast horde of Hymenoptera which frequent them during the warm
part of the day. The showy white corolla and the nectary are the
chief attractions for insects as the flowers are practically odorless. Cross
pollination is favored by most writers, and such a view seems Ukely
from the fact that the bee in order to obtain the nectar must get far
into the interior of the flower. In so doing it seems hardly possible
for it to miss touching the stigma.

Knuth states that automatic self-pollination appears regularly to
take place in hermaphroditic flowers of the Pruneae should insect visits
fail. Whether this is effective or not seems doubtful, as Kirchner states
that numerous bushes observed by him rarely set fruit. The true ex-
planation, however, seems to be found in the andro-dioecism as described
above by the writer.

4. Fruit
The fruits exhibit by far the most striking variation to the eye,
and it was this variation which caused Dr. Macfarlane (18) to first
make a preliminary study of the Beach Plum. As I have inferred under
materials and method, I began making observations and collecting
material when the bushes were in fruit as the variation at that time
was the most easily recognized. The fruit, which is usually produced
in large quantities, is borne on the younger branches mostly of the
previous year's growth. They begin to ripen about the first of August
in the earlier varieties and continue to produce an abundance of fruit

A STUDY IN PLANT VARIATION

245

until the latter part of September when the best varieties have fallen.
There is, however, a variety of the small purple, which does not mature
all its fruits until frost overtakes it early in October.

It is interesting to note that the fruits remain upon the bushes about
three weeks after they have ripened, a much longer time than we find
among our cultivated varieties. Likewise, it may be of interest to know
that one rarely finds any of the fruits rotting on the bushes; and after
they have fallen, the fruits, if not carried away by birds, or other animals
that distribute fleshy fruits, are palatable for another week, after which
the pulp cells begin to disintegrate, but the leathery skin prevents
them from rotting until they finally dry into a shrunken shriveled mass.
The fruit is sub-globose or slightly oval. The skin is tough and leathery,
of a blue, purple, red or yellow color, covered with a light waxen bloom,
and flecked with numerous small light-colored spots. The flesh is of
a watery yellow color, juicy, astringent and in most varieties decidedly
free from the stone. The stone is rounded or flattened, pointed at both
ends, rigid on the ventral, and sHghtly grooved on the dorsal suture.
The taste depends largely upon the presence or absence of the tannin
content. Sugar is evidently distributed throughout the pulp, while
tannin is deposited in a layer just beneath the skin. Mucilage is
very abundant, and is contained in ducts or canals imbedded in the
pulp. The yellow fruits seem to be rather free from the tannin and
consequently do not have a sharp, astringent taste. That they vary
in size is shown from the following table.

VARIETY

Large Blue

Small Blue (1)...
(2)...

Large Purple

Small Purple (1)
(2)

Large Red

Small Red

Large Yellow

Small Yellow

DIMENSIONS

1.7 X 1.6 cms.
L5 X 1.4
L2 X 1.3
1.7 X 1.6

1.3 X 1.2
.9 X 1.0

1.6 X 1.5

1.4 X 1.3

1.7 X 1.6
1.4 X 1.3

LENGTH OF STEM

1.4

cms.

1.9

1.1

1.4

1.7

1.4

1.5

1.2

1.2

1.5

The length of the stem, as shown above, is quite variable, and while
it is not constant and does not seem to explain much, yet it serves to
show the great variation which this plant exhibits. Small Purple
(2) noted in the table above illustrates the variety which does not ripen
until October.

246 PENNYP ACKER— ON THE BEACH PLUM

Color. The red, purple, and blue colors seem to be due to one or
more pigments added to the yellow color. This became apparent when
I added a dilute solution of NaOH to the skin of the red variety. The
red dissolved pigment was readily attacked by the alkali and quickly
changed its color to purple, to green and finally to yellow. Small yellow
chromatophores were easily distinguished after the disappearance of
the dissolved pigment, and the pulp, which was treated in a simi-
lar manner, though containing only a small amount of pigment,
had the same general appearance as the yellow variety. This seems
to prove, as far as color goes, that the yellow is the more primitive t>^e
and that the red has taken on an added pigment. In the purple and in
the blue varieties there appears a sub-epidermal layer which in addition
to the red dissolved pigment (same as in the red varieties) gives to the
fruits their respective colors. This pigment is contained in small oval
or rod shaped bodies which are considerably smaller than the cells which
contain the red dissolved pigment. The skin of the purple and of the
blue varieties resembles the skin of the red after the inner pigmented
layer has been removed. This further emphasizes the fact that the
yellow is the more primitive type, since the purples and the blues have
evolved from the reds by the increasing alkalinity of a dissolved red
pigment, and again the reds have evolved from the yellows by the for-
mation of a dissolved acid pigment, and the disappearance of the yellow
color from the chromatophore. To say that these varieties overlap,
and that the colors grade into each other i.e., that the purples grade
into the reds on one side, and into the blues on the other, may appear
to weaken the value of the varietal characters given later. But when one
considers the sum total of varietal characters typical of each fruiting type,
the distinctions above indicated for the fruit seem to hold fairly well
for other parts, and to constitute sufficiently stable diagnostic differences.

The majority of the fruits in any one locality are of the dark blue
variety, and comprise about 75% of all the plants. The yellow is the
least common and forms about 1%, the reds about 5% and the purples
not over 20%. The only locality where I have seen the yellow variety
growing was at Cape May Point, although it has been reported as grow-
ing at Island Heights Junction, N. J., and at Falmouth, Mass. The other
varieties are quite common and can be found in almost any locality
where the Beach Plum grows.

The earliest ones to ripen are the small blue var. praecox, which are
fully matured by August first. These, as I have noted earlier in this
paper, blossom about ten days earlier than any of the other varieties.

A STUDY IN PLANT VARIATION

247

The next to ripen, and the first at Cape May Point, are the large blues
which are fully matured by August fifth, and continue until about
August twenty-fifth. They ripen gradually on the bush like our cul-
tivated plum, and there may be found on any of the bushes fruits in all
degrees of maturity. The yellows, however, differ from the others in
this respect, as the fruits appear to all mature about the same time.
As the blues ripen, they first turn pinkish on one side and this gradually
deepens into a deep red. At this stage, they closely resemble the large
red variety, and could easily be mistaken for them if they matured at
the same time, but in about four days they will have turned to a dark
blue and are now covered with a light waxen bloom. About the time
that the blues are fully matured, the yellows begin to produce abundant
fruits, which are characterized by their sweet, pleasant taste. The
large purple variety ripens next, followed shortly by the red. Finally,
about the last of September, the small purple ripens, the coarsest and
smallest fruited variety. They are by far the richest in tannin and
decidedly unpleasant to the taste. Usually frost overtakes them before
they attain what may be termed ripeness.

VARIETY

PERIOD OF MATURATION

Large Blue

Small Blue (1)

(2)

Large Purple..

Small Purple...

Large Red

Small Red

Large Yellow..
Small Yellow..

Aug.
Aug.
Aug.
Aug.
Sept.
Aug.
Aug.
Aug.
Aug.

5— Aug. 25

10— Aug. 30

1— Aug. 20

15— Sept. 5

1— Oct. 1

25— Sept. 15

25— Sept. 15

10— Aug. 30

10— Aug. 30

The fruit is subject to the sting of an insect, the plum weevil or plum
curculio as it is sometimes called. The stinging occurs either when the
blossom is on or shortly after it has fallen. The very young fruits can
then readily be detected when punctured by the insect, owing to a mass
of mucilage exuding from the wound. The blues and the purples are
the most generally attacked, and often more than half of the fruits are
stung. The reds are occasionally attacked while the yellows are seldom,
if ever, stung. Contrary to expectation, only a few of the stung fruits
fall to the ground, while the majority of them remain upon the bush
even to maturity. There results from the sting a hard black core of
hardened mucilage, which becomes a localized center for the secretion
of tannin, and this gives to the fruit a most unpleasant taste.

248

PENNYPACKER— ON THE BEACH PLUM

During the early stages of growth, the young fruit is subject to a
fungus attack by Exoascus Pruni (Plum pocket). This is the same
fungus that attacks our cultivated plum and which has been fully de-
scribed by De Bary and Duggar.

Fruit Weight: — It is possible under this head to give statistics which
will emphasize genetic peculiarities and with which I was able to correlate
varietal characters in the leaf and flowers.

VARIETY

Large Blue

Small Blue (1)....
(2)....

Large Purple

Small Purple (1)
(2)

Large Red

Small Red

Large YeUow

Small Yellow

WEIGHT or

FRUIT

3.61
1.74
1.29
3.52
1.40
.55
2.82
1.71
2.92
1.70

grams

WEIGHT OF
STONE

.54 grams

.22

.18

.43

.22

.18

.40 "

.28

.40 "

.22

WEIGHT OF
FLESH

3.17 grams
1.52
1.11
3.09
1.18
.37
2.42
1.43
2.52
1.48

PEE CENT OF
FLESH

88
87

86

86.5

84

67

86

84

86

87

These variations in weight are very constant for each variety. The
weights were obtained by weighing 10 plums and then estimating the
average. The per cent of flesh ranges from 84 to 88 per cent with the
exception of the small purple late variety which does not ripen until
October. The table shows that the stone has attained nearly its fuU
size and weight while the flesh development is quite scanty, owing to
the lateness of its maturity.

Stone Variation: — By way of comparison the stone variation is given
and seems to markedly correspond with the variation as shown above
in the size and weight of the fruits.

VARIETY

LENGTH

THICKNESS

Large Blue

12.5 m.m.
11.0 "

7.8 "
13.5 "

8.5 "
12.5 "

9.0 "
10.0 "

8.5 "

11.0 X 7.3 m.m.

Small Blue (1)

7.5 X 5.8
6.5 X 5.4
10.0 X 6.5
7.8 X 5.5
9.5 X 6.5
8.5 X 5.8
9.5 X 7.5
7.0 X 6.0

(2)

Large Purple

Small Purple

Large Red

Small Red

Large Yellow

Small Yellow

A STUDY IN PLANT VARIATION 249

Cell Contents and Glandular Stntctures

1 . Calcium oxalate is excreted into cells of this plant as a waste pro-
duct in considerable abundance. The evidence of it is shown by the
rosette aggregate of crystals found in different parts of the plant. They
are of various sizes and usually occupy an entire cell. The greatest
abundance of them is found in the cortex of the stem, and in the loose
parenchyma cells of the leaf mesophyll, where they are irregularly scat-
tered throughout. Likewise, they are found in the sepals and receptac-
ular tube. Again in the fruit, they are found not only in the soft pulpy
cells and in the cells which are thickening their walls to form the stone,
but also in the developing embryo inside. That they may be found in
other parts of the plants is, in the opinion of the writer, very probable,
as no special attempt was made to study them, their presence simply
being noted in the histological study of the above named parts.

2. Mucilage is likewise excreted in great abundance, evidence of
which is found in the cells of the upper and of the lower epidermis of the
leaf. Mucilage canals are also found in the fleshy part of the fruit,
and when the fruits are stung by the plum weevil masses of the mucilage
exude from the puncture which soon hardens into a plug heaUng up
the wound. When the ripe fruits are plucked from the stalks Macfar-
lane (18, p. 226) refers to the considerable amount of "bleeding" which
occurs when they are heaped in a basket or on a dish for a few hours.
This secretion, which is sticky to the touch, is evidently the same mucilage
as that previously referred to. If the stalk is plucked with the fruits
this bleeding is entirely prevented. Frequently when a branch is broken
from the bush or an injury to the bark occurs, mucilage is exuded to
cover the exposed wound.

3. Tannin which gives to the fruits their bitter astringent taste
is deposited in a layer beneath the skin. The yellow varieties are almost
free from it, while the late maturing ones contain the most; the small
purple being extremely bitter. The intermediate types vary in tannin
content. The softer and larger varieties when ripe contain only small
traces of tannin.

4. Glands of the most interesting character are found on various
parts of the plant, viz., at the junction of the lamina with the petiole,
on the lobes of the deciduous stipules and on the teeth of the lamina.
These are all fairly large — visible to the unaided eye — of a dark brown
color, and very dense. The glands at the junction of the lamina with
the petiole are borne on short stalks. Normally, there are two — one
on each side of the petiole — but frequently one of these may be absent,
while very rarely leaves are found with no glands developed. The

250 PENNYP ACKER— ON THE BEACH PLUM

glands on the stipules are very conspicuous, but probably rank second
in importance as the stipules are very early deciduous and consequently
the glands are seen only for a short time. A reproduction of them is
shown in Plate LXVIII, Fig. 11. The glands on the teeth of the lamina
are more numerous near the apex of the leaf, and seem to be more com-
mon on the young leaves. Many of the older leaves do not show them.

III. Classification of Varieties

In deahng with this phase of the subject, I will give a detailed taxo-
nomic description of the species as a whole after the style of Sargent.
Such a description has not been attempted so far as the present hterature
goes and for that reason is needed all the more. Then I will give a de-
tailed description of each variety as I have found them. To have estab-
lished new species on the basis of the variation presented seemed to me
rather to confuse, and for that reason I have decided to describe them
as varieties. I might add that some of my varietal characters are equally
as strong as the characters which were used in separating P. Gravesii,
which grows in the immediate neighborhood and under precisely the
same conditions as P. maritima.

Prunus maritima, Marsh.
(Beach Plum)
A small slender shrub, occasionally 12 to 15 feet high, with a main
stem which is sometimes 3 1/2 inches in diameter, and which divides
into numerous erect rigid branches; usually 2 to 5 feet high, erect or
prostrate, forming dense thicket like clumps. The bark of the trunk
is dark brown to gray, 1/8 to 1/4 inches thick, the fissured surface
being broken into thin persistent scales. The branches are stout, rigid,
marked with minute pale lenticels, puberulous during the first summer,
and unarmed with occasional lateral spinescent or spur-like branchlets.
The winter buds are about 1/8 inch long, obtuse, and covered with chest-
nut brown scales. The leaves are oval, ovate or obovate, 4.0 to 6.8
cms. long by 2.0 to 3.5 cms. wide, on short, stout pubescent petioles,
7 to 12 mm. long; acute, finely and sharply serrate with glandular tipped
teeth when young; at the base of the blade are two, occasionally one,
rather large, round, stalked, dark glands, sometimes eglandular; slightly
puberulous to glabrous on the upper surface, while on the lower surface
quite pubescent along the midrib and in the axils of the primary veins,
slightly along the cross veins of certain varieties. Stipules linear, lobed,
pubescent, glandular serrate and very early deciduous.

A STUDY IN PLANT VARIATION 251

The flowers, which appear from the middle of April to the middle of
May before or with the unfolding of the leaves, are borne in 2 to 4 flowered
umbels, 10 to 15 mm. broad when fully expanded; pedicels and calyx
rather strongly pubescent, pedicels 7 to 15 mm. long; calyx tube campan-
ulate 2 to 4 mm. deep, calyx lobes ovate, rounded at the apex with in-
flexed edges, pubescent on both sides. Petals white or pinkish, oblong
or oblong ovate, 7 mm. long and 4 mm. broad, contracted to claws at
the base. Stamens 25 to 35, filaments of two lengths; style stout, 10
to 15 mm. long, glabrous and truncate at the tip. Carpel ellipsoid,
glabrous containing two pendulous ovules. The fruit which is frequently
produced in great quantities, ripens from August 1 to October 1; it
is borne on short stout stems, usually of the previous year's growth
sub-globose to slightly oval, and varies in size from 1.0 to 1.7 cms. in
diameter to 1.0 to 1.6 cms. long. The skin is thick and rather tough,
blue, purple, red or yellow, covered with a light waxen bloom and abund-
dantly flecked; the flesh is of a watery greenish yellow, juicy astringent
and decidedly free from the turgid stone which is rounded or compressed.
The stone varies in size from 6.5 x 5.4 mm. to 11. 0x7. 3 mm in diameter
and 7.8 to 12.5 mm. long, pointed at both ends, ridged on the ventral,
slightly grooved on the dorsal suture.

1. Primus maritima var. coerulea magna
(Large Blue Black)

Form A low spreading bush 3 to 6 feet high forming dense thicket

like patches, branching very irregular with short, stout
branches which usually stand erect, giving it the aspect
of gnarled and stunted in its habit of growth.

Bark Old stem rather stout, rough, cracked into thin dark gray

plates which tend to persist, mostly lichen covered. Young
branches smooth, dark purplish brown, numerous light
colored lenticels, and numerous spinescent lateral branches.
Youngest twigs smooth, numerous, elongate and rather
stout, puberulous, epidermis silvery gray and when shed
uncovers a light reddish brown bark.

Lenticels Numerous, small, conspicuous, irregularly distributed over
the branches, rounded and more elevated on the youngest
twigs, darker and flattened on the young branches.

Buds Alternate, conical shaped, sharp pointed, brown to gray

covered with numerous triangular scales, puberulous along
the margins; covered with a silvery gray waxy coating
which is shed in early spring.

252 PENNYP ACKER— ON THE BEACH PLUM

Flower Buds more precocious in early spring, chestnut
brown, rather abundantly scattered along the twigs,
frequently compounded with a single leaf bud.
Leaf Buds smaller, more pointed, dark gray and terminal,
bursting after the bush is in full bloom.

Leaf Scar Crescent shaped, with three small inconspicuous fibro-
vascular bundle scars.

Leaf Ovate, alternate, 3.5 cms. wide 6.8 cms. long, petiole

LO cms. long and finely pubescent; acute finely and sharply
serrate with small dark brown glandular tipped teeth;
rounded and biglandular at the base with two dark brown
stalked glands; slightly pubescent along the midrib, very
slightly along the primary veins on the upper surface,
pubescent along the midrib and in the axils of the primary
veins on the lower surface; stipules lobed, glandular-
serrate, and early deciduous. At maturity thick and firm,
dark green above, paler below.

Flower Appearing from May 1 to 15 slightly before the leaves, 15-

20 mm. broad on pubescent pedicels 15 mm. in length,
usually in 3 or 4 flowered umbels, sepals ovate, rounded
at the apex, pinkish along the margins, pubescent; calyx
tube campanulate 2 mm. deep, glandular on the inner sur-
face, pubescent on the outer; petals white turning pinkish
on fading, ovate, rounded at the apex, contracted at
the base into short claws; stamens 35, from 8 to 12 mm.
long, glabrous, turning pink on fading, inserted on the
calyx tube; style 15 mm. long, glabrous with truncate
stigma; carpels simple, ellipsoid, glabrous.

Fruit Ripening August 5 to 25 of a deep blue black color, spheri-

cal or slightly flattened at the blossom end 17x16 mm.
covered with a light waxen bloom, flecked with numerous
light colored spots; weight 3.0 to 3.9 grams, flesh of watery
green color; pulp free from the stone, sweet and pleasant,
abundance of tannin in a layer beneath the skin; stone
oval, compressed, reddish in color, 12.5 mm. long and
11.0x7.3 mm. thick, weight .54 gram, acutely ridged on
the ventral, shghtly grooved on the dorsal suture.

A STUDY IN PLANT VARIATION

253

Form

Bark

Lenticels

Buds

Leaf Scars

Leaf

Flower

2. Prutius maritima Var. coeruJea parva
(Small Blue Black)
Low spreading bush 3 to 6 feet high, branching irregular,
gnarled and stunted in its habit of growth.

Old stem rather slender, elongated, rough, cracked into
thin dark gray plates which tend to persist. Young
branches smooth dark purplish brown with numerous
light gray flattened lenticels, and few spinescent branches.
Youngest twigs smooth, numerous, elongate, slender,
puberulous, epidermis silvery gray and when shed uncovers
a light reddish brown bark.

Numerous, small, rather inconspicuous, irregularly dis-
tributed over the branches, rounded and elevated on the
youngest twigs, darker and more flattened on the younger
branches.

Alternate, conical shaped, sharp pointed, brown to gray
covered with numerous triangular scales, puberulous along
the margins; covered with a silvery gray waxy coating
which is shed in early spring.

Flower buds lateral, more precocious in early spring, chest-
nut brown, very abundantly scattered along the branches,
frequently compounded with a single leaf bud.
Leaf buds smaller, more pointed, dark gray and terminal
bursting after the bush is in full bloom.

Crescent shaped with three small inconspicuous fibro-
vascular bundle scars.

Ovate, alternate, 3.0 cms. wide 5.7 cms. long, petiole 1.0
cms. long and finely pubescent; acute, finely and sharply
serrate with small dark brown glandular tipped teeth;
rounded and biglandular at the base with two dark brown
stalked glands; pubescent only along the midrib on the
upper as well as the lower surface, blade glabrous; stipules
lobed, glandular-serrate, and very early deciduous. Thick
and firm at maturity, dark green above, paler below.
Appearing from May 1 to 15 slightly before the leaves, 14
to 18 mm. broad on pubescent pedicels 13 mm. in length,
usually in 3 or 4 flowered umbels, sepals ovate, rounded
at the apex, pinkish along the margins, calyx tube cam-
panulate 2 mm. deep, glandular on the inner surface,

254 PENNYP ACKER— ON THE BEACH PLUM

pubescent on the outer; petals white turning pinkish on
fading, ovate, rounded at the apex, contracted at the base
into short claws; stamens 35, from 8 to 10 mm. long,
glabrous, turning pink on fading, inserted on the calyx
tube; style 12 mm. long, glabrous with truncate stigma;
carpels simple, ellipsoid; glabrous.
Fruit Ripening August 10-30, of a deep blue color, sub-globose

to slightly oval, 15 x 14 mm. covered with a light waxen
bloom, flecked with numerous light colored spots; weight

1.4 to 1.85 grams; flesh of a watery green color, pulp free
from the stone, sweet and pleasant, abundance, of tannin
deposited in a layer beneath the skin; stone oval to elongate
slightly compressed reddish in color, 11 mm. long and

7.5 x 5.8 mm. thick, weight .22 grams, acutely ridged on
the ventral, slightly grooved on the dorsal suture.

3. Prunus maritima Var. praecox
(Small Inland Blue Black)
Form A low spreading bush 3 to 5 feet high forming dense thicket-

like patches, branching irregular with slender short bran-
ches, erect or prostrate often drooping forming a broad
crown.
Bark Old stem rather slender, rough, cracked into thin dark

gray plates which tend to persist, mostly lichen covered.
Young branches smooth, dark purpHsh brown, numerous
light colored lenticels and numerous spinescent branches.
Youngest twigs smooth, numerous, short and stout, puberu-
lous, epidermis silvery gray and when shed uncovers a
light reddish brown bark.
Lenticels Numerous, sm.all, conspicuous, irregularly distributed
over the branches, rounded and elevated on the youngest
twigs, while darker and flattened on the young branches.
Buds Alternate, conical, sharp pointed, brown or gray, covered

with numerous triangular scales, puberulous along the
margins, covered with a silvery gray waxy coating which
is shed in early spring.

Flower buds more precocious in early spring, chestnut
brown, very abundantly scattered along the branches,
frequently compounded with a single leaf bud near the
end of the twigs.

A STUDY IN PLANT VARIATION 255

Leaf buds smaller, more pointed, dark gray, and terminal,
bursting when the bush is in full bloom.

Leaf Scar Crescent shaped, with three small inconspicuous fibro-
vascular bundle scars.

Leaf Ovate, 5.5 cms. long and 2.3 cms, wide, petiole .7 cm. long,

apex acute, finely and sharply serrate wtih small dark
brown glandular teeth; rounded and biglandular at the
base with two dark brown stalked glands; pubescent along
midrib and primary veins above, very pubescent below
along midrib, primary veins and cross veins; stipules lobes,
glandular-serrate and early deciduous. At maturity thick
and firm, dark green above, paler below.

Flower Appearing from April 20 to May 5, slightly before the

leaves, 10 to 15 mm. broad on pubescent pedicels, 10 mm.
in length, in 2 to 4 flowered umbels, sepals ovate, rounded
at the apex, pinkish along the margins, pubescent, calyx
tube campanulate 2 mm. deep, glandular on the inner
surface, pubescent on the outer, petals white turning
pinkish on fading, ovate, rounded at the apex, contracted
at the base into short claws; stamens 25, from 5 to 10
mm. long, glabrous turning pinkish on fading, inserted
on the calyx tube; style 12 mm. long, glabrous with trun-
cate stigma, carpels simple, ellipsoid, glabrous.

Fruit Ripening August 1 to 20, of a deep blue black color, spheri-

cal or slightly flattened at the blossom end, 11.5 x 12.5
mm. covered with a light waxen bloom, flecked with num-
erous light colored spots; weight 1.1 to 1.6 grams, flesh
of a watery green color, pulp free from the stone, sweet
and pleasant, layer of tannin beneath the skin; stone oval
to rounded, reddish in color, 7.8 mm. long and 6.5 x 5.4
mm. thick, weight .10 to .15 grams, ridged on the ventral,
slightly grooved on the dorsal suture.

4. Prunus maritima Var. purpurea magna
(Large Purple)

Form A low spreading bush 3 to 6 feet high, forming dense

thicket like patches; branching rather regular and erect,
with numerous spinescent lateral branches, general ap-
pearance less gnarled and stunted in its habit of growth.

256
Bark

Lenticels

Buds

Leaf scar
Leaf

Flower

PENNYP ACKER— ON THE BEACH PLUM

Old stem rather stout, rough, cracked into thin dark gray
plates which tend to persist, frequently lichen covered.
Young branches smooth, dark purplish brown, numerous
light colored lenticels, and numerous spinescent lateral
branches. Youngest twigs numerous, elongate and stout,
puberulous, epidermis silvery gray rather persistent, when
shed uncovers a light reddish brown bark.
Numerous, large, very conspicuous, irregularly distributed
over the branches, rounded and more elevated on the
youngest twigs, darker and flattened on the young branches.
Alternate, conical shaped, sharp pointed, brown to gray
covered with numerous triangular scales, puberulous along
the margins; covered with a silvery gray waxy coating
which is shed in early spring.

Flower buds more precocious in early spring, chestnut
brown, rather abundantly scattered along the twigs,
frequently compounded with a single leaf bud near the
end of the twigs. Leaf buds smaller, more pointed, dark
gray and terminal, opening after the bush is in full bloom.
Crescent shaped, with three small inconspicuous fibro-
vascular bundle scars.

Ovate, alternate, 6.8 cms. long and 3.4 cms. wide; petiole
L2 cms. long and finely pubescent; acute, finely and sharply
serrate with small dark brown glandular teeth; rounded
and biglandular at the base with two dark brown stalked
glands; pubescent along the midrib on the upper surface,
along the midrib and in the axils of the primary veins on
the lower surface; stipules lobed, glandular-serrate, and
early deciduous. At maturity thick and firm, dark green
above, paler below.

Appearing from May 1 to 15 slightly before or with the
leaves 15 to 20 mm. broad on pubescent pedicels 15 mm. in
length, usually in 3 or 4 flowered umbels, sepals ovate,
rounded at the apex, pinkish along the margins, pubescent;
calyx tube campanula te 2 mm. deep, glandular on the
inner surface, pubescent on the outer; petals white turning
pinkish on fading, ovate, rounded at the apex, contracted
at the base into short claws; stamens 30, from 8 to 12 mm.
long, glabrous turning pink on fading, inserted on the
calyx tube; style 20 mm. long glabrous with truncate
stigma; carpels simple ellipsoid, glabrous.

A STUDY IX PLANT \ARIATION

257

Fruit

Form

Bark

Lenticels

Buds

Leaf Scar

Ripening August 15 to September 5, of a reddish purple
color, spherical to slightly oval 17 x 16 mm. covered with
a light waxen bloom, flecked with numerous light colored
spots; weight 3.25 to 3.6 grams; flesh of a watery greenish
yellow color, pulp decidedly free from the stone, sweet
and pleasant, tannin layer deposited beneath the skin;
stone oval and compressed, reddish in color, 13.5 x 10.0
mm. and 6.5 mm. thick, weight .40 to .50 grams, acutely
ridged on the ventral, slightly grooved on the dorsal suture.

5. Primus maritima Var. purpurea parva
(Small Purple)
A low spreading bush 3 to 6 feet high forming dense thicket
like patches; branching rather regular, erect, with numerous
spinescent lateral branches. General appearance less
gnarled and stunted.

Old stem rather stout, rough, cracked into thin dark gray
plates which tend to persist, mostly lichen covered. Young
branches smooth, dark purphsh brown, numerous light
gray lenticels and numerous spine-like lateral branches.
Youngest twigs smooth, numerous, elongate and rather
stout, puberulous, epidermis silvery gray and when shed
uncovers a light reddish brown bark.

Numerous, large, very conspicuous, irregularly distributed
over the branches, rounded and more elevated on the
youngest twigs, while darker and more flattened on the
young branches.

Alternate, conical shaped, sharp pointed, brown or gray,
covered with numerous triangular scales, puberulous along
the margins; covered with a silvery gray waxy coating
which is shed in early spring.

Flower buds more precocious in early spring, chestnut
brown, rather abundantly scattered along the twigs,
frequently compounded with a single leaf bud near the
end of the twigs.

Leaf buds smaller, more pointed, dark gray and terminal,
bursting after the bush is in full bloom.

Crescent shaped, with three small inconspicuous fibro-
vascular bundle scars.

258 PENNYP ACKER— ON THE BEACH PLUM

Leaf Ovate to obovate, alternate 2.0 cms. wide 4.0 cms. long,

petiole 1.0 cms. long and finely pubescent, acute, finely
and sharply serrate with small dark brown glandular tipped
teeth, rounded and biglandular at the base with two dark
brown stalked glands; pubescent only along the midrib
on the upper and on the lower surfaces, blade glabrous;
stipules lobed, glandular serrate, and early deciduous. At
maturity thick and firm, dark green above, paler below.

Flower Appearing from May 1 to 15 slightly before or with the

leaves, 10 to 15 mm. broad on pubescent pedicels 15 mm.
in length, usually in 3 or 4 flowered umbels, sepals ovate,
rounded at the apex, pinkish along the margins, pubescent;
calyx tube campanula te 2 mm. deep, glandular on the inner
surface, pubescent on the outer; petals white turning pink-
ish on fading, ovate rounded at the apex, contracted at
the base into short claws; stamens 30, from 5 to 10 mm.
long, glabrous, turning pinkish on fading, inserted on the
calyx tube; style 10 mm. long, glabrous with truncate
stigma; carpels simple, ellipsoid, glabrous.

fruit Ripening September 1 to October 1, or until frost overtakes

it, of deep purple color, spherical or sUghtly oval 13 x 12
mm. and Q x 10 mm. in the late variety, covered with a
light waxen bloom, flecked with numerous light colored
spots; weight .50 to 1.50 grams, flesh of a watery greenish
yellow color, pulp in the late variety tending to cling to
the stone and with a very bitter astringent taste, dense
layer of tannin beneath the skin; stone oval, reddish in
color, 8.5 mm. long and 7.8 x 5.5 mm. thick, weight .15
to .25 grams, ridged on the ventral, slightly grooved on
the dorsal suture.

6. Prunus maritima var. rubra magna
(Large Red)

Form A low spreading bush 3 to 6 feet high forming dense thicket

like patches; branching irregular and erect, with few
spinelike lateral branches. Some bushes appear quite
stunted in their habit of growth.

Bark Old stem stout, rough, cracked into thin dark gray plates

which tend to persist, mostly lichen covered. Young
branches smooth, dark purplish brown, numerous light

A STUDY IN PLANT VARIATION

259

colored lenticels, and few spine like lateral branches.
Youngest twigs smooth, numerous, elongate and rather
stout, puberulous, epidermis silvery gray and when shed
uncovers a light reddish brown bark.
Lenticels Numerous, small, inconspicuous, irregularly distributed
over the branches, rounded and more elevated on the
youngest twigs, darker and more flattened on the younger
branches.
Buds Alternate, conical shaped, sharp pointed, brown or gray,

covered with numerous triangular scales, puberulous along
the margins; covered with a silvery gray waxy coating
which is shed in early spring.

Flower buds lateral, more precocious in early spring,
chestnut brown, rather abundantly scattered along the
branches, frequently compounded with, a single leaf bud
near the ends of the twigs.

Leaf Buds smaller, more pointed, dark gray and terminal,
bursting after the bush is in full bloom.
Leaf Scar Crescent shaped, with three small inconspicuous fibro-

vascular bundle scars.

Leaf Ovate, alternate, 3.2 cms. wide and 6.8 cms. long, petiole

LI cms. long and finely pubescent; acute, finely and sharply

serrate with small dark brown glandular tipped teeth;

rounded and biglandular at the base with two dark brown

stalked glands; pubescent along the midrib on the upper

surface and along the midrib and in the axils of the primary

veins on the lower surface; stipules lobed, glandular-serrate,

and early deciduous. At maturity thick and firm, dark

green above, paler below.

Flower Appearing May 1 to 15, slightly before or with the leaves,

15 to 20 mm. broad on pubescent pedicles 10 mm. m

length, usually in 3 or 4 flowered umbels, sepals ovate,

rounded at the apex, green, pubescent; calyx tube cam-

panulate 2 mm. deep, glandular on the inner surface,

pubescent on the outer; petals white, ovate, rounded at

the apex, contracted at the base into short claws; stamens

30, from 8 to 12 mm. long, glabrous inserted on the calyx

tube; style 15 mm. long, glabrous with truncate stigma

carpels simple, ellipsoid, glabrous.

260
Fruit

Form

Bark

Lenticels

Buds

Leaf Scar
Leaf

PENNYPACKER— ON THE BEACH PLUM

Ripening August 25 to September 15, of a dark red color,
spherical or slightly flattened at the blossom end, 16 x 15
mm. covered with a light waxen bloom, flecked with nu-
merous light colored spots; weight 2.40 to 3.0 grams,
flesh of a watery reddish yellow color; pulp free from the
stone, taste sweet, tannin less abundant than before;
stone oval, compressed, reddish in color, 12.5 mm. long
and 9.5 x 6.5 mm. thick, weight .40 to .50 grams, acutely
ridged on the ventral, sUghtly grooved on the dorsal suture.

7. Primus maritima var. rubra parva
(Small Red)
A low spreading bush 3 to 6 feet high forming dense thicket
like patches, branching rather regular and erect, with
few spine-like lateral branches; bushes frequently appear
gnarled and stunted in their habit of growth.
Old stem slender elongate, rough, cracked into thin dark
gray plates which tend to persist, mostly lichen covered.
Young branches smooth, dark purplish brown with few
light colored lenticels, and few spine-like lateral branches.
Youngest twigs smooth, numerous, elongate, slender,
puberulous; epidermis silvery gray, persistent, when shed
uncovers a light reddish brown bark.

Few, small, inconspicuous, irregularly distributed over
the branches, rounded and more elevated on the younger
twigs, darker and more flattened on the young branches.
Alternate, conical shaped, sharp pointed, brown to gray,
covered with numerous triangular scales, puberulous along
the margins; covered with a silvery gray epidermis which
is shed in early spring.

Flower buds more precocious in early spring, chestnut
brown, rather abundantly scattered along the twigs, fre-
quently compounded with a single leaf bud.
Leaf Buds smaller, more pointed, dark gray and terminal,
bursting after the bush is in full bloom.
Crescent shaped, with three small inconspicuous fibro-
vascular bundle scars.

Ovate, alternate, 2.0 cms. wide, 4.5 cms. long, petiole
0.9 cms. long and finely pubescent, acute, finely and sharply
serrate with small dark brown glandular tipped teeth;

A STUDY IN PLANT VARIATION

261

Flower

Fruit

rounded and biglandular at the base with two dark brown
stalked glands; pubescent only along the midrib on the
upper and on the lower surfaces, blade glabrous; stipules
lobed, glandular-serrate, and early deciduous. At maturi-
ty thick and firm, dark green above, paler below.
Appearing May 1 to 15, before the leaves, 10 to 15 mm.
broad on pubescent pedicels 15 mm. long, usually in 3 or 4
flowered umbels, sepals ovate, rounded at the apex, green,
pubescent, calyx tube campanulate 2 mm. deep, glandular
on the inner surface, pubescent on the outer; petals white,
ovate, rounded at the apex, contracted at the base into
short claws; stamens 30, from 5 to 10 mm. long, glabrous,
inserted on the calyx tube; style 10 mm. long, glabrous,
with truncate stigma; carpels s mple, ellipsoid, glabrous.
Ripening August 25 to September 15, of a dark red color,
spherical or slightly flattened at the blossom end, 14 x 13
mm. covered with a light waxen bloom, flecked with,
numerous light colored spots; weight 1.50 to 1.90 grams,
flesh of a watery reddish yellow color; pulp free from the
stone, sweet, tannin layer under the skin; stone oval,
compressed, reddish in color, 9.0 mm. long and 8.5 x 5.8
mm. thick, weight .25 to .30 grams, acutely ridged
on the ventral, slightly grooved on the dorsal suture.

Form

Bark

Lenticels

8. Primus maritima var. lutea magna
(Large Yellow)

A low spreading bush 3 to 6 feet high forming dense thicket-
like patches; branching irregular and elongate, giving the
bush a rather straggling appearance as to its habit of
growth.

Old stem rather stout, rough, cracked into thin dark gray
plates which tend to persist, mostly lichen covered. Young
branches smooth, dark purplish brown, with numerous
light gray lenticels. Youngest twigs smooth, numerous,
elongate, slender, puberulous, epidermis very gray,
persistent, when shed uncovers a light reddish brown bark.
Numerous, small, conspicuous, irregularly distributed over
the branches, rounded and more elevated on the youngest
twigs, darker and more flattened on the younger branches.

262 PENNYP ACKER— ON THE BEACH PLUM

Buds Alternate, conical shaped, sharp pointed, brown to gray,

covered with numerous triangular scales, puberulous along
the margins; covered with a silvery gray epidermis which
is shed in the early spring.

Flower buds lateral, more precocious in early spring, chest-
nut brown, rather abundantly scattered along the twigs,
frequently compounded with a single leaf bud.
Leaf buds smaller, more pointed, dark gray and terminal,
bursting after the bush is in full bloom.

Leaf Scar Crescent shaped, with three small inconspicuous fibro-
vascular bundle scars.

Leaf Ovate to obovate, alternate, 3.5 cms. wide 6.8 cms. long,

petiole .8 cms. long and finely pubescent; acute, finely
and sharply serrate with small dark brown glandular tipped
teeth, rounded and biglandular at the base with two dark
brown stalked glands; pubescent along the midrib and
slightly along the leading veins on the upper surface, while
on the lower surface pubescent only along the midrib;
stipules lobed, glandular serrate, and early deciduous. At
maturity thick and firm, dark green above, paler below.

Flower Appearing May 1 to 15 before the leaves, 15 to 20 mm.

broad on pubescent pedicels 15 mm. in length, usually in
3 or 4 flowered umbels, sepals ovate, rounded at the apex,
green pubescent, cayx tube campanulate, 2 mm. deep,
glabrous on the inner surface, pubescent on the outer;
petals white, rounded at the apex, contracted at the base
into short claws; stamens 30, from 8 to 12 mm. long, gla-
brous, inserted on the calyx tube, style 15 mm. long, gla-
brous, with a truncate stigma, carpels simple, ellipsoid,
glabrous.

Fruit Ripening August 10 to 30, of a deep yellow color developing

a pinkish tinge when fully mature, spherical to slightly
flattened at the blossom end 17 x 16 mm. covered with a
light waxen bloom, flecked with numerous light colored
spots; weight 2.7 to 3.25 grams, flesh of a watery yellow
color, pulp decidedly free from the stone, sweet and pleas-
ant, tannin layer slight, much less than in the preceeding
varieties, stone oval, slightly compressed, yellowish in
color, 10 mm. long and 9.5 x 7.5 mm. thick, weight .35 to
.45 grams, slightly ridged on the ventral, slightly grooved
on the dorsal suture.

A STUDY IN PLANT VARIATION

263

Form

Bark

Lenticels

Buds

Leaf Scar
Leaf

Flower

9. Prunus maritima var. Intea parva.
(Small Yellow)

A low spreading bush 3 to 6 feet high, forming dense thick-
et-like patches; branching irregular and elongate, giving
the bush a rather straggUng appearance as to its habit
of growth.

Old stem slender, elongate, rough, cracked into thin dark
gray plates which tend to persist, mostly lichen covered.
Young branches smooth, dark purpHsh brown, with num-
erous light colored lenticels. Youngest twigs smooth,
numerous, slender, elongate, puberulous, epidermis silvery
gray, persistent, when shed uncovers a light reddish brown
bark.

Numerous, small, conspicuous, irregularly distributed over
the branches, rounded and more elevated on the youngest
twigs, darker and more flattened on the younger branches.

Alternate, conical shaped, sharp pointed, brown to gray
covered with numerous triangular scales, puberulous along
the margins; covered with a silvery gray waxy coating
which is shed in early spring.

Flower hiids lateral, more precocious in early spring, chest-
nut brown, rather abundantly scattered along the twigs,
frequently compounded with a single leaf bud near the
end of the branch.

Leaf buds smaller, more pointed, dark gray and terminal,
bursting after the bush is in full bloom.
Crescent shaped, with three small inconspicuous fibro-
vascular bundle scars.

Ovate, alternate, 2.5 cms. wide 5.7 cms. long, petiole 1.0
cms. long and finely pubescent; acute, finely and sharply
serrate with small dark brown glandular tipped teeth;
rounded and biglandular at the base with two dark brown
stalked glands; very pubescent along the midrib, veins and
cross veins on the upper surface, slightly less pubescent
along the midrib, veins and cross veins on the lower sur-
face; stipules lobed, glandular-serrate, and early deciduous.
At maturity thick and firm, dark green above, paler below.
Appearing from May 1 to 15, slightly before, or with the
leaves, 10 to 15 mm. broad on pubescent pedicels 20 mm.

264 PENNYPACKER— ON THE BEACH PLUM

long usually in 3 or 4 flowered umbels, sepals ovate, rounded
at the apex, green, pubescent, calyx tube campanulate 2
mm. deep, glandular on the inner surface, pubescent on
the outer; petals white, ovate, rounded at the apex, con-
tracted at the base into short claws; stamens 30, from 5
to 10 mm. long, glabrous, inserted on the calyx tube;
style 10 mm. long, glabrous, with truncate stigma; carpels
simple, ellipsoid, glabrous.
Fruit Ripening August 10 to 30, of a deep yellow color, spherical

to slightly flattened at the blossom end, 14 x 13 mm.
covered with a Hght waxen bloom, flecked with numerous
light colored spots; weight 1.45 to 1.85 grams, flesh of a
watery yellow color, pulp decidedly free from the stone,
sweet and pleasant, tannin layer slight; stone oval, slightly
compressed, yellowish in color, 8.5 mm. long and 7.0 x 6.0
mm. thick, weight .25 to .35 grams, slightly ridged on
the ventral, slightly grooved on the dorsal suture.

IV. Economic Value

One is immediately impressed, on seeing the vast expanse of P. mari-
tima growing on the sandy soils of the Atlantic coast, and fruiting in
abundance, by the thought that it should have a high economic value.
The characters which would seem to warrant its horticultural possibiU-
ties are its great hardiness, late blooming, enormous productiveness,
and ability to withstand adverse conditions. These were recognized
by Burbank several years ago and every effort was made by him to
cross it with some of the larger and finer species. This cross, however,
was not effected for several years because the Beach Plum blossoms
very late, long after all other plums have shed their bloom. Burbank
early recognized the value of the introduction of foreign blood into the
plum family, and from his hybridizing experiments — commenced in
1885 and continued up to the present time — has produced 65 varieties.
Of these, 38 have been developed from Asiatic, 14 from American, and
13 from European stock.

It is interesting to note, that of the six American species of Prunus
used in his experiments, maritima should have proved to be the most
important, as it is the one to which the greatest interest is attached. Of
the American species used all are unusually hardy. Cold does them
no harm even in the most northern parts of the United States. It

A STUDY IN PLANT VARIATION 265

seems therefore that the best horticultural results can be obtained if
hardy fruits are more and more developed and selected for the colder
sections of our country.

The hybridizing experiments of Burbank included the crossing of the
present species with hybrids of the Japenese Plum (P. tr if lor a). There
resulted from these crosses three interesting varieties of which the
"Giant Maritima" seems to be the most marked. The fruit begins to
ripen in California early in July and when ripe is of a deep crimson
covered with a thin pale bloom. The flesh until fully ripe is very firm
and solid, but it breaks down quickly when ripe. It is honey yellow,
with a pale greenish tinge. The quaUty is good. The fruit is fragrant,
and as large as any other plum known in 1905. It was grafted into
numerous older trees and appears to be a strong grower. He states that
it will never prove of much commercial value as it lacks firmness of
texture. The second variety was called "East." It was not as good
as he had anticipated, it was too soft for shipping, but proved to be a
desirable variety for home consumption. In quality of fruit, it was
probably inferior to the best Japanese hybrids. It ripens from August 1
to 15. "Pride," the third variety, is of little value as a shipping plum.
It ripens too quickly so that it will not stand shipping to any great dis-
tance. It is apple shaped, dark red, a good grower, an excellent bearer
and ripens about August 20th. He further states that in addition to
these, nearly 2000 other promising maritima hybrids are now being
grow^n from these crosses. Many of them are excellent in habit, produc-
tiveness, and hardiness and from them he hopes to introduce many
new forms.

The writer has grown in the Botanic Gardens of the University of
Pennsylvania P. maritima budded on domestic stock, which at the pre-
sent time is growing very rapidly, although no blooms have appeared
as yet. It is the writer's hope that interesting forms may result from
proper cultivation and selection of the budded stock, and in this connec-
tion he desires to express his personal gratitude to Messrs. Hoopes Bro.
and Thomas, West Chester, Pa. for their skill in efifecting these graft
unions for him.

This work was conducted under the direction of Professor John M.
Macfarlane, of the University of Pennsylvania, to whom the writer is
indebted for many helpful suggestions and criticisms, and for which he
expresses his full appreciation.

266 PENNIT ACKER— ON THE BEACH PLUM

V. Summary

The following summarizes the body of new facts obtained by the
writer during the present investigation.

1. While P. maritima is typically a seaside plant that grows in loose
sandy soil, there are inland localized areas along the eastern sea board
where it occurs. These probably represent isolated, and stranded patches
of individuals, left from shore lines that once existed in previous geologi-
cal periods.

2. While the average blossoming period has been determined to be
the 2d of May, varieties are observed along shore lines where they are
exposed to sweeping sea breezes and which do not open until May 15th.

3. The most noteworthy characteristics of P. maritima are shown by
the writer to be its marked variation in size, in mode of branching, and
in vigor of the shoots; in time of appearance, in size and in hairiness of
the leaves; in size and in color of the petals, as well as in their blossoming
before, during, or after the commencing expansion of the foliage leaves.
The above is in line with the extreme variability of the fruits as already
noted by previous observers, and would suggest that the species is under-
going marked mutational variations.

4. The flowers are shown to be andro-dioecious and this evidently
affords a key to the conditions seen in autumn when some bushes are
fruitless, others sparsely fruitful, and still others bear abundantly. The
practical value of this observation in connection with possible future
cultivation by man is evident.

5. As a rule, those bushes which bear the darker or deeper colored
fruits of purple and blue tint have a general purplish color that extends
even to the petals and stamens.

6. Interesting petiolar glands have been found to occur near the
junction of the petiole with the lamina as well as laminar tooth-glands.

7. A concave nectary lines the interior of the receptacular cup that
surrounds the pistil and this excretes a large amount of nectar.

8. The writer confirms the views previously expressed regarding the
striking variability and resulting types of the fruit.

9. The extreme variation in size, shape, and weight of the stones in
the different varieties has been determined and compared.

10. The variation forms in this highly variable plant recognized by
the writer have been grouped under nine heads, each of which has been

A STUDY IN PLANT VARIATION 267

given a definite varietal name, and all of these tend to appear in regions
where the Beach Plum grows from Cape May Point to Cape Cod pen-
insula.

11. The primitive color was probably greenish yellow or greenish red.
Through increasing transformation of the chlgroplasts into bright
chromoplasts pure yellow fruits were secured along one evolutionary
Une; through development of a red, purple and eventually blue color the
chloroplasts became concealed, and so the climax of combined size and
color evolution was reached in the large blue black.

12. Many bushes belonging to several varieties, particularly the
small blue black and the small purple, are often destructively punctured
by a weevil during the early stages of fruit maturation, or about two
weeks after the flowering period. This produces a hardened protective
secretion rich in tannin but which causes deteriotation in the quaUty
of the fruit.

13. While the purple and the blue black fruits are often rich in tannin,
the yellow fruits are comparatively poor in this.

14. The Beach Plum, like the cultivated plum, {P. insititia) and the
Sand Cherry {P. Besseyi), is a mutational species, which through the
agency of environmental factors, that are still hard to determine accur-
ately, is undergoing change in individual plants, not along one but along
several lines of variation.

15. The author agrees with the views put forth by Macfarlane as to
the possible high economic value of the shrub for the future in its dif-
ferent varieties. He also records the interesting hybridizing experi-
ments of Burbank with the present species and the Japanese Plum {P.
triflora), as holding out hopes for origin of many new forms by hybridi-
zation.

268

PENNYPACKER— ON THE BEACH PLUM

1. Bailey, L. H.

2. Beck, L. C.

3. Bigelow, Jacob

4. Britton, N. L.

5. Britton and Brown

6. Chapman, A. W.

7. Curtis, S. and
Hutchinson, J.

8. Darby, John

9. De Candolle, A. P.

10. Don, George

11. Eaton and Wright

12. Ehrhart, Fred.

13. Elliott, Stephen

14. Emerson, G. B.

15. Farlow, W. G.

16. Gray, Asa

17. Loudon, J. C.

18. Macfarlane, J. M.

19. Marshall, Humphrey

20. Michaux, Andreas

21. Nuttall, Thomas

22. Pursh, Frederick

23. Roemer, M. J.

24. Small, J. K.

25. Spach, Edouard

26. Sprengel, Curtis

27. Stone, Witmer

28. Torrey, John

29. "

30. Torrey and Gray

VI. Bibliography
An Account of the Cultivated Native Plums and Cherries.
Bull. 38, Cornell Exp. Sta., 1892.
Botany of the Northern and Middle States, 1833, 95.
Florula Bostoniensis. 2d edition, 1824, 193.
Manual of the Flora of the Northern States and Canada,
1901, 524-5.

An Illustrated Flora of Northern United States, Canada,
etc., 1913, Vol. 2, 325.

Flora of the Southern United States. 3d edition, 1897,
131.

Botanical Magazine, 4th series. Vol. 5, 1909, Tab 8289.

Botany of the Southern States, 1855, 299.

Prodromus Systematis Naturalis Regni Vegetabilis, Vol.

2, 1825, 533.

Miller's Gardeners Dictionary, Vol. 2, 1832, 499.

North American Botany. 8th edition, 1840, 189, 377.

Beitrage zur Naturkunde und den damit verwandten

Wissenschaften besonders der Botanik, Chemie, Haus und

Landwirthschaft und Apotheker kunst. Book 4, 1789, 17.

A Sketch of the Botany of South Carolina and Georgia.

Vol. 1, 1821, 543.

Trees of Massachusetts, 1846, 449.

Bulletin of the Bussey Institution, 1876, 440-453.

New Manual of Botany, by Robinson & Fernald. 7th

edition, 1908, 498.

Arboretum et fruticetum Britannicum, Vol. 2, 1838, 691.

The Beach Plum, Viewed from Botanical and Economical

Aspects. Contributions from Bot. Lab. U. of Pa., Vol.

2, 1904, 216-230.

Arbustum Americanum, 1785, 112.

Flora Boreali-Americana, Vol. 1, 1803, 284.

The Genera of North American Plants, 1818, 302.

Plants of North America, Vol. 1, 1814, 332.

Synopses Monographicae. Fas. 3, 1847, 57.

An Apparently Undescribed Species of Prunus from

Connecticut. Bull. Torr. Bot. Club. 24, 1897, 44-45.

Histoire Naturelle des Vegetaux, Vol. 1, 1834, 397.

Systema Vegetabihum, Vol. 2, 1825, 476.

Report of the New Jersey State Museum, 1910, 491.

Compendium of the Flora of the Northern and Middle

States, 1826, 199.

Natural History of New York. Part 2, Botany, Vol. 1,

1843, 194-195.

Flora of North America, Vol. 1, 1838, 408.

A STUDY IN PLANT VARIATION 269

31. Wangenheim, F. A. BeschrcibunReiniscrNordamerikanischerHolz-und Busch-

arten mil Anwendung auf deutsche Forsten, 1781, 103.

32. " " Beitrag zur teutschen holzgerechten Forstwissenschaft,

die Anpflanzung Nordamerikanischer Holzarten, mit
Anwendung auf teutsche Forste, betreffend, 1787.

33. Wight, W. F. Native .\mcrican Species of Prunus. Bull. 179. Bur.

Plajit Ind. U. S. Dept. of Agri., 1915, 55.

34. Willdenow, K. L. Enumeratio Plantarum, Vol. 1, 1809, 519.

VII. EXPLAN.\TION OF PlATES

Plate LXVT
Fig. 1 Branching form of P. maritima var. coeridea magna, in winter condition.

Fig. 2 Branching form P. maritima var. purpurea magna, in winter condition.

Fig. 3 Branching form P. maritima var. rubra parva, in winter condition.

Fig. 4 Branching form P. maritima var. lutea magna, in winter condition.

Plate LXVTI
Fig. 5 Blossoming branch P. maritima var. coerulca magna.

Fig. 6 Blossoming branch P. maritima var. purpurea magna.

Fig. 7 Blossoming branch P. maritima var. rubra parva.

Fig. 8 Blossoming branch P. maritima var. lutea parva.

Plate LXVIII
Fig. 9 Cross section of stem P. maritima var. purpurea parva, showing rings of

cork and phloem. X 25.
Fig. 10 Cross section of stem P. maritima var. coerulea magna, showing elements

of structure. X 25.
Fig. 11 Stipule of P. maritima var. coerulea magna, showing glandular serrate

structure. X 10.
Fig. 12 Cross section of midrib P. maritima var. rubra magna. X 25.

Fig. 13 Lower epidermis of P. maritima var. coerulea magna, showing stomatic

islands and conglomerate crystals of calcium oxalate. X 160.

Plate LXIX
Fig. 14 Fruits (natural size) P. maritima var. coeridea magna and var. coertdea

parva.
Fig. 15 Fruits (natural size) P. maritima var. purpurea magna and var. purpurea

parva.
Fig. 16 Fruits (natural size) P. maritima var. praecox showing bloom.

Fig. 17 Fruits (natural size) P. maritima var. lidea magna and var. lutea parva.

Plate LXX

Fig. 18 (a) Seeds (edge and end views) P. maritima var. coerulea magna.

Fig. 18 (b) Seeds (edge and end views) P. maritima var. coerulea parva.

Fig. 19 Seeds (edge and end views) P. maritima var. praecox.

Fig. 20 (a) Seeds (edge and end views) P. maritima var. purpurea magna.

Fig. 20 (b) Seeds (edge and end views) P. maritima var. purpurea parva.

Fig. 21 (a) Seeds (edge and end views) P. maritima var. rubra magna.

Fig. 21 (b) Seeds (edge and end views) P. maritima var. rubra parva.

Fig. 22 (a) Seeds (edge and end views) P. maritima var. lutea magna.

Fig. 22 (b) Seeds (edge and end views) P. maritima var. lutea parva.

^

Bot. Contrib. I'niv. Pain.

Vol. IV, Plate LXVl

L

Fig. 1

Fig. 2

Fig. 3 Fig. 4

PENNYPACKER ON BEACH PLUM.

Bot. Contrib. L'liiv. Pom.

Vol. IV, Plate LXVII

Fio. 5

Fig. 6

Fig. 7

Fig. 8

PENNYPACKER OX BEACH PLUM

Bot. Contrib. Univ. Penn.

Vol. IV, Plate LXVIII

Fig. 11

Fig. 13

PENNYPACKER ON BEACH PLUM

Bot. Contrib. Univ. Penn.

Vol n\ Plate LXIX

Fig. 14

Fig. 15

-^■s^.

Fig. 16 Fig. 17

PENNYPACKER ON BEACH PLUM

Bot. Contrib. Univ. Penn.

Vol. IV, Plate LXX

Fig. 18 a

A

#

« %

# #

« •

•

#

t #

Fig. 20

Fig. 18 b

• #

#•

Fig. 19

I

# #

• • '

♦ •

• •

B

• •

• •

Fig. 22 a

• t

• #

Fig. 21 b

PENNYPACKER ON BEACH PLUM

Fig. 22 b

•%

ON THE PRODUCTION OF
NEW CELL FORMATIONS IN PLANTS.

BY

William Randolph Taylor, B.S., M.S.
With Plates LXXI to LXXVIII

CONTENTS

Introduction 271

Experiments and Observations on Castanea 274

Material and Methods 274

Normal Histolog>' 274

New Tissue Formations in Phloem 276

Summary 281

Experiments and Observations on Herbaceous Plants 282

Material and Methods 282

Summary of Experiments 283

Effects of Injections on the General Growth 286

Discussion of the Graphs 288

Effects of Injections on the Stem Histology 289

Summary 293

Conclusions 294

Bibliography 294

Description of Plates 295

Introduction

For five years, from 1911-1916, Dr. Caroline Rumbold studied the
Chestnut Blight Disease in the Botanical Laboratory of the University
of Pennsylvania, endeavoring especially to develop a method of com-
bating the disease by the injection of chemicals into the trees. While
examining microscopically the trunks of certain of these trees, she dis-
covered that the structure of the bark in certain places differed strikingly
from thei. normal. Further study convinced her that these abnormalities
were dtie to the action of the injected substances and were formed in
response to some unusual physiological stimulus furnished by them.
At the suggestion of Dr. John M. Macfarlane, she embodied these ob-
servations in a paper (3) presented at the meeting of the American
Philosophical Society held on March 3, 1916. The importance of this
contribution consisted of the demonstration of unsuspected tissue-
formative powers located in the soft-bast region. As Dr. Rumbold

271

272 TA\"LOR— ON THE PRODUCTION OF

stated that she had no desire to institute further experiments to investi-
gate the meristematic power of plant tissues, it was suggested by Dr.
Macfarlane that the work be continued by the writer, to which Dr.
Rumbold agreed. Work was begun about the first of April 1916, and
sufficient material having accumulated to pe;rmit the drawing of just
conclusions, there is here presented a summary of the experiments
carried on by the writer during the past year. The writer is greatly
indebted to Dr. John M. Macfarlane for suggesting this work, and for
his constant supervision and encouragement while it was in progress.

Past studies of tissue abnormalities have been largely concerned
with wound-healing, gall-formation and so forth. The literature on
this subject has been admirably condensed by Kiister (1), The ma-
jor part of the work has been concerned with the general ability
to produce protective wound tissue, rather than to observe the capabilities
of change of each particular type of cell.

The most extensive tissue-production seems to have been obtained
by Simon (5) from the ends of twigs of Populus nigra kept in a water-
saturated atmosphere. Pustular outgrowths on leaves have been re-
ported, and also similar intumescences on stems. The mesophyll of
the leaves seems in the first case to have been the active layer. Methods
of reproducing these experimentally have been found by Kiister and
others, and Erwin F. Smith (6) has published a paper on the production
by simple chemical means of tumor-like outgrowths that extends this
line of investigation. In addition to producing superficial outgrowths,
he caused internal pith proliferations in the stems, these developing
internal vascular structures. His experiments on this particular line
he states to have been begun in June 1916. Schilberszky (4) by mechan-
ical means produced extrafascicular bundles in Phaseolus. This showed
that there existed in the layer of cells outside of the hard-bast tissue
formative powers that under extraordinary conditions might be used
by the plant to furnish itself with additional vascular tissue. These
cortical cells that gave rise to the meristem were primary tissue, not
secondary tissue derived from a cambium. On the contrary, the cells
which developed the xylem studied by Dr. Rumbold in Castanea were
soft-bast cells in secondary or metaphloem derived from a cambium.
This is one important difference between the two conditions; another
is the comparative age of the tissues involved in the two plants. It
would generally be considered that the cells in the bark of a woody
perennial, (in some of the cases observed several years old), would be
far less responsive than those of an annual.

NEW CELL FORMATIONS IN PLANTS 273

This notable contrast suggested that other tissues, if treated in the
right way, might respond and form highly interesting structures. Some-
thing of this sort is indicated in the production of certain galls, as shown
by Houard. The formation of wood from the pith has been reported
by Jaccard, Prilheux, and especially Maiile. Production of bundle-
cambia from non-vascular tissues is not unknown as a normal feature
of certain plants.

The work of the writer on the general subject of tissue modifications
may be divided into two parts. The first was an examination of a
large amount of Casianea material injected by Dr. Rumbold, to check
up her findings in that species. The second was a series of original
experiments, planned to extend the observations to other plants.

In pursuance of the second part of this program, apparatus was con-
structed for the injection of trees, and the first experiment (on California
Privet) was set up on April 18, 1916. Subsequently several species
of trees were used, including Populus carolinensis, Catalpa bignonioides,
Ailanthus glandulosa, and Paulownia tomentosa. These experiments did
not yield very satisfactory results, due in part at least to excessive dilution
of the substance injected. From the first the writer determined to con-
duct similar experiments on herbaceous plants, and the injection of
these was begun the first week in July, 1916. It was hoped that the
reactions of these would furnish the key to the very complex conditions
present in Castanea, and the results secured from Ricinus cut by the
end of September verified these conjectures. However, the number
of plants available during the early summer was not very large, so during
the winter of 1916 roots of Polygonum Sieboldii were started in the
greenhouses and the shoots used. This plant was further experimented
on outdoors the next spring (1917). Altogether, about a hundred in-
jections of various kinds have been made, and suitable controls main-
tained.

Briefly expressed, the object of this study was to produce special
cellular growth of all tissues capable of it, to determine what these
were, and what the morphology of the products of this growth might
be. In extension of the work of Dr. Rumbold, it was determined to
do this as far as practicable by the injection of various chemicals in
solution. As may easily be understood, it was not possible to eliminate
entirely the factor of mechanical injury. The mere puncture of the
needle in the work on herbaceous plants caused a slight reaction and com-
pUcated the result to some extent. Details of the behavior of the tissues
are given in the descriptions of individual cases.

274 TAYLOR— ON THE PRODUCTION OF

Castanea — "Paragon"

Material and Methods

The material used in the study of tissue modifications in the Chestnut
consisted of small grafted "Paragon" trees obtained from a large orchard
near Martic Forge, Lancaster County, Penna. In the hope of thereby
controlling the Chestnut Blight, Dr. Rumbold had injected into different
trees here during 1915 and previously, a large variety of chemicals (2).
On the 7th and 8th of November 1916, the writer went with Dr. Rumbold
to Martic Forge and cut down a number of trees. Sections of the trunks
of these were sent in to the laboratory and preserved for study.

The collection comprised trees that had been injected with various
chemicals, others that had been badly infected with the blight, some in
which large cankers had been carefully cleaned away a year or more previ-
ously, and others with minor injuries or normal structure for comparison.
As a result of the various vicissitudes that they had undergone, many
showed an extremely gnarled and twisted exterior with but little Hving
wood, while a few were completely dead.

In all cases a hand-lens examination of the material preceded prepara-
tion for sectioning on the microtome. Each specimen was sawn thru
at the point to be examined, and the surface smoothed with a sharp
chisel, when examination of the surface with a good hand-lens
(X 10) sufficed to show clearly the distribution and general nature
of any abnormal structures in the peripheral wood or the bark. Material
so prepared and coated with glycerin photographed easily (Plate LXXIV,
Fig. 11). Representative specimens of the material were cut out,
treated with hydrofluoric acid to remove the siKca, washed, and embedded
for sectioning in celloidin in the usual way.

Normal Histology of Castanea — "Paragon"
Before describing in detail the various alterations in the tissues
found in the neighborhood of the wounds of Castanea, it is necessary
that the normal structure of the stem be discussed.

The Pith in stems of the age used had long been lignified and dead.
Lignification seems to occur during the first year.

The Wood consists of elongated fiber cells and of prominent pitted
vessels. These latter are mostly formed in the spring, producing a
distinctly ring-porous wood. The walls of these pitted vessels are not
very greatly hardened, and the pits are small and transverse (Plate

NEW CELL FORMATIONS IN PLANTS 275

LXXIII, Fig. 9). In the fibers there are pits of a special form, similar
to those found in other members of this family. They are called by
Solereder (7) ''Bordered Pits." In form they are circular, the wall
being here rather heavily thickened. The central pore is slitlike, and
the direction of the slit on one side is generally the reverse of that on
the other, causing under low power the appearance of a cross (Plate
LXXIII, Fig. 9). This is more evident in the smaller pits, where the
slits cross each other at right angles, than it is in the largest ones, the
larger slits seen in these crossing more obliquely.

The Medullary Rays are very narrow, generally being but one cell
wide (Plate LXXII, Fig. 5). In passing thru the bark they cross directly
thru the hard-bast strands. The walls of the cells become lignified,
and have many simple pits in them.

The Phloem is of two sorts. Some of the cells are soft, in cross-
section somewhat rectangular, and in longitudinal section about twice
as long as broad. There is a secondary differentiation of these into
some that contain much tannin and into others that are comparatively
free from it. The tannin may be demonstrated by treatment with
ferric chloride in the usual way. In position the tannin cells may be
scattered, or they may form long rows two or three cells wide and circhng
the stem for a considerable distance. One or two of these rows (rarely
more) may lie in the ordinary soft bast between the hard-bast strands
of one year and those of the year preceeding. Many of the soft-bast
cells contain conglomerate crystals.

The other type of phloem or bast cell is lignified, and forms in the
older regions of the bark plates or strands of hard-bast cells. These
form each year a broken ring encircling the stem. The size of these
hard-bast strands varies considerably, altho they are fairly uniform
in thickness in any one year's ring. Around the margin of each
strand is a layer of cells containing tabular crystals. Stone-cells fre-
quently appear, and their presence cannot be attributed to any special
injury to the tissues. They are not distributed in any regular manner.
The Cortex seemed to be largely gone from the specimens examined,
but when it was present it showed a ring of sclerenchyma embedded
in the soft parenchyma. The outer cortex in young stems consists
of a collenchyma. Under the lenticels there are formed wedge-shaped
areas with the cells rounded instead of rectangular as is usual elsewhere.
Cork forms early. According to Solereder the cambium originates
from the outer layer of cortex cells. In the particular variety of Chestnut
studied, the surface of the bark remains smooth for a number of years,

276 TAYLOR— ON THE PRODUCTION OF

but little splitting occurring. When the age of the tree or injuries necessi-
tate it, cork-cambia form from the deeper soft-bast cells and cause the
shedding of fragments of the bark.

New Tissue Formation in the Phloem of Castanea "Paragon"

In all of the specimens examined the special development of lignified
tissues in the soft-bast occurred in the neighborhood of rather extensive
injuries to the cambium. These injuries of course killed the bark and
the cambium over a considerable area, and in addition stopped further
growth from the cambium for a short distance above and below this
area. This distance was never more than five centimeters, except
where an injected chemical, passing up just inside of the cambium,
caused a more extensive inhibition of growth. The general result did
not seem to vary with the cause of the wound. The simplest conditions
were those seen in trunks which were healing around a clean mechanical
wound (Plate LXXI, Fig. 3). The condition after an injection was
similar. Very weak solutions were ineffective. When the solutions
were toxic, however, the cambium and bark adjacent to the path of the
flaiid were injured, became in time segregated from the healthy bark by
a cork cambium, and finally dried up. Healing progressed as in the
first case, the abnormalities appearing above and below the wound as
before (Plate LXXI, Fig. 2). A more moderate action killed only the
cambium and the innermost part of the soft-bast, in which case the
abnormal tissues were formed in the bast external to the dead region.

In the region of inhibited cambial growth it was seen that the cells
between the wood and the inner side of the youngest hard-bast ring
eventually became lignified, largely forming wood fibers. Except in
the limited regions just described, the cambium continued normal growth.
Just outside the area of wound inhibition cell division was unusually
active, and there was tendency to form along the border of this area
a tissue growth similar to the callus that formed rapidly on the sides
of the wound. The effect of this activity would be to force aside the
bark over the inhibited area, unless in this area there was some means
of compensating for the lack of normal cambial growth. To a great
extent this is effected thru the formation of lignified patches in the soft-
bast.

The simplest explanation of the origin of the complicated tissue
masses produced can best be afforded thru a separate consideration
of the nature of the changes in each of the tissues involved.

NEW CELL FORMATIONS IN PLANTS 277

The Xylem or Wood was stained by the passage of the injection
fluid. If the cells adjoining them are not killed by the fluid, the pitted
vessels may be filled with tyloses, due in part at least, to the effect of
the injection (3). Some substances caused the wood to disintegrate,
killing the cambium and stimulating the formation of abnormal xylem
patches external to the dead region. The medullary rays in the wood
showed no particular reaction.

The Bundle-cambium as has already been mentioned, stopped growth
and matured into lignified elements, mainly fibers (Plate LXXIV,
Fig. 12).

The Phloem formed the greater width of the bark in trees of the age
of those examined, and was the part that showed the most pecuHar
developments.

There resulted an active formation of lignified tissues (a) from the
simple lignification of soft bast-cells, and (b) by the development from
soft-bast cells of meristematic areas that subsequently formed true xylem.
These meristematic areas, from their function may appropriately be
termed Xyleni-cambia. Since the formation of areas of simple lignifica-
tion frequently takes place subsequent to the formation of the xylem-
cambia, and in a definite relation to them, the discussion of the xylem-
cambia will be given before that of the simpler case.

The formation of the xylem-cambia in the soft-bast is a condition
of fundamental importance. It is interesting from the contrast it pre-
sents with Wistaria and other cases where a bundle-cambium originates
from the pericambium, as well as with the discontinuous bundles of some
of the Curvembryonae, such as the Chenopodiaceae and the Phytolac-
caceae, or Mucuna of the Leguminosae (8). In the present case the xylem-
cambium formed is not a bundle cambium, producing both phloem and
xylem, but forms xylem only, and eventually is completely matured into
xylem. Activity begins with the cells adjacent to the outer face of the
hard-bast strands only, and proceeds progressively outward. This
active zone is continually renewed on the outside by the activation of
more soft-bast cells, while on the inside the cells that have been dividing
mature into xylem. There was no evidence that the phloem ring of
one year was more sensitive than that of any other, except that the
outermost part of the phloem was never affected. This remarkable
activity may continue for some time, but finally ceases. There may
remain a considerable strip of unmodified soft-bast cells between the
newly formed xylem patch and the hard-bast strand next outside, or this
strip may become lignified as described below. In no case did the entire

278 TA\TOR— ON THE PRODUCTION OF

soft-bast between one hard-bast strand and the next outer become
completely re-formed to xylem. There seemed always to be the inter-
vening zone of lignified (or unHgnified) non-xylem elements. In large
and well-developed xylem-patches the distribution of elements resembles
that in the year-ring of wood, the majority of the pitted vessels being
inside, with the fibers on the outer face (Plate LXXIV, Fig. 12).

The elements of the xylem formed in this peculiar manner resemble
those of the normal xylem very closely, differing from them in certain
minor points only. They are somewhat smaller in size, and show the
peculiar pit markings more prominently (Plate LXXIII, Fig. 8). Fre-
quently they are twisted and deformed, and occasionally the pitted
vessels are completely bent over at right angles to the normal. The
pitted vessels also show the manner of their formation from several
cells fused together end to end. These peculiarities are seen along
the junction of two modified patches, where the wood forming normally
around to one side of the transforming region meets and joins with that
formed from the temporary xylem-cambia, and also in the first xylem
formed from the new bundle cambium in the manner described below.

Before touching upon this, it is necessary that the matter of simple
lignification be considered. This phase was very generally present.
The cells on the inner side of the hard-bast strands were those that
acted in this manner most constantly. Here the space between one
hard-bast strand and the xylem area already formed on the outer face
of the next inner hard-bast strand is often entirely filled by these lignified
soft-bast cells. Lignification does not proceed rapidly, and may be
completed in the deeper parts of the abnormal area only. Frequently
there is to be seen a mixture of Hgnified and of unlignified elements
(Plate LXXII, Fig. 6). The lignified cells are prominently pitted. Al-
tho the tannin-filled cells do not take any initiatory part in the transfor-
mation, they as a class become more completely hgnified than the ordinary
soft-bast cells. The circumferential lines of these cells, lignified, may
cross considerable areas of unchanged soft-bast. What has been said
of these applies equally to the medullary rays of the bark, which also
contain much tannin.

The soft-bast cells in the region of simple lignification show but
little tendency to change their form. They appear considerably larger
than the unchanged cells, but this is in part due to the shrinkage of
embedding, which affects only the soft, unchanged cells.

In one place where the soft bast lignifies there does appear a change
of form. The outward growth of the individual abnormal xylem areas

NEW CELL FORMATIONS IN PLANTS 279

would cause a rupture of the cells between them, unless these kept pace
with this growth. This occurs in part by cell multiplication, but mainly
by cell growth, and is followed by lignification.

Before passing to the closing phase of development the stone-cells
may be mentioned. They were found in various locations in normal
and in transforming bark, and they were seen in the patches of hgnifying
phloem, but not in the xylem patches forming in the phloem. They had
no evident relation to these. The hard-bast strands were much dislo-
cated, and sometimes were broken by the unequal growth of adjacent
xylem-patches, but otherwise they were unmodified.

In time there arises a new bundle-cambium from soft-bast cells
outside of the abnormal xylem areas. Unless there was something
of this sort formed, there would be no means for growth to continue
other than by the extension in number and size of the abnormal patches,
and there would be no phloem formed in these regions. Instead of
this happening, a meristematic zone arises on the outer face of hard-
bast strands and having essentially the same origin as the xylem-cambia
described above. This zone in time forms xylem and phloem (Plate
LXXIII, Fig. 8) of the type ordinarily seen on the sides of a wound.
Whether there originates first a xylem-cambium of the extraordinary
type, which later becomes a bundle-cambium, or whether there is phloem
formed from the first as well as xylem, the material at hand failed to show.
Following the establishment of this new bundle-cambium, natural growth
progressed as usual. There is a peculiarity of the hard-bast formed
by this cambium (at least during the first year or two) that is also charac-
teristic of that formed by the rapidly dividing cambium on the edges of
a wound. The abnormal strands of hard-bast are not flat and regularly
disposed, nor are they formed as in normal tissue, one ring each year,
they are larger and round in transverse section, and are somewhat scat-
tered (Plate LXXIV, Fig. 12). This, in the normal type of wound, is
shown in the second figure of Dr. Rumbold's paper (3).

The abnormal development of cork-cambia that was reported by
Dr. Rumbold seems, in all cases that the writer has examined, to be
about what might have been expected. These cork layers cause the
shelUng oflf of bark which is seen around all wounds, and eventually
over all parts of the trunk of the tree. They also occur isolated in the
bark, and here are probably due, not so much to the injection, as to
mechanical injuries. They may possibly be due to, and surround, Uttle
internal splits caused by strains set up by the deeper furrowing of the
bark. An interesting phenomenon seen in one case was the stimulation

280 ^ TAYLOR— ON THE PRODUCTION OF

to division of soft-bast cells and the causing of the cells so formed to
line up in a radial manner around the cork-bounded cavity. When
the injection caused the death of a strip of bark, this was segregated
by a cork-cambium that sometimes took a peculiar step-like course,
running between the hard-bast strands of the successive annual rings.

Since this work was begun as a continuation of that of Dr. Rumbold's
with a view to the amplification and confirmation of her findings, it
is necessary to compare the conclusions to which we have come.

The most important new point is the discovery of the same modifi-
cations in uninjected trees as she had found in injected ones. This
ehminates the necessity for explaining them on any such basis as that
propounded by her on page 491 of her paper (3). They are probably
a normal wound-reaction of the species.

On page 486, heading 5, she states that "The wounded tissue was
abnormal in that its position was reversed from the one in which it is
customarily seen. " So far as the writer could determine from the mater-
ial at hand the direction of xylem formation from the xylem-cambia was
the same as from ordinary cambia, that is, from the inner face of the
cambium.

The overgrowth wound- or callus-tissue that was formed at the
sides of the wounds was perfectly normal in position. In some cases
the ingrowing margin was so bent over that there is produced the appear-
ance in transverse section of a finger of wood largely surrounded by
bark, and new tissue being formed on the face turned toward the trunk
as well as outside. This, being merely a mechanical matter, can hardly
be interpreted as a physiological reversal of orientation. As for the
special xylem formed from the xylem-cambia in the soft-bast, this also
the writer cannot consider to be reversed in position. A reversed
relation of the xylem and phloem would imply that it was produced
on the outside and not on the inside of the formative meristem, inside
here meaning in the direction of the pith. Instead of this, it is quite
evident that the meristem was outside and that the elements were
produced on its inner face. A point emphasizing this is the relative
position of what appears to represent spring and autumn wood in these
abnormal patches, the "spring" wood being innermost.

On page 487 she states that " Quick killing was not followed by stimu-
lation other than the formation of normal wound tissue (callus) to cover
the wound. " The overgrowing side callus which she used in her Figure
2 to illustrate this could hardly be expected to show the abnormal areas.
The writer failed to find them here in all specimens that he examined.

NEW CELL F0R1\L\TI0XS IN PLANTS 281

In one or two cases isolated hard-bast strands were seen at the base
of the callus where it joined the wood, these probably having been left
behind after some little local disarrangement of the usual course of
events, such as irregularities in the form of the wound. The abnormal
areas might have been present close above or below the killed area and
have escaped detection.

It is to be observed that the term "bark" is here used to denote
all outside of the bundle-cambium. The hard-bast elements in the
abnormal areas would come in time to be surrounded by wood, and,
by the formation of the new bundle-cambium come to be cut off from
the bark of the tree. The term "bark cambium" as used on pages
488 and 489 corresponds to the term "cork-cambium" used here. Dr.
Rumbold on this latter page also states that the presence of abnormal
lumps of cork visible on the surface of the trunk denoted the presence
of a disturbance of the cambium and phloem tissues beneath, by which,
in connection with her Figure 6 referred to, the writer understands the
production of abnormal xylem patches. In the specimen indicated
to the writer by her in the field as a case in point, and brought in for
examination no such relationship was seen. There was merely a little
superficial patch of cork.

On page 489 she says "The inhibitory effect on the cambium was
transitory. In time a new cambium layer was formed, arising from
phloem cells." The effect on the original cambium was as permanent
as could be desired, since the cells became strongly lignified. It
was only the effect on cambial growth in general that was transitory,
tissue formation being taken up by the new cambium.

The last point which the writer wishes to note is Dr. Rumbold's use
of the term "xylem." Thruout her paper it includes all lignified tissues
except the hard-bast and the stone-cells. The writer would strongly
favor limiting the meaning to those tissues that are composed of fibers,
pitted vessels, etc., and which only seem to arise from a secondary meris-
tem: that is, wood in the ordinary sense. This would exclude those
areas of phloem modified by Ugnification without special change of cell
form.

Castanea — Summary

The work on Castanea consisted of a microscopical examination
of the tissues of trees injected with various chemicals, and of the tissues
forming in the neighborhood of mechanical wounds, with a view to
the determination of the exact genesis and nature of the abnormalities

282 TAYLOR— ON THE PRODUCTION OF

reported by Dr. Rumbold, their behavior during the second year of
growth, and the condition of the tissues prerequisite for their formation.

As a result of an extensive wound, the growth of the normal wood
cambium is stopped for a short distance above and below the wound.
In the region thus affected, there may form in the soft-bast of the bark
two principal types of lignified tissue not normally present there.

The simplest type consists of lignified soft-bast cells, the other is a
xylem resulting from the assumption by soft-bast cells of a meristematic
character.

In the latter case the first cells to become active are those on the
outer faces of the hard-bast strands, and the cells external to these
become involved successively.

These meristematic cells divide, and the products become matured
into the elements of a xylem which closely simulates normally formed
xylem.

A considerable part of the tissue between two or several rings of
hard-bast strands may be re-formed in this manner over a considerable
area, the patches of soft bast-cells that remain either being unaltered in
structure, or the walls become lignified without marked change in the
shape of the cells.

The cells between the adjacent strands of hard-bast of the same ring
come to take part in the changes, and in such regions the cells become
much distorted, and may appear intermediate in form between lignified
bast cells and xylem elements.

The medullary rays in the region of the formation of the xylem patches
grow with them and form rays similar to those in the normal xylem.

Eventually a fairly normal cambial region forming phloem as well
as xylem is developed from the outermost of these meristematic areas,
and, so far as the material indicated, further growth proceeded from
this in much the usual fashion.

The Injection of Herbaceous Plants

Material and Methods
As has been stated, this part of the work was instituted to obtain
evidence explanatory of the conditions found in Castanea. Naturally the
first method attempted was that used by Dr. Rumbold on Castanea (2).
The softness of the stem in the actively growing regions rendered difficult
the obtaining of a water-tight connection between the stem and the

NEW CELL FOR^L\TIONS IN PLANTS 283

tubes from the reservoir. This was partially accomplished thru the
use of bolts and wing-nuts in place of the springs used by Dr. Rumbold.
It only succeeded with fairly mature stems. In order to take advantage
of the greater activity of young stems, another method was devised.
A h>TDodermic syringe was used to introduce the fluid, generally inta
the pith cavity. After the solution was injected, the needle punctures
were covered with grafting wax. As an indicator of the path taken by
the injected fluid Anilin Black was used for some of the injections. For
the other chemicals used, see page 285. The plants used in the experi-
ments were Phytolacca decandra, Heliantlms anmms, Ricimis sp., and
Polygonum Sieboldii. The plants were allowed varying lengths of time
for such growths as might result from the injections to mature, and then
were cut down. After cutting, the specimens were brought in to the
laboratory and either sectioned fresh or preserved in alcohol for subse-
quent examination. Mounts of the Anihn Black injections were made
immediately, care being taken not to extract the dye. Most of the other
specimens were sectioned under alcohol, a few were embedded in Celloidin,
and still fewer in Parafiin. Safranin and Methyl Green was the com-
bination of dyes used on the unembedded material, while Safranin and
Delafield's Haematoxyhn served for the Celloidin and Paraffin prepa-
rations.

Summary of Experiments on Herbaceous Plants

The work on Phytolacca was begun early in July, 1916, and was
largely in the nature of preliminary tests to determine the best methods
to use on the Helianthus and Ricinus material, which was not yet ready.
The gravity-method injections (2) yielded Httle, but the hypodermic
needle gave some interesting results. The injections of Anilin Black
showed that the fluid passed up thru the protoxylem region. This is
what might have been expected, for the protoxylem region formed by
far the largest part of the conducting tissue at the time of the injection.
The only histological effect of this substance was to decrease the rapidity
of xylem lignification for a few centimeters up the stem and to reduce
the number of pitted vessels formed.

Picric Acid was the only other substance injected into Phytolacca, and
as a strong solution, in small quantities. If this solution had been in-
jected in large amounts into an internodal cavity as the weaker solutions
were in later experiments, the poisonous effect would have been extreme.
Internodal cavities were absent from stems of the size and age used.
The shoots were not as young as those used in the other plants. The

284 TAYLOR— ON THE PRODUCTION OF

effect of the treatment was to produce a local breaking down of the
tissues, pith proUferations, and a proHferation of the cortical tissues
to form a wound callus. The Picric Acid was not solely responsible
for the tissue growths here, since similar and nearly equal growths oc-
curred in the control experiments, where the stems were merely punc-
tured with a dry needle. By means of serial celloidin sections the re-
lation of the wound to the tissues produced was followed out quite care-
fully in this plant.

When these experiments were planned it was arranged that the largest
number should be done on Helianthus. In the late summer this plant
became so hard in the part near the ground that it was impossible to
obtain good, thin sections of it. Therefore it was possible to give much
of this material only a superficial examination.

Gravity injections with Anilin Black showed the same conditions as
in Phytolacca. In the older stems the pitted vessels as well as the pro-
toxylem served for the conduction of the fluid.

The main injections were made with Picric Acid, and in one of the
experiments this produced a very active pith-cambium (Plate LXXVI,
Fig. 18 and Plate LXXVII, Fig. 23).

The work on Ricinus was the most successful done during the summer
of 1916. This plant was found to be a favorable one by Erwin F. Smith
(6). The plants used were grown from seed listed by the dealer under the
names Ricinus borhoniensis arboreus, R. purpureus, and R. cambodgensis.
They were injected with Picric Acid both by the gravity and the needle
methods, the latter yielding highly interesting results. The solutions
in the latter case were injected into the internodal cavity and had their
greatest effect on the wall of the cavity. Some of the solutions did get
into the vessels, however, and caused decided changes in the tissues
surrounding them.

The most notable features of the sections are, first, the well developed
cambium formed in the pith, and second, the proUferations around the
vascular bundles. Hardly less interesting are the proliferations of the
pith cells into the internodal cavity. While preparing a set of sections
of an uninjected stem, proliferations similar in character to those into
the pith cavity of the injected specimens were found, altho less in extent.
They were not accompanied by any evident wound to the stem, and possi-
bly show that artificial stimulation is not requisite for their production.

A feature that was very prominent around the wound was the pro-
duction of cork. This was greater in amount than that normally pro-
duced at the base of the plant. The formation of the cork occurred

NEW CELL FORMATIONS IN PLANTS 285

from a cambium that generally originated from the second layer of corti-
cal cells inside of the epidermis. It extended for two or more centimeters
on all sides of the wound.

The work on Polygonum Sieboldii was started at the suggestion of
Dr. Macfarlane in order that advantage might be taken of its rapid
growth by forcing in the greenhouses, and that injections might be made
at a time of year when the plants outside were still dormant. About
thirty experiments were so conducted, from January 1917 on. The
first experiments outdoors were made on the 16th of April 1917, and of
these fifty were eventually set up, as well as twenty controls.

Stages have been obtained showing the formation of the cambium
developed from the pith cells, with the laying down of xylem by this
cambium. The points shown in the other experiments:— cell-outgrowths
into the pith cavity, local proliferations of cortical cells, etc., have been
duplicated. A new feature appearing here for the first time in the
writer's experiments is the formation of xylem from the inner cortex,
which has already been reported by Schilberszky to follow his operations
on Phaseolus. This in the present experiments only appeared when
the treatment had been unusually severe.

The work on Helianthiis, Phytolacca and Ricinus was limited to the
injection of Anilin Black and Picric Acid. The choice of Picric Acid
was made because it was thought to have been especially effective in Dr.
Rumbold's injections. The chemicals used on Polygonum were:— Anilin
Black, Picric Acid, Ammonium Hydrate, Lithium Carbonate, Copper
Sulphate, Sodium Carbonate, Potassium Chlorate. Ammonium Car-
bonate, Chloroform and Urea, all in aqueous solutions and in most
cases in two or more dilutions. All injections were made with the
hypodermic syringe into the internodal cavity.

The action of the different chemicals differed only in rapidity and
extent. The writer could trace no specificity in action, beyond the fact
that some failed to produce the internal pith proliferations. There is
here given a summary of the reactions in some of the more important
groups of Polygonum injections : —

Distilled Water. Both simple proHferations into the cavity and
a pith cambium were present. No formation of xylem from the pith
cambium was evident.

Chloroform Water. There was here an advanced formation of
the pith cambium following the injection of water saturated with Chloro-
form.

286 TAYLOR— ON THE PRODUCTION OF

Ammonia. All stages in the formation of the pith cambium were
found, including, in the neighborhood of the wound, the formation of
xylem. Pith proliferations into the cavity also took place.

Lithium Carbonate. The stages in the formation of the pith
cambium, including the formation of xylem, were seen as a result of the
injections made with this substance. Superficial proliferations also
occurred.

Copper Sulphate. The pith cavity was much browned in this
series and the formation of a pith cambium occurred.

Picric Acid. The formation of the pith cambium was active in
some of this series. When the injection was so severe as to kill a part
of the internode, extrafascicular cambia and bundles were produced.

The Effect of the Injection on the General Growth of the Plant

In order to fully appreciate the importance of the histological modi-
fications produced by the injections and by the attendant wounds,
it is necessary to consider the effect of these on the general health of the
plant.

In general, unless the toxicity was quite high, there was no immediate
effect visible on the surface of the stem. When this was the case however,
the internode treated became flaccid. Later this effect extended up
and down the stem and death generally followed. If a violently corro-
sive substance had been injected the effect would have been even more
rapid.

When the substance was only strong enough to have a locally toxic
effect, the first evidence of this effect was a browning around the injec-
tion holes. This gradually spread up and down the stem. This, in
time, might involve three or four internodes.

The age of the shoot and its vigor at the time of the injection had
considerable to do with the effect of the injection. The more vigorous
and active the shoot the greater the spread of the injected fluid thru
the vascular system, and also the greater the response of the pith cells.
A stem that had not a vigorous habit tended to give a local reaction.

Besides these local effects, due directly to the toxicity of the injected
substance, there was an effect on the growth of the injected internode
that occurred even tho the solution was so weak that tissue destruction
did not take place.

This was first suspected when it was found that the injected internode
was shorter than those above and below it. Measurements were made
to determine to what degree this took place. The degree of variation

NEW CELL FORIVLVTIONS IN PLANTS 287

in the rapidity of elongation of the internodes depends on so many
factors, such as weather, soil moisture, age and position on the shoot,
that it was not possible to arrange series that did not show a great amount
of variation. This indicates that great care should be exercised in
the interpretation of the data. All facts considered, however, the general
conclusions that have been reached are as follows:

(1) The length of the injected internode normally would be inter-
mediate between that of the one above and the one below, but the differ-
ences between them would be slight. This does not hold if the parts
measured were too near the ground, for the internodes near the ground are
much shorter than those a little higher up.

(2) After the injection there followed a short period of normal length-
ening, subsequent to which the injected internode slackened growth.

(3) The internodes above and below the injected one continued
to lengthen normally or nearly so. A point must here be noted that
cannot be determined until the plant has been cut down and dissected.
Sometimes, either thru the action of the injected substance or because
of a deficiency in the internodal septum, the fluid finds its way into
the internode below that injected, affecting that as well as the one
originally treated.

(4) The stunting of the injected internode was permanent. No
reassumption of activity tending to increase of length appeared. The
diameter increase is about as normal.

On examination of the graphs one notices that the internode above
the one injected lengthens rapidly, until it is of the long^t of the three
considered. That this is due to the effect of the injection is doubtful.
The position on the stem was probably the most potent factor in causing
this increase. The injected plants were not rendered sickly by the treat-
ment; they were quite as healthy as the controls. This is only true
of those where the tissue destruction was not great. One measurement
was made of the plants about two weeks after the close of the regular
measurements, and it showed an increase in length of the upper internode
only, and that of but two to five millimeters. The point of importance
in the graphs is the dwarfing of the injected internode. That there was
some slight effect on the growth of the adjacent internodes was very prob-
able, but the data are insufficent to demonstrate it.

Besides the cases used to illustrate these points, a number of addition-
al observations were made, all on Polygonum.

288

TA\T.OR— ON THE PRODUCTION OF

[ength ef

Internoae

in

millnetert

/ GRRPH I- nCEPLt PUnCTURE

n- PISTILLEP WRTER
Hr-UTHlUn CRRBOriRTE

90

eo

70
60
£0
40
30
20
10
0

0--0

I

-=^rZ

'CSGSJfI>-

Tnternoda bolov Injection Q
Xntarnede Injaofd. 0~~

Intsrnoda above Injection Q- — — — — ^

The first date given t» that
upon which th9 Injeotton was luidtf.

Mea$itrement» made lote in i/te
$uaKtr afiou no jiroittfi ooer that
ahott/n tn thgse graphs.

Discussion of the Graphs — Polygonum
Needle Puncture Controls
In addition to untreated plants, a series was run on plants that had
been punctured as if making an injection, and the wounds similarly
protected with grafting wax. It will be seen that the punctured internode
is longer than the internode below, as in the normal stem. Compare
with the injected specimens. Specimens 57, 58, 59, 60, 61.

Distilled Water Injections
The injected internode is seen to be the shortest of all. The plants
seem to have averaged somewhat older than the preceding. Specimens
62, 63, 64, 65, 66.

NEW CELL FOR^LATIONS IN PLANTS 289

Lithium Carbonate Injections. 1 :10 Saturated Aqueous Solution

This series of specimens shows a very prominent stunting of the in-
jected internode. Specimens 33, 34, 35.

Tissue Reactions Following the Injections of Herbaceous Plants

When the injection was made, two holes were opened in the stem.
The liquid, injected thru one hole, ran down the side of the cavity till
this was full up to the hole left for the escape of the air. Then the
excess fluid escaped thru this hole. The needle was withdrawn, the
outside of the stem wiped dry, and a piece of gauze bandage smeared
with melted grafting wax was wrapped around the wound. Such
were the conditions when the fluid began to act on the stem. Some-
times less fluid was injected, and some of the other arrangements were
different in the earlier experiments, but the general conditions were
the same thruout.

It will be seen that there were two main regions upon which the sub-
stance could act:— the surface of the cavity and the tissues surrounding
the vascular bundles which were cut by the introduction of the needle.
That these were the only two practicable points of attack was fairly
well indicated by the fact that only in these two places, the cavity of
the internode and the neighborhood of the vascular bundles, was the
Anilin Black seen after injection. To this there is an exception. In
the neighborhood of the wound there was a considerable amount of the
dye that seeped into the intercellular spaces. This seepage had no
effect on the major part of the tissues. The amount of fluid that got
into the vessels was much less in the hA-podermic needle injections than
in those where the gravity method was used, for in this latter case the
number of vessels cut open was far larger.

The Epidermis. It is now necessary to turn to a detailed considera-
tion of the tissue modifications produced by the injections. So far as
the material went, there was no evidence of any tissue production from
the epidermis, even tho in the young specimens injected the development
of a cuticle could not have advanced very far.

The Cortex. The cortex usually only shows the effect of the injection
near the wound. This may either be due to the mechanical irritation
or the percolated injection fluid. The extrafascicular bundles will
be considered later.

In spite of the fact that the outer cortex was a well developed collen-
chyma, it proliferated freely in Phytolacca, Polygonum and Ricinus.
It proliferated in Helianthus also. The product was sometimes an

290 TAYLOR-ON THE PRODUCTION OF

irregular mass of cells, while on the other hand there was sometimes
formed a distinct cambium. This cambium often split oflf cells like
the most typical cork-cambium (Plate LXXVIII, Fig. 28). This is
a different condition from that so often seen, where the outer layer of
the cortex gives rise to the cork cambium. It was only found around
wounds of injection or fissures in the epidermis resulting from mechanical
injuries.

The inner cortex, which was parenchymatous, showed in all species
a ready ability to prohferate. The causes which operated on the outer
cortex were effective here also. The cork-like formation was present as
in the outer cortex, both often taking part in the same reaction. It is
improbable that much tissue could be laid down from the proliferation of
these cortical cells.

The scleroid patches, when present, took no part in the reaction.
No new scleroid tissue was produced.

The most prominent and interesting tissue production was from the
innermost cortical layers. This formed the extra-fascicular bundles
of the type mentioned by Schilberszky (4). When some great disturb-
ance of the function of the stem occurred, such as the killing of a large
part of the internode by the toxic action of the injection, the stem re-
sponded by the formation of extra-fascicular bundles. When the pith
was killed, (and possibly also the xylem affected,) these bundles were
often produced. In the former case as active a reaction would probably
have been produced by the simple operation of removing a part of the
internode, thus exposing the interior to drying. This is practically
what Schilberszky did. The xylem produced was quite abnormal in
appearance, the small pitted vessels being much distorted and displaced
from their normal orientation. They became well lignified, however,
and were quite unmistakable. These were produced in Polygonum
alone, for only there were the solutions sufficiently strong to cause the
necessary tissue destruction. The production of these bundles was
limited to the neighborhood of the wounded region (Plate LXXVIII
Figs. 25, 26, 27).

Schilberszky (4) describes the formation of both phloem and xylem.
In the writer's preparations there was no well-defined phloem shown.
Outside the xylem, and dipping into it in places, there was a soft tissue
which took the Methyl Green stain readily, but it was not possible to
difi'erentiate it into a meristematic and a matured layer.

The Phloem is the next tissue to be considered. Generally it showed
no response to the injections, but in at least one case of Polygonum

NEW CELL FORMATIONS IN PLANTS 291

where the injury had been great, there appeared a pecuhar proHferation
of the vascular cells. The cells that responded were apparently all
from the young, unlignified cells of the phloem, cambium and probably
the xylem. In this case there was a partial resumption of the formation
of lignified tissue, with the production of a few pitted vessels. Unfor-
tunately the specimen was cut too soon for much growth to have oc-
curred, and the extent to which the cambium might have regenerated
and resumed its normal function was not shown. (Plate LXXV, Fig. 16).

The Cambium was often retarded in its growth by the injection,
this being especially noticeable when the fluid had passed up or down
the vascular bundle. Except that they were possibly of a smaller
size, the elements produced were not abnormal. See also the condition
described under Phloem.

The Xylem was generally little affected by the injection, this being
especially the case if the elements had become the least lignified. If
the fluid passed up the protoxylem while the secondary elements were
still soft, they took part in the general proliferation around the bundle.
The condition is well shown in Plate LXXVIII, Fig. 29. This case
was an extreme one. The cambium continued growth, forming a con-
siderable amount of secondary wood. If secondary wood had been
formed at the time of the injection, there would have been fibers and
pitted vessels visible in the mass of dead xylem included in the center
of the proHferated area, instead of spiral protoxylem elements only.
It is to be noted that the cells which were directly around the vessels
were killed by the toxic action of the fluid, and that the cells further
back proUferated. This suggests strongly that the effect of the chemical
was an indirect one, and that if by some mechanical means a mass of
bundle and surrounding pith cells could have been killed the same result
would have been produced.

The Pith was the region of the stem that gave the most interesting
results in the way of tissue formation. This was in those specimens
that had been injected by the needle method.

If the fluid was weak, the inner pith cells bordering on the cavity
were not killed. If any reaction at all took place as the result of such
an injection, it was a simple budding of the cells into the pith cavity.
This condition has already been reported, and has been produced by
the use of gases (6). The condition of affairs is seen in Plate LXXVII,
Fig. 21.

If, now, the strength of the solution was somewhat greater, there
was formed a layer of dead cells Uning the cavity. From behind this,

292 TAYLOR— ON THE PRODUCTION OF

the underlying cells may bud out into the pith cavity. That the growth
is not limited to the elongation of one or of a few adjacent cells is shown
by Plate LXXVII, Fig. 24.

Coincident with the condition just described, there was another and
more important development. This was not generally preceded by any
proliferation of the pith cells into the cavity, it being strongest when the
inner cells, bordering on the cavity, had been killed by the injection.
It consisted of the formation of a cambioid zone from pith cells. This
has been already described for Ricinus, but not for the other species
worked upon by the writer. The procedure was as follows : —

The cells underlying the dead area first elongate radially and tend
in so doing to decrease the size of the internal cavity. Then they divide,
and the resulting division walls are fairly transverse to the direction of
elongation (Plate LXXV, Figs. 13, 14, 15).

Next, the cells in a portion of this prohferating region divide more
rapidly than the others, thus forming a meristematic zone which takes
the form of a sheath often wholly surrounding the part of the cavity
touched by the injected fluid. The nuclei in this meristem and in the
cells of the proliferated region are often unusually prominent. The
protoplasm is not dense, as in an embryonic layer in a growing point,
but is scant, as in the cambia of ordinary stems.

Sometimes the cells just inside of this meristem have the appearance
of having been cut off from it. As they are often yellowish, and stain
readily with Safranin, it might be thought that they were cork, but
repeated applications of tests for suberized membranes have failed to
demonstrate its presence.

An interesting fact in this connection was found in Polygonum. In
some of the untreated plants there had appeared deep slits or furrows
in the pith. About the bases of these sHts, near the protoxylem, the
pith cells showed a tendency to divide, and the areas thus formed re-
sembled the early stages in the formation of the pith-cambium.

The question now arises as to what is the true nature of the meri-
stematic zone. Is it a true cambium, forming both phloem and xylem,
or does it only form xylem? There is first formed a belt of cells that
are like xylem parenchyma, there being gradually added to these in the
later formed zone, pitted vessels. The formation of this zone seems
to start near the insertion of the needle, and progress around the stem
from there. The form of the elements is not as regular as in the normal
wood, nor are they as large, but they are undoubtedly xylem elements.
However, cells with thickened pitted walls were seen in some of the

NEW CELL FORMATIONS IN PLANTS 293

sections, which must be carefully distinguished from the abnormally
formed xylem. They were formed from the pith cells by simple lignifica-
tion, this occurrence, tho not normal to the species used, being nothing
unusual. It was most apparent in a plant of Polygonum that had had
a large part of the internode exposed to the air by the corroding action
of Picric Acid introduced in the solid form.

The formation of phloem was not as clearly shown as was the forma-
tion of xylem. The cells inside of the cambium region were all so uni-
form in appearance (the cells formed by the original segmentation of
the pith cells are to be carefully distinguished here) that there seems
to be little if any true phloem present in the sections examined. It
may be that a httle longer time for maturation would have shown that
some of the questionable regions were really phloem. The formation
of phloem in this manner has already been reported for Ricinus, (6),
by Dr. Erwin F. Smith, who also reports the formation of "concentric
medullary bundles" in the pith. The writer is disinclined to assert
that this pith-cambium is a true bundle cambium on the basis of his
own preparations.

Herbaceous Plants — Stimmary

In order to obtain material that would serve to aid in the interpre-
tation of the conditions present near the wounds in the trunks of
Castanea, experiments were started on herbaceous plants.

These mainly consisted of the injection into the pith cavity by means
of a hypodermic syringe, of various substances in aqueous solution.
The injections served to stunt the growth of the injected internode.

Modifications in the tissues of the plant were produced both as a direct
action of the chemical agent as a stimulating agent, and indirectly as
a result of its toxic action on the cells.

The reactions that can be considered as primarily due to the first
cause consisted mainly of prohferations into the pith cavity of the cells
bordering on it.

Those due to the poisonous action of the injected fluid cannot always
be clearly distinguished from these, nor can they always be distinguished
from wound tissue reactions caused by the insertion of the injection
needle.

As a result of the action of one or more of these agencies proliferations
of all the tissues except the epidermis and the lignified elements were
produced.

294 TAYLOR— ON THE PRODUCTION OF

These proliferations were frequently irregular, and then from them
no differentiated tissue arose.

At times under favorable conditions there was formed in these pro-
liferated regions a cambioid zone, that laid down tissue in a manner
resembling that of normal cambia.

From zones of this kind that were formed from cortical cells, cork-
like protective tissues were produced.

As a result of severe injury to the stem, the innermost cortex pro-
duced xylem patches that seemed to correspond to the " extra-fascicular
bundles" of Schilberszky.

The only proHferation of phloem cells that was produced was an
irregular one and concerned also the cambium and probably the young
xylem cells. There was a suggestion of the resumption of cambial
activity in this region.

The xylem, besides being concerned in the above reaction, was able,
when it had not been lignified, to proliferate.

The pith responded readily while young, and there was formed
after proliferation a cambium that laid down xylem, and possibly also
phloem, around the pith cavity.

Conclusions

As a result of these experiments, the writer would consider that all
the elements of the normal stem are capable of extensive multiplication
unless they have been modified by cuticularization, lignification, or
suberization. Cells that are collenchymatous are, notwithstanding,
able to prohferate freely. From these proliferated areas there may be
formed cambioid zones that give rise to cork, to xylem, and possibly,
to phloem. The initial multiplication may be started by mechanical
or chemical means. The chemical irritation, if it suffices to cause tissue
destruction, may have an ultimate effect similar to a mechanical irrita-
tion.

BIBLIOGRAPHY
This list contains only those works found most helpful to the writer. Where an
author has been mentioned in the text, but is not referred to in the bibliography, it
is to be understood that a synopsis of his work is contained in Kiisters "Pfianzenana-
tomie."

(1) Kuster, Ernst "Pathologische Pfianzenanatomie" Jena, Gustav Fischer.

Zweite Auflage, 1916.

NEW CELL FORMATIONS IN PLANTS 295

(2) Rumbold, Caroline "Methods of Injecting Trees." Phytopathology 1915-

V, 225-229.

(3) Rumbold, Caroline "Pathological Anatomy of the Injected Trunks of Chest-

nut Trees." Proc. Amer. Phil. Soc, 1916: LV, 485-493.

(4) Schilberszky, Karl "Kunstlich hervorgerufene Bildung secundarer (e.xtra-

faszikularer) Gefassbiindel bei Dicotyledonen " Ber. d.
D.botGes., 1892: X, 424.

(5) Simon, S. "Experimentelle Untersuchenen uber Differenzierungs-

vorgange im Callusgewebe vom Holzgewachsen. " Jahrb.
f. wissen. Bot., 1908: XLV, 351.

(6) Smith, Erwin F. "Mechanism of Tumor Growth in Crown-Gall." Jour.

Agr. Research, 1917: VIII, 165-186.

(7) Solereder, Hans "Systematic Anatomy of the Dicotyledons." English

Translation, Revision of D. H. Scott, Clarendon Press,
Oxford, 1908.

(8) Strasburger, Edward "Textbook of Botany. " 4th English Edition, Macmillan,

London, 1912.

(9) deVries, H. "Uber Wundholz." Flora, 1896: LIX, S. 2, 17, 38,49,

81, 97, 113, 129.

EXPLANATION OF PLATES

PLATE LXXI— TYPICAL CASTANEA TRUNKS FROM AMONG THOSE

EXAMINED

Fig. 1. The hole made for the injection may be seen at "a." The effect of the in-
jection was barely visible as far up as "b." The saw cuts for the pre-
liminary examination may be readily seen. Injection 32.

Fig. 2. Injected in 1914 with Lithium Hydrate 1 :100 Gram-Molecular concentra-
tion. This killed the bark for a considerable distance above and
below the points of injection 'V"-"c." Xylem-cambial activity was
visible as far as the top of the part of the specimen shown, and material
for histological examination was taken from "d." It is shown in Plate
LXXIII, Fig. 8. Injection 20.

Fig. 3. From a part of the specimen not reached by the injection. There is shown
here a healing wound resulting from the removal of a blight infection.
Xylem formation from xylem-cambia had occurred, showing well in material
from "e." Injection 21.

Fig. 4. A similar case. The normal callus is shown at "/." The included hard-
bast strands were to be seen to the top of the part of the specimen figured.
Injection 31.

PLATE LXXII— THE PHLOEM, NORMAL AND MODIFIED— C^^jT^A^^^

Fig. 5. Normal Phloem— Transverse Section. The hard-bast of two rings is shown
at " a. " The dark row of cells that surrounds each of the hard-bast strands
is the layer of crystal cells. The soft-bast, "b," fills all of the space be-

296 TAYLOR— ON THE PRODUCTION OF

tween the hard-bast strands. The medullary rays are marked "c," the
tannin-containing soft-bast cells "d " Three of the rows that these form
are shown, the lowest only being marked. X 50.

Fig. 6. Injected in 1916 with Sodium Hydroxide, 1:1000 G.M. This shows, in
transverse section, phloem in the process of lignification. The growth in
this specimen was not sufficiently active to produce pitted vessels. In
this figure "j" indicates a medullary ray. A narrow hard-bast strand is
marked "j"-"j"; the cells did not stain heavily and therefore do not show
clearly. They may be distinguished by the dark central spot caused by
the refractivity of the wall. The lignified phloem cells, "/, " may be
distinguished by their thick walls, the unmodified cell walls appearing
dark, but much thinner. These are shown at "m." Three of the rows
of tannin cells may be seen, the uppermost passing thru the uppermost
"i." Injection 28. X 180.

Fig. 7. Lignified Phloem, Longitudinal Section. This is a tangential section, show-
ing at "o" a medullary ray in end view. Injection 30. X 130.

PLATE LXXIII— XYLEM FORMED FROM THE BUNDLE-GAMBIA—

CASTANEA

Fig. 8. Longitudinal Section of Xylem from the New or Re-formed Bundle Cambium.
The hard-bast strand, outside which the xylem was formed is seen at "h."
The wood fibers appear at "d." A medullary ray has been marked "e. "
At "/" is shown one of the pitted vessels, the wall of another shows at " g."
Injection 20. X 50.

Fig. 9. The Bordered Pits of the Normal Wood Fibers. Longitudinal Section.
The field shown here was an unusually favorable one for the demonstration
of these pits. The round area about the pore is shouTi. X 610.

Fig. 10. The Structure of an Individual Xylem Patch. At "a" is located a hard-
bast strand. The direction of the pith is toward the bottom of the sheet.
It will be easily seen that the xylem is continuous outward from the hard-
bast, and that the pitted vessels are mostly in the earlier formed wood, as
in the normal stem. The darker area "b" outside of the .\>'lem patch
is the region of lignified phloem cells. The lignification here is com-
pleted, no cambial zone remaining between the xylem and the lignified
phloem. Injection 28. X 75.

PLATE LXXIV— XYLEM PRODUCTION ORIGINATING IN PHLOEM

TISSUE— C^ 5r.l NBA

Fig. 11. Appearance of the Modified Regions when viewed as Opaque Objects.
"WTien dealing with a wood as hard as that of Castanea, it is advantageous
to be able to recognize the important features without resorting to the
preparation of microtome sections. This photograph shows part of a large
area as it would be seen if prepared as described on page 274 and viewed
with a high-power hand lens. It is difficult to show the hard-bast strands

NEW CELL FOR^LVTIONS IN PLANTS 297

in such a photograph as this, they are too similar in color to the surrounding
soft-bast. In the photograph "c" is just outside the region of the re-
formed cambium. A little below and to the right of "c" there is a lighter
patch separated from the .\ylem by a dark line, and other smaller patches
are to be seen in similar positions. These are hard-bast patches corres-
ponding to "g" of Plate LXXIV, Fig. 12. The included hard-bast strands
are marked "d" and "e." The group of hard-bast strands of which "e"
is a part is not of the same appearance as the more normal shaped ones
{"d") farther out. This irregularity was caused by an injury to the
cambium prececding the removal of the blight infection, which subsequent
operation in turn caused the formation of the .\ylcm cambia. From "/"
downward is wood formed from the old cambium before its growth was
stopped. X 11.

Fig. 12. General View of a Transformed Area. It is verj- difficult to show the complex
details of the transformed areas in a low-power photograph. In the illus-
tration all above "//" is normal phloem. Just below " li" there is a
little local disturbance of the growth. The reformed bundle-cambium
stretches across the picture between the phloem and the somewhat darker
x>'lem. Just outside of it are located ("g") the hard-bast patches formed
from this cambium. At "j" is the row of hard-bast strands outside of
which the new bundle-cambium started activity. At "i" is the only
hard-bast ring which had formed outside of it the temporary xylem-cambia.
X 14.

KEY TO THE LETTERING OF THE FIGURES OF THE HERBACEOUS

PLANTS

(Other signs than these are explained in the individual descriptions.)

"cp" ParenchjTna of Cortex "mx" Metaxylem of Wood

"hb" Hard-bast of Phloem "/>" Pith

PLATE LXXV— THE PITH CA:MBIUM IN POLYGONUM

The first two of these figures hardly need any explanation: they show the pith
region just bordering on the ca\aty, and the increase in the number of the cell walls
laid down in the formation of the pith-cambium. The description of this process
is given on page 292.

Fig. 13. First Stages in the Formation of the Pith Cambium. From Injection No.
3, where Picric Acid was introduced in solid form into the treated internode.
X 55.

Fig. 14. Well Formed Pith Cambium. From in jection No. 1, Distilled Water. X 55.
Fig. 15. This is a general \aew of one of the pith regions and part of the bundle ring

in Injection No. 1, and shows the point of entrance of the injection needle.

The letters "f"^"f" mark the position of the pith cambium. X 9.

298 TAYLOR— ON THE PRODUCTION OF

Fig. 16. Proliferation in the Phloem-Cambium Region, Polygonum. From Injection
No. 6, where Calcium Nitrate in solid form was introduced into the treated
intemode. This figure and the next show two phases of the same action.
In this one there are two lines of activity, one along "b"-"b," the other
more external at "d"-"d." The cells in between were not in an active
condition at the time the plant was cut down. At "/" there is a group of
wood elements, probably pitted vessels, that have most likely been formed
from the iimermost cambial zone. X 55.

Fig. 17. Xylem Formation from the Pith Cambium. Because of excessively deep
staining the details of this preparation were somewhat obscured. At
"a" is seen a large pitted vessel. The axis of the vessel is at right angles
to the axis of the stem. This is often the case near the wound. The
active zone was located at about "b"-'b." The pith cavity is above "d,"
where the innermost cells are dead and stained dark. The more normal
pith is seen to the left. From Injection No. 12, where 1% NH4 OH was
used. X 55.

PLATE LXXVI— THE MAXIMUM DEVELOPMENT OF THE PITH

CAMBIUM

Fig. 18. Helianthiis. This specimen was taken from Injection No. 7, treated with
1 :1000 G.M. Picric Acid. It shows the point of entrance of the needle.
There was no pith cavity formed. There was a considerable reaction here,
and the pitted vessels formed from the pith-cambium generally had their
axes parallel with that of the stem. The large cell shown at "a" is not a
pitted vessel with a horizontal axis, but one of the pith cells that, though it
elongated, did not divide. See also Plate LXXVII, Fig. 23. X 11.

Fig. 19. Ricinus. From Injection No. 1, injected with Picric Acid, 1 :100 G.M. Here,
as in Helianthus, the intemode was nearly soHd at the time of the injection,
and the little cavity that originally existed was filled by the multipHcation
of the pith cells. The xylem shows as a dark band, and is marked "b."
The split caused by the needle is marked "(i." X 11.

PLATE LXXVII— PROLIFERATIONS INTO THE PITH CAVITY AND
PHLOEM-CAMBIUM PROLIFERATION

Fig. 20. Ricinus Pith Proliferations into the Cavity. From Injection No. la, treated
with 1:100 G.M. Picric Acid. Note the tuberculate appearance of the
proliferated cells. The dark masses ("a") at the bases of some of the
enlarged cells are dead cells. X 55.

Fig. 21. Polygonum Pith Proliferations into the Pith Cavity. From Injection No. 9,
treated with 1:50 G.M. Picric Acid. This specimen shows a moderate
development of the pith cells. The ends of the cells do not show the
tuberculate appearance that was developed in Ricinus. Below the en-
larged cellg there may be seen the beginning of a pith cambium. X 55.

NEW CELL FORMATIONS IN PLANTS 299

Fig. 22. This shows a type of proliferation in Ricinus that involves young meta-
xj'lem as well as pith. Below "w.v" is showTi metaxylem that has not
been affected. Compare with Plate LXXVIII Fig. 29. X 45.

Fig. 23. The Xylem Formed from the Pith Cambium in Ilclianlhus. From Injec-
tion No. 7, treated with 1:100 Picric Acid. The character of the pitted
vessels formed from the pith-cambium is hard to show. The wall of a
pitted vessel is shown at "^." X 85.

Fig. 24. Proliferation of a Large ^lass of Cells in Polygonum. From Injection No.
19, treated with NIL OH 10% Solution. This shows that not only can the
superficial cells proliferate singly or in small groups, but large masses of
deep-lying cells can act together and form nodules on the sides of the
cavity. X 17.

PLATE LXXVni— THE FORMATION OF THE EXTRAFASCICULAR BUN-
DLES IN POLYGONUM— ALSO PROLIFERATIONS AROUND
THE PROTOXYLEM AND CORK FOR-
MATION IN RICINUS

Fig. 25. The First Stage in the Formation of the Extrafascicular Bundle. From
Injection No. 23, see above. There are seen here the first cell walls laid
down in the formation of an extra-fascicular bundle. Note especially
"a." X55.

Fig. 26. A Well Developed Extrafascicular Cambium, Polygonum. From Injection
No. 23, see above. The cambium shown here had just begun to lay down
wood when the specimen was cut. There is seen a pitted vessel at

"b." X55.

Fig. 27. Extra-fascicular Bundle in Polygonum. This shows one of the most advanced
cases that these experiments have produced. A large group of pitted
vessels is shown at "d." X 70.

Fig. 28. Cork Formation in Ricinus. From Injection lb, for data see Injection la,
above. This photograph shows the manner of cork formation in Ricinus.
The interest attaches to the manner in which the different layers of the
collenchymatous outer cortex take part in the process. The needle hole
was above the top of the picture. It is easy to see the individual layers of
cells, as "a," "b," "d," along the lower margin of the field, and to trace
each row upward and determine what part it had in the formation of the
cork-cambium. The position of this cambium is along "/"-"/. " It will
be noted that the nearer the region under observation is to the puncture, the
deeper are the layers that take part in the cork formation. X 45.

Fig. 29. An Advanced Case of Proliferation in Ricinus. From Injection 2c, treated
with Picric Acid 1:100 G.M. The tissue destruction here was much more
considerable than in the preceeding case. Considerable numbers of
dark conglomerate crystals are seen that are located in a tissue that must
have come from the cambium, either directly or by proliferation of cells
previously formed from it. The cambium had resumed normal growth,
and the normal metaxylem is seen near the lower part of the picture. X 20.

Bot. Coiiirib. Univ. Fain.

Vol. IV, Plate LXXI

I— t

H
<

o

Bot. Conlrib. Univ. Pcnn.

Vol. IV, Plate LXXII

Vic. 6

.M. -^^ V

^^f-i

t,

l4

fe^ «»^- .--»... ^„- .- ^».-

'v"'£S'': '-^ ?^^ ' --Sill

Fig. 5

Ftg. 7

TAYLOR ON CELL FORMATIONS

Bot. Coutril. L'niv. Pom.

*V^ ^Jkj

' e

IW. /r, P/a/t' LXXIIl

Fig. 8

1... 9

FiG. 10
TAYLOR ON CELL FORMATIONS

Hot. Conlril). L'liiv. Pom.

Vol. I\\ Plate I.XXIV

Fig. 11

f,' '^^\

'v*r. 'n't'' ■*

'^'"' "■ ij\

¥..

i1

Fig. 12
TAYLOR ON CELL FORMATIONS

'\

Bot. Contrib. I'niv. Pom.

Vol. I\\ Plate LXXV

Fig. 15

+^•3 ..

i^fi;;^^

"^''3^

l-IG. 17

TAYLOR ON CELL FORMATIONS

Bot. Contrib. Univ. Pom.

Vol. TV, PI ale LXXVI

Fic. 18

Fig. 19
TAYLOR ON CELL FORMATIONS

Bol. Conlrib. ink'. Pom.

Vol. I\\ Phi'.c LXXVII

Fig. 20

Fig. 21

c-.4«rT~^

Fig. 22 F^g. 24

TAYLOR ON CELL FORMATIONS

Bot. Contrib. L'niv. Penti.

Vol. n\ Plate LXXVJII

Fig. 25

Fig. 26

/.-'.•

^'^Sj.

\. :

""^m^^^^^

^

^jiif^.

y>:

r^

.^•

Fig. 27

Fig. 28 Fig. 29

TAYLOR ON CELL FORMATIONS

THE PEANUT {Arachis hypogaea) — ITS HISTORY,
HISTOLOGY, PHYSIOLOGY, AND UTILITY

BY

Ralph Augustus Waldron, B.S., M.S.
With Plates LXXIX and LXXX

Thesis Presented to the Faculty of the Graduate School of the University of Penn-
sylvania, May 1918, in partial fulfilment of the requirements
for the degree of Doctor of Philosophy.

CONTENTS

Introduction •^

History -^

Introduction ■^

Early American Records ^°^

Lack of Evidence as to its extra-American Origin 3°S

Distribution • ^

Recent Literature -^

General Morphology ^^°

Histology -^ -^

Root ^^^

Stem ^^^

Leaf ^\^

Fruit -^^^

Physiology ^

Root Hairs ^

Observations and Experiments with Root Hairs 3^^

Absorption ■^ ^

Development -^ ■'

Biological Considerations Concerning the Fruit and Gynophore 3^7

Conclusions from Physiological Studies ^^^^

Uses of the Peanut ^^

Summary of Results 335

Literature Cited ^•^'

Explanation of Plates ^3

302 WALDRON— THE PEANUT

Introduction

"Few plants present as great interest or diversity of problems for
botanical study as the peanut. Agriculturally it is of great import-
ance as a soil renovator and forage plant. From an economic stand-
point, its products are of great value, every part of the plant being
of some direct or indirect use. The peanut, potato, cotton, tobacco
and Indian corn, — five plants which are exerting great influence in
the world's commerce and industries — were contributed by the new
world. The peanut, although still in the background of some of
these, promises to rival, if not surpass, them in importance. It is
often grown in the economic house of botanical gardens and as a
novelty out of doors. After a brief botanical study of the plant from
the morphological standpoint, certain problems were presented
which bear an important relation to its physiology, and are rather
striking in their bearing on some well known ecological problems.
A study of the history of the plant, and brief discussions concerning
its economy in nature and its utilization by man, have now been
taken up in succeeding pages. The writer, in here recording these
observations, feels that he has made but a beginning and hopes to
continue investigations in the future.

The writer wishes to express his grateful appreciation to Professor
John M. Macfarlane, of the University of Pennsylvania, for his
many suggestions and kind guidance in working up this treatise.

History

Introduction. With few exceptions, authors agree that the ori-
ginal home of the peanut {Arachis hypogaea) is uncertain. In the
mind of the writer, there are sufficient facts at hand to state definitely,
as have one or two already, that it is a native of Brazil, although,
as with many other extensively cultivated plants, it has never been
recognized in the truly wild state. There is no evidence to contra-
dict the view that it is a native of this part of South America. The
writer gives below a synopsis of his studies regarding the native
home of the plant and its history so far as known in relation to man.

Early American Records. The earliest mentions in any existing
literature are those pertaining to Brazil and Peru, and these ante-
date any found in European works. Acosta , in his work published
in 1598, refers to it along with other plants which are native to
Brazil, and calls it "mani," a name still applied to it among Spanish

WALDROX— THE PEANUT

303

speaking people of South America. Monardes", according to
Marcgraf and Piso^, indicated its presence in Peru about this time,
giving it the name of "anchic." Aside from these and other early
mentions in literature, fruits of the plant were found in tombs at
Ancon, Peru. Their presence there undoubtedly antedates the
Spanish conquest, and so, also, any written record. According to
Dubard*, it was taken from Brazil to Peru sometime before the

Figure i (after Marcgraf & Piso). Mundubi Braziliensis.

sixteenth century and there "was cultivated from an early unknown
date." Among European works, Parkinson^ in his celebrated
"Theatrum Botanicum" published in 1640, gives an illustration of
the fruit, which is very likely the first. A few years later, (1648)
Marcgraf and Piso were the first to figure the whole plant (Fig. i).
It seems worth while to quote parts of Parkinson's quaint descrip-
tion as follows:

"ArACHUS VTroyeio- AmERICANUS. UNDERGROUND CiCHELING OF

America or Indian Earthnuts."
"The Indian Earth-nuts (the figure whereof, I give you together
as they are termed to us by them that have brought them us) are

304 WALDRON— THE PEANUT

very likely to grow from such like plants as are formerly described,
(Species of Vicia) not onely by the name but by the sight and taste
of the thing it selfe, for wee have not yet scene the face thereof
above ground, yet the fruit, or Pease-cods (as I may so call it) is
farre larger, whose outer huske is thicke and somewhat long, round
at both ends, or a little hooked at the lower end, of a sullen whitish
color on the outside, striped, and as it were wrinkled, bunching out
into two parts, where the two nuts (for they are bigger than any
Filberd kernell) or Pease doe lie joyning close one unto another,
being somewhat long, with the roundnesse firme and solide, and of a
darke reddish colour on the out side, and white within tasting sweet
like a Nut, but more oily."

Concerning the introduction into Europe Parkinson's discussion
indicates that they were introduced into Portugal. He received
specimens sent from Candy and Lisbon. To quote him further he
states that the Indian earthnuts are found in "most places of Am-
erica, as well as to the South, as West parts thereof, both on the
Maine and Islands; and generally called by our English Sea-men
that goe into those parts Earth-nuts, erroneously enough, as they do
most other things that they there meete with."

Eight years after Parkinson's reference the plant was described as
follows by Marcgraf and Piso — "Mundubi Brasiliensis Herba, in
pedalem aut bipedalem altitudinem adsurgit, caule quadrato aut
striato, ex viridi rufFescente & piloso. Hinc inde enascuntur ramuli
primo quasi caulem amplectentes & foliolis angustis, acuminatis.
stipati; mox habent nodum ac trium vel quatuor digitorum longi-
tudine extenduntur; continetq; quilibet ramulus quatuor folia,
duo semper sibi opposita paulo plus quam duos digitos longa sesquidi-
gitum lata superne, laete viridia, instar trifolii, inferne paulum canes-
centia, nervo conspicuo & subtilibus venulis quasi parallelis dotata,
raris quoque pilis vestita. Ad exortum ramulorum qui folia gerunt
prodit pediculus sesquidigitum circiter longus, tenuis, flosculumgerens.
flavum .& per oras rubentem duobus foliolis constantem, more vic-
iorum aut trifolii. Radix illius baud longa, tenuis, contorta, filam-
entosa, cui adnascuntur folliculi ex albicantegrysei,figura minimae
cucurbitae, oblongae, magnitudinae Myrobalani fragiles: quilibet
autem continet in se duos nucleos, pellicula saturate purpurea vestitos,
carne intus alba, oleaginosa, sapore pistaceorum, qui comeduntur
cocti & inter bellaria aponuntur. Multum tamen comesti capitis
dolores causare ajunt. Fructu integro quassato nucle intus strep-
unt

WALDRON— THE PEANUT 305

Linnaeus^ says it inhabits Surinam, Brazil and Peru, but does not
state whether it is wild or cultivated.

Lack of Evidence as to its extra-American Origin. Among old world
literature antedating the i6th century, no mention is made. It
was thus unknown there before the discovery of the new world.
According to Watt^ all Greek, Latin, Bengalese and Arabian writers
are silent concerning the plant. This is very significant since the
peanut is too valuable a plant to have been known to Sanskrit
speaking peoples and not be used by them. Such a plant could
hardly have an antiquity among them without some record being
kept. Until quite recently a mention by Theophrastus of an Egyp-
tian grown plant was thought by some, to be a reference to Arachis,
but this has since been disproved. If it had ever existed in Egypt
it could still be found there. Furthermore no mention is made of it
in the works of Forskal*^ or Delile^ It is not recorded by any
early writer on the flora of India. According to De Candolle in
Dr. Bretschneider's study of Chinese works'', the statement is
made that its introduction into that country was in the sixteenth
century. It is not mentioned in ancient Chinese literature. This
suggests the possibility of its introduction there from Peru by such
expeditions as Magellan's.

De Candolle states: "The antiquity of its cultivation in Africa
is an argument of some force which compensates to a certain degree
its antiquity in Brazil." The only points offered to indicate an
African antiquity are (i) the statement by Sloane'^ that it was
used as food on the early slave ships sailing between Africa and
America, and (2) that it now has a wide area of cultivation there,
both of which could very readily have occurred after its introduction
from Brazil. The writer would suggest that the very earliest ships
to sail from Brazil to Africa took sctdsoi Arachis to that place, and
the environment of the west coast being ideal for its growth, its
cultivation very early became widespread; and this has, during the
last century, developed into a great industry in the French colonies.
Pison, in his early Brazilian work figures a somewhat similar plant,
in its habit of fruit production, to Arachis, but states it to be African,
while he says Arachis is Brazilian. This was a species of Voand^eia.
To quote De Candolle again, he states, "the silence of Greek, Latin
and Arab authors, and the absence of the species in Egypt at Fors-
kal's time lead me to think that its cultivation in Guinea, Senegal,
and the east coast of Africa is not of very ancient date; neither has

306 WALDRON— THE PEANUT

it the marks of a great antiquity in Asia. No Sanskrit name for it
is known, but only a Hindustani one. Rumphius^^ says that it
was imported from Japan to several islands of the Indian Archipelago.
It would in that case have borne only foreign names, like the Chinese
name, for instance, which signified "earth bean." At the end of the
last century (19th) it was generally cultivated in China and Cochin-
China. Yet, in spite of Rumphius's theory of an introduction into
the islands from China or Japan, I see that Thunberg does not
speak of it in his Japanese Flora. Now, Japan has had dealings
with China for sixteen centuries, and cultivated plants, natives of
one of the two countries, were commonly early introduced into the
other. It is not mentioned by Forster among the plants employed
in the small islands of the Pacific. All these facts point to an Am-
erican origin." No authors speak of it v/ild or uncultivated in either
hemisphere. Those who speak of it in x'\sia or Africa carefully say
it is cultivated. Piso, in writing of Brazil, says the species is planted.
Marcgraf does not mention it as cultivated, indicating however
that it may have been. As to foreign names such as the Chinese,
meaning "earth-bean," they are all such as would occur to any one
upon seeing the plant. Contrary, therefore, to the suggestion of
some authors, little significance need be attached to the American
names not having accompanied it in its travels to Japan and the
Orient.

According to Watt, Sir George Birdwood in his Bombay products
gave it a Sanskrit name meaning earth-gram. This name has been
repeated by some subsequent writers without the authenticity of
it being inquired into.

Distribution. According to Bentham^* in "Flora Brasiliensis,"
there are seven species of Arachis found in Brazil, six of which are
found in the wild state. The other {A. hypogaea) he states is generally
cultivated in all warm parts of the world. Now, De Candolle well
remarks in this relation that: "A genus of which all the well known
species are thus placed in a single region of America can scarcely
have a species common to both hemispheres; it would be too great
an exception to the law of geographical botany." By referring to the
accompanying outline map (Figure 2) which shows the reported
distribution of these species it will be noted that A. prostrata is the
most widespread, and A. pusilla the most restricted of the group.

The writer asks if varieties of this plant, so generally grown in
such environment as exists in Brazil, may not be cultivated forms of
one or more Brazilian species?

WALDRON— THE PEANUT

307

Of the several cultivated varieties grown today there are recog-
nized two general types of plants as follows: (i) The bunch type,
growing erect and bearing its fruit around the base of a single stem.
The Spanish variety is an example. It can withstand considerable
moisture conditions and its erectness suggests a shade loving ten-
dency. (2) The trailing type, with its several branches spread on
the soil, succeeds best in a hot sandy soil, indicating greater xero-
phytic tendencies. The Jumbo variety is an example. Now, the
wild Brazilian species A. pusilla, is an erect plant, simulating the

Figure 2
Outline Map of Brazil indicating the reported distribu-
tion of the different species of Arachis.
I. A. pusilla.
1. A. prostrata.

3. A. villosa.

4. A. glabrata.

5. A. marginata.

6. A. tuberosa.

bunch variety and is reported as growing in dry woods and shady
places. Another, A. prostrata^ is more trailing and grows in open
sandy places and so, simulating the prostrate cultivated variety, is

308

WALDRON— THE PEANUT

more xerophytic. Thus the possibility is suggested, first, that the cul-
tivated bunch varieties are derived from such a species as A. pusilla
and second, that the prostrate varieties are derived Jrom A. prostrata.
Other evidence in support of this theory is seen in the marked differ-
ence in the histology of the fruits of the two domesticated varieties.

Figure 3 (after Dubard)
Types of peanuts taken at random from

1. Tombs at Ancon, Peru.

2. Java.

3. Tonkin.

4. Madagascar.

5. Madagascar.

6. Spain.

7. Dahomey.

8. Senegal.

9. Spain.

The following summary taken from Dubard further indicates the
possibility of our present day varieties being derived from two such
wild plants. He states that there were two types of peanuts dis-
tributed over the world from South America; one being a two-
seeded Brazilian type and the other a three-seeded Peruvian type.
The first form was carried to West Africa by Portuguese negroes,

WALDRON— THE PEANUT 309

according to this author. It was taken to Europe by early travelers,
the Portuguese being there the first propagators, as indicated by
Parkinson and others. Further evidence in support of this east-
ward distribution, which seems more natural from the geographical
standpoint, is, that (i) all early illustrations of European and
Brazilian works show two seeds, and (2) those of West Africa are
two-seeded. The Peruvian type was a variety created in Peru from
a form carried there from Brazil some time before the sixteenth cen-
tury. The question now arises as to the possibility of this plant
being a different species from the common two-seeded type. Ben-
tham states the number of seeds to vary from one to three in this
genus. This being true, any species might well have been selected.
However, Dubard later describes differences in the structure of
these which are sufficiently marked to separate them into species.
About Magellan's time, this three-seeded form was carried from Peru, to
the Moluccas, Phillipines, Indo-China, Asia and Madagascar. If this
is true, we should find the three-seeded fruits in these Pacific local-
ities, and this Dubard proves to be true, by comparing specimens
taken at hazard from Java, Indo-China and Madagascar. (Figure 3.)

Finally, if peanuts of to-day from Spain and North America be
examined, the above two types will be found, indicating a meeting
again of these, after having been carried around the world in opposite
directions, yet remaining distinct in character. The Peruvian form
was undoubtedly carried north and east, but at a date much later
than its westward spread. The supposition, according to record,
that the two were present at about the same time, — the one in Africa
and Europe, the other in the Orient, — further supports this view.
Watt states that one name given it in India, where it is much culti-
vated, is "Manilla-Kottai." This suggests its introduction there
from the Phillipines. All these facts relating to the distribution
are suggestive, when the lack of evidence of its presence during ancient
times in China, India, Africa and Europe is considered.

Since Dubard does not describe the plants of his two forms, it is
impossible to determine whether or not they might correspond to
the erect and prostrate types. He does describe the fruit and seed
of each, however, and it is found that his Brazilian form corresponds
to some of the erect, and his Peruvian form to some of the prostrate
varieties of to-day. If such a history is possible, which does not
seem unlikely, the reported distribution of the two wild species —
A. prostrata and A. piisilla — is again significant. The former,

310 WALDRON— THE PEANUT

found in all parts of Brazil, would be the one most naturally carried
to Peru rather than the more restricted latter species that is found
only in the eastern part. The lack of full descriptive literature on
the distribution and morphology of wild and cultivated varieties
opens up opportunities for further investigation.

To-day, varieties are rapidly being increased In number by man.
This fact, with the ease of intercourse between different countries,
explains in part at least why there are such varieties as the Virginian,
Spanish, African, Asiatic, etc. It does not indicate in any way the
native home of the peanut, which is undoubtedly Brazil.

Recent Literature. During the last 25 years most of the literature
on the peanut has been largely concerned with either its culture, uses
or chemistry. The writer does not attempt to summarize these and
includes titles of but a few of the more recent publications in his
bibliography. By referring to the Experiment Station Record of
the United States Department of Agriculture, many such references
may be found. During the past few years, publications concerning
its culture and varieties have been issued by several of the Agri-
cultural Experiment Stations of those southern states in which the
peanut is becoming an important crop.

Concerning the study of the plant from a morphological stand-
point, little has been done. In 1895, Pettit^'^ worked on the fruit
stalk. She described its structure in detail and discussed Its phys-
iological relation to the plant.

Winton'^ published the results of a histological study of the
mature fruit, undertaken especially to secure data for use in the
microscopical examination of peanut products. In this relation he
described and illustrated the cell structure of the fruit, testa and
cotyledons. Adam^ in 1908 published a fine work concerning Its
history, growth habits, varieties, culture, products and Industry In
western Africa.

General Morphology

Description. Since there are two well recognized forms of the
peanut plant, the author suggests a division of the Linnaean species
Into the two sub-species — (i) Jastigiata for the bunch type, and (2)
procumbens for the prostrate type, and including under each a num-
ber of varieties. Adam gives the full species name asiatica for the
former, and qfricana for the latter. Such are not only names of
varieties and localities, but suggest an erroneous origin. The fol-

WALDRON— THE PE.ANUT 311

lowing is a description of the species, the suggested sub-species and
the common varieties.

Arachis hypogaea
Family Leguminosae Sub-family Hedysareae

An herbaceous annual.

Roots — fibrous, delicate and white when young. Root hairs
usually in rosettes at the base of the side roots — rarely with normal
tip hairs. Nodules spherical; surface gray, interior pink; appear-
ing when plants are 8-10 weeks old.

Hvpocotyl — 2-8 cm. long, sometimes slightly swollen at base dur-
ing germination.

Stems — 30-80 cm. long, erect, or prostrate, more or less hairy,
tough, flexible, slightly quandrangular. Main stem usually branch-
ing early into cotyledonary branches.

Cotyledons — low epigeal, green, with short, thick petiole; remain-
ing fleshy for two to three weeks, when they dry and drop off.

Leaves — sensitive to light, 8-12 cm. long, alternate, stipulate,
pinnately compound. Stipules linear-lanceolate, erect, striate. Pet-
iole straight, firm, with a single groove along the upper side; a pulvinus
at its base. Leaflets 2-5 cm. long, four in two pairs, oblong to obo-
vate; apex rotund and tipped by a tiny spine; veins pinnately
arranged; under surface slightly hairy; each attached to the petiole
by a short pulvinus which causes them to close together vertically
in pairs at night.

Inflorescence — an axillary, usually three flowered, fascicled and
reduced head.

Flowers — yellow, the larger more terminal ones usually sterile and
adorning the plant for some time; the more axial numerous, basal
ones usually fertile, smaller, more or less hidden, and born on short
peduncles which elongate after fertilization. Calyx forming a long
stalk-like tube with one narrow lobe as a lower lip, the upper broad
and four-toothed. Corolla with a large yellow, orange-striped
standard, two small wings, and a tiny, incurved, beaked keel. Sta-
mens ten, monadelphous, versatile; five often with fertile anthers
attached near their base; the alternate ones absent or short and
fixed at their center. P/j///monocarpous; style long, slender; ovary
small, at base of the long calyx tube, one celled with one to six ovules;
after fertilization the floral envelopes drop away, and the ovary,
now sharp-pointed and strengthened, is pushed by the rigid, recurv-

312 WALDRON— THE PEANUT

ing pedicel one to three inches into the soil; after penetration it
begins to swell and ripen into a fruit.

Fj'uit — an indehiscent legume, oblong, reticulated, thick, coria-
ceous, beaked, swollen around the contained seeds, provided with
absorptive hairs when nearing maturity. Fruit stem or gynophore
reddish and slightly hairy above the soil — white and matted with
absorbing hairs below.

Seeds — i-i ^ cm. long; cotyledons thick, fleshy, oily; radicle
short, straight. Testa thin, papery, membranous, varying from
cream to pink to dark red color.

Sub-species — -fastigiata Waldron

Main stem erect and branches all in an upward diagonal position,
giving the plant a bushy appearance. Fertile flowers axially group-
ed near base of plant. Fruit clustered below the main stem. Seeds
usually small and oblong.

Variety — White Spanish

Plant 20-30 cm. tall in average soils — foliage abundant and heavy.
Pods small, adhering well to the plant, entirely filled by two seeds
with pink to brownish testa. Very productive with high oil content.

Variety — Red Spanish

Similar to the White Spanish, except that the seed coats are red
and the pods somewhat larger — less productive than the White
Spanish.

Variety — Valencia

Plant 25-50 cm. tall. Pods long, medium thickness, clinging
poorly to the plant, and containing two to four closely crowded
seeds with red testas.

Variety — Tennessee Red

Plants similar to those of the Spanish varieties. Pods long, cling-
ing to the plants and containing two to six seeds with dull red seed
coats.

Sub-species — prociimbens Waldron

Loosely branching, spreading plants (semi-erect in one variety);
the stems often later ascending. Flowers and fruit considerably
scattered along the prostrate stems. Seeds large, more or less
pointed.

WALDRON— THE PEANUT 313

Variety — North Carolina, Florida Runner, African, Wilmington, etc.

Rank growing plants with dark green massive foliage. Pods not
clinging well, medium sized, containing two, sometimes three, mod-
erate-sized seeds with reddish testas.

Variety — Virgin ia Ru n n er

Similar to the North Carolina, but with much larger pods and
seeds.

Variety — Virgin ia Bunch

Plants semi-erect with light foliage, often appearing among plants
in fields of the Runner type. Pods large, adhering well to the plant,
bright, clean, with two, sometimes three, seeds covered by light
brown coats.

Variety — Jumbo

The same as, or possibly sometimes a strain of, the Virginia Run-
ner or Virginia Bunch varieties.

Histology

Root

The internal structure of the root of Arachis is of the normal di-
cotyledonous type. The epidermal relation, however, of young roots
and the production of root hairs is quite striking. It has been re-
ported by Pettit and Richter^'^ that no hairs are produced on the
plant. The author, however, found them on all plants examined
but usually in different position from the normally produced tip
hairs. Although the tip hairs were found, they were rare and appear-
ed only on young vigorous plants. Usually they are present in the
form of rosettes on and at the base of newly formed side roots. As
seen in Plate LXXIX, Fig. 4, those produced nearest the base are
comparatively long, but are gradually reduced in length until none
appear. Aside from the position relation, these rosette hairs are of
the normal root-tip hair structure. The tip hairs when present are
usually rather short and scattered, occurring on young delicate, usu-
ally few-branched, roots. No hairs of either type have been ob-
served on the main root either during germination or later.

Young elongating roots of the plant which bear no root hairs often
have their cuticle mucilaginized causing the soil particles to adhere
as if hairs are present. These roots are often very white, delicate

314 WALDRON— THE PEANUT

and semi-transparent. Placing a portion of one on a slide for exam-
ination under the microscope must be done with care, for the epider-
mal and outer cortex cells seem readily to fall apart like so many
poorly cemented bricks becoming loosened. This is due, in part at
least, to pressure from within, since, as noted by Pettit, the inner
meristems develop much faster than the dermatogen. This is mark-
edly evident on the primary root where the surface cells are continu-
ally peeling off in rows or patches. Sufficient protection, according
to Pettit, is afforded by a cutinized outer wall being formed by the
cells which become exposed. Pettit also states that the lateral
roots act in a similar way, but the present writer did not find this to
be the case with his plants. These roots are always normal in this
regard, although very delicate. As they increase in age, their outer
surface is supplied with a regular periderm, and it is only in the
early stages of radicle growth that this peculiar habit is observed.

Stem

The stem is normal for dicotyledons. The epidermis is composed
of a layer of small, thickly cuticularized cells interrupted by stomata
in young stems and by a corky lenticel proliferation in old ones.
Three-celled hairs are scattered along young stems. These are
typical of many other Legiminosae, each hair having a long pointed
terminal cell with two tiny, flattened basal cells. There are also
small crystal cells arranged in groups of two to four, each containing
one rectangular crystal. The cortex is six to eight layered and com-
posed of much larger thin walled cells. At the outer extremity of
the primary bundle areas of the vascular system are large patches
of highly indurated hard bast. This, along with the quite extensive
and considerably indurated xylem, give a tough and flexible char-
acter to the stem. The pith in young stems is composed of round
extremely thin walled cells filled with starch. As the stem matures,
however, the pith becomes more or less broken down resulting in a
semi-hollow condition. Concerning the cambium, a very interesting
relation is noted later as to its development in the fruit stalk (see
page 314).

Leaf

The leaf structure of the peanut is most striking, especially when
the xerophytic tendencies of the plant are considered. There are
numerous average-sized stomata on both surfaces that are neither

WAI.DRON— THE PEANUT 315

raised nor sunken. They average from fifteen to twenty per square
millimeter, and each is surrounded by two unequal subsidiary cells.
Both epidermal layers are also supplied with numerous specially
formed crystal cells (Plate LXXIX, Fig. 5.) which, when the leaf is
young, are small, each containing a single rounded crystal. Later,
these cells become fused into what might be called an "epidermal
vessel," irregular in shape and containing two to thirty crystals
arranged in clusters or irregular rows. Solereder^^ notes the pres-
ence of these, but does not mention the later fusion into one. In
attempting to determine the origin of these, a stem apex with a
young bud attached was sectioned. It was found that in very
young leaves all the epidermal cells were alike, and devoid of crystals.
Immediately after the last cell division, however, some of the cells
increased in size to form the normal epidermis, while others remained
small to become the crystal-containing cells. In some of these it
was noted that the small crystal seemed to be a part of the nucleus,
possibly formed within and by it. It was in a pellicle-like projection
which later separated away. In all such young cells the crystal was
imbedded in a protoplasmic matrix which gradually became mucil-
aginized in old cells. The mesophyll of the leaf is composed of a
two- to five-layered palisade tissue immediately below the upper
epidermis and a single layer of water storage cells next to the lower.
These two features are more typical of a xerophytic plant than is
the presence of the above mentioned stomata. A loose, comparative-
ly thin layer of spongy mesophyll separates the palisade and water
storage layers. The petiole and lower epidermis of the leaflets bear
the typical three celled hairs already mentioned as found on the stem.

Fruit

The anatomy and physiology of the fruit were carefully studied
for the purpose of discovering, if possible, some facts concerning its
hypogeal development. Among other members of the Leguminosae
to share this peculiarity are Amphicarpa monoica, Trijolium subter-
raneum^ species of Voandzeia and of the new African genus Ker-
stingiella. Amphicarpa bears two kinds of flowers, and accordingly
two forms of fruit, only one of which develops underground. The
flower which gives rise to this subterrenean fruit is formed and al-
ways remains underground. Trijolium subterraneum bears but one
type of flower formed in heads. The peduncle, after flowering.

316 WALDRON— THE PEANUT

lengthens and sinks into the soil carrying the head with it. The
seeds do ripen above ground, however, and will germinate.

Differing from both of these, Arachis and the other two genera
mentioned above have a nearly uniform type of flower, the ovary of
which is pushed into the ground by a growth at its base. The elong-
ating fruit stalk is called a "gynophore."

Anatomy of the Young Gynophore. Observations by the writer
correspond to most of those of Pettit, whose article is chiefly concerned
with the structure and development of this organ. Sections made
longitudinally through young flower buds reveal a nearly sessile
ovary usually containing two parietal ovules each, on a short funi-
culus (Plate LXXIX, Fig. 8). There are ii to 13 bundles which ex-
tend through the base of the ovary to the tip, branching more or less-
in their course. Along the inner edge of each bundle are tannin
pockets. After the egg is formed and fertilized, the reduced ovarian
axis begins to elongate to form the gynophore. The ovary and em-
bryo sac remain unchanged in this condition until the gynophore is
mature. This will be further discussed under physiology. The
later cytological study of the embryo was not attempted.

The meristematic tissue which gives rise to this growth is mostly
situated just below, and around the base of the ovary. That below
the ovary forms the pith, while that around the base forms the bundle
tissue and outer cortex. A few dividing cells forming the latter
were found well up around the ovarian cavity. The epidermis of
the tip becomes sharp pointed and highly lignified in its outer walls.
In Plate LXXIX, Fig. 9, is seen the area at one side of the tip where
the style was formerly attached. This style is terminal in very young
buds, but the later lateral position of the scar is prearranged for
by a special development of a few large lignified epidermal cells at
one side of its base. (Plate LXXIX, Fig. 8). These grow forward
a little, and form the sharp point of the ovary.

Anatomy of the Mature Gynophore. While the structure of this
fruit stalk corresponds to that of the stem of any herbaceous dicoty-
ledon, its manner of development resembles that of ordinary roots.
There are no lateral appendages, and so no nodes and internodes.
There are two distinct divisions of this organ (i) the epigeal part,
with smooth, red pigmented surface bearing a few three celled hairs
similar to those on other parts of the plant; (2) the hypogeal white
part, whose surface produces single-celled absorptive hairs. (Plate
LXXIX, Fig. 7.) The surface of the aerial portion is covered with

WALDROX— THE PEANUT 317

stomata and lenticels. The epidermal layer of this has a few scat-
tered crystal cells similar to those found on the stem.

The cortex is composed of six to eight layers of round thin walled
cells. Internal to this is the vascular system composed of a ring of
bundles, each with a large patch of highly lignified bast on its outer
side.

Concerning the cambium layer, Pettit says, "There is an indication
of the formation of a cambium ring, although it never occurs even
in the oldest portion of the organ." She goes on to explain the ap-
parent presence of meristematic tissue between the bundles, which
the writer considers interfascicular cambium. She continues, "The
cells are, in their early stages, no larger than the pith cells, but as
they become older they increase rapidly in both tangential and radial
diameters. This process, however, appears insufficient to keep pace
with the growing intrafascicular cambium, and they now become
meristematic, forming new walls which are at first tangential; later
radial walls are found. In this manner arise clusters or bands of
relatively small cells extending from bundle to bundle. While
these small cells appear like the ordinary meristematic tissue of
stems whose cambium is formed after the bundles appear, they do
not continue meristematic; at least in the organs studied there is
little evidence that these small cells produce lasting tissue of any
kind, and none whatever of the formation of phloem and xylem ele-
ments." The above description is correct for the cause and manner
in which they develop and appear, but the writer feels that this
tissue is in all respects meristematic as a part of a continuous cam-
bium ring. The author has seen xylem and phloem elements cut
off from it to form secondary bundles, and thus proving that it has
the ability to form these. Also, the presence of such cells in a nearly
continuous line with the intrafascicular cambium suggests an hered-
itary tendency to form it even though it is less active than in other
stems. It does apparently connect the xylem patches, but so does
the corresponding tissue of many stems. The fact that they divide
even once is sufficient proof, since the resulting cells are permanent.
The pith is composed of thin walled cells stored with starch until
the fruit begins to form. Later the pith breaks down as does that
of the stem and the gynophore becomes more or less hollow.

The anatomy of the subterranean part of the gynophore differs
from that above ground in the following ways: (i) All of the epi-
dermal cells become extended outward into long absorptive hairs

318 WALDRON— THE PEANUT

simulating typical root hairs. (2) a growth in thickness occurs by
a process similar to that of periderm formation, by which the diam-
eter of the subterranean part is somewhat increased. Pettit notes
another difference in the absence of what she calls plasmolytic cells
situated in the center of the cortex of the epigeal part. These cells
have not been observed by the author.

The absorptive hairs (Plate LXXIX, Fig. 7) are large, unbranched,
one celled and average nearly a millimeter in length. Each is
slightly enlarged at the base which represents the size of the original
cell from which it springs. No stomata were seen, but lenticels
were present which developed into a white proliferation of cells
when exposed to a moist atmosphere as did some of those of the
epigeal portion.

The examination of a cross section of this hypogeal area showed
the outer layer of the cortex dividing into two to four layers of cork-
like cells. The cells of this sub-epidermal layer are somewhat larger
than those of the rest of the cortex. Although these apparent phel-
logen derived cells have tbe appearance of periderm, according to
Pettit, they are free from suberin, as would be expected from the
presence of absorption hairs on their exterior.

Anatomy of the Yoimg Fruit. As noted above (see page 316) the
ovary, situated at the tip of the gynophore remains inactive until
the time comes for fruit maturation. The epidermis at this time
is composed of much deeper and narrower cells radially, than that
of the non-hairy part of the gynophore. They become deep, tap-
ering and lignified at the tip forming a hard, but not capped apex.
The lumen of those at the tip contain numerous granules that are
not evident further back and suggest a relation to the geotropic
reaction of the gynophore. Three or four hypodermal layers, that
later form the outer mesocarp, are composed of markedly cylindric
cells. A branching bundle system within this is a continuation of
that of the gynophore which gradually disappears toward the tip.
Just interior to this are a few layers which later assume marked
appearance and importance as tissue which becomes gradually lig-
nified to form the strengthening inner shell layer of the fruit. The
innermost tissue next to the ovarian cavity is composed of several
layers of tiny nearly square cells arranged in radiating rows.

Anato7ny of the Developing Fruit. When the ovary begins to en-
large, the epidermal cells elongate longitudinally and later become
ruptured. This epidermis with the subjacent layer, is rubbed or

WALDRON— THE PEANUT 319

Stripped off, which is a very unique event in fruit maturation. Ap-
parently the epidermis does not keep up with the growth from within.
Within the epidermis several layers of periderm have already ap-
peared, similar to, and continuous with, that of the lower end of the
fruit stalk (gynophore). Till this occurs and until the fruit is one
half to three quarters full size, no hairs are formed on any part of
the ovary or young fruit. About this time, however, and lasting
until the fruit is mature, there appear on this layer irregular, often
branched, one-celled absorptive hairs (Plate LXXIX, Fig. 6).

The developing bundles have increased in size, and with their
connecting branches, form ridges which later give the reticulations
to the fruit. Meantime a few layers, just interior to these, are be-
coming remarkably indurated to constitute the solid enclosing
chamber of the mature fruit. The inner endocarp area of small
cells has enormously thickened and from the time the fruit has be-
gun to swell until nearly ripe, this area is composed of very large,
thin walled, pith-like cells which contain sugar. This area remains
thick and the developing seed small until a short time before maturity.
It forms a large part of the fruit at this time.

Anatomy of the Mature Fruit. Winton notes the presence of an
epidermis and states that it is not easily seen. As noted in the dis-
cussion of the developing fruit, this could only be a pseudo-epidermis
that he has mistaken for the already shed epidermis, and the fact
that it is the small-celled third laver of the ovarian tissue mav ex-
plain why it is distinguished with difficulty. As in the nearly ma-
ture fruit, absorbing hairs are present as wall extensions of it. These
hairs were not observed to be as often branched as in the younger
fruit. One or two appeared to be septate.

Just below the outer absorbing layer are several rows of brick
shaped, thin walled cells, simulating those of the hypogeal gynophore
in appearance and position. Within this are several layers of thin
walled cells that have collapsed. Apparently new cells were not
produced for this area to allow for the expansion of the fruit. These
surrounded a few layers of unbroken, rounded cells in which are em-
bedded the branching vascular bundles. Large bundles are arranged
longitudinally, and are connected by short smaller cross bundles,
the whole system forming the reticulation of the peanut shell. The
xylem and phloem of these have become highly lignified.

Attached to the inner edge of the bundle network is the solid hard
enclosing shell of the peanut, composed of the now mature lignified

320 WALDRON— THE PEANUT

mass of cells mentioned in connection with the maturing fruit. The
cells forming this are now remarkable in their shape, size, wall thick-
enings, and branches. (See Winton's article.)

The structural relation of the carpellary wall to that of a leaf,
from which it is modified, is recognized, but with more difficulty
than that of many aerial leguminous fruits. By carefully splitting
open a fruit, as one ordinarily shells a peanut, it will be noted that
the seeds are attached to the somewhat convex side opposite that of
the beak. Thus the dorsal is recognized from the ventral suture.
Histologically these areas are not discernable in young ovaries,
except by noting the location of the ovule attachment, or that of
the style. This also locates exactly the position of the beak in
the mature fruit. In mature fruit shells there is a ventral suture
along which they quite readily split, due to a weaker, less lignified,
loose line of tissue as seen under the microscope.

Concerning the exocarp, mesocarp and endocarp and their origin,
there is such a marked change and fusion of parts that any sharp
line of demarcation is impossible. The exocarp, which is typi-
cally derived from the lower leaf epidermis, is lost during fruit ma-
turation. The name endocarp, derived from the lower epidermis,
might be applied to the soft, internal tissue called inner parenchyma
by Winton. Practically the whole of the shell then would be the
mesocarp and can be subdivided into hypoderm, bundle area and
brs layer.

The anatomy and cytology of the embryo not having been at-
tempted, that of the mature seed is also omitted. For details oj
the latter the reader is referred to Winton's article.

Physiology

Root Hairs

To the writer, the thought suggested by others, as noted in the
histological discussion, that Arachis bore no root hairs, seemed con-
trary to expectation. The plant, with a semi-xerophytic tendency,
and growing well in a warm, loose soil, would be expected to have
them, at least when moisture is sufficient to stimulate their produc-
tion.

Plants started in the greenhouse and carried on in flower pots
were therefore examined. Of twenty-five plants, one showed hairs
present near the tips of two of the young vigorous growing roots.

VVALDRON— THE PEANUT 321

They were, however, very short and comparatively few in number.
It was noted, however, with some surprise, that on the roots of
nearly every plant, tufts or whorls of hairs were present as rosettes
at the base of some of the side rootlets. These were much longer
than the first-mentioned type, as is set forth in the figure. The
possibility that the influence of potting may have in some way
caused the development of these rosettes as well as tip hairs, led to
an examination of the roots in the center of pots, and of roots on
plants which had been carried forward in boxes (2x3x1^ ft.)
Such roots have less air drainage than those along the inside wall
of a flower pot. Plants were removed, their roots carefully washed
out, and tufts of hairs were found at the base of many newly formed
side roots.

The presence of normally produced tip hairs was carefully watched
for, but none was found. The only plant mentioned which had these
was one, the roots of which were in a moist air compartment, formed
by the drainage hole of the bottom of the flower pot, the broken
crocks just above, and the cinder bench below. This suggests,
therefore, that optimum oxygenation is a necessary factor for the
tip hair growth. None were found on the roots of any plants in the
soil, either under dry or moist conditions. Since the plant seemed
to have at least a hereditary tendency under certain conditions
to produce these normal tip hairs it was considered worth while to
determine, if possible, the exact causes or stimuli which affect their
growth and that of the rosette type. Observations and experi-
ments were, therefore, carried on with this in view, as well as to
determine the causes and method of production of, and differences
between, the two forms. The results of this investigation are de-
scribed in succession to the following discussion of the works of
others on this subject.

Concerning the presence and absence of root hairs, Strasburger
states that in some few instances as in some conifers plants bear no
root hairs. Jost"' says that few plants produce none, probably re-
ferring to aquatic types. Haberlandt" refers to two stages in the
specialization of absorptive tissue in plants, (i) Some plants are
content to increase their absorptive surface by a greater output of
side roots, the epidermal cells of which are flat or slightly convex in
their outer walls. But he states that this type includes marsh and
aquatic plants, and is less advanced in this regard. If this is a stage
in specialization, would it not be possible to think of it as a reduc-

322 WALDRON— THE PEANUT

tion change following a former evolved condition in root hair pro-
duction due to adaptation to changed environment? Corn and
many other plants produce none when put in water. (2) Other
plants produce root hairs near the tips of their roots by elongation
of the outer epidermal cell walls. Since many hair producing plants
cease to bear these when in contact with water or saturated soil,
it would seem to the writer that instead of being two stages of spe-
cialization, it is an example of two types of absorptive tissue depen-
dent on ecological factors. It is a more or less epidermal surface
extension for absorption, dependent upon the amount of causing
stimuli present. The hairless aquatic plants may have had hairs
at some time, and some do produce them when growing in dry soil
again.

As to the cause for root hair production, PfefFer^^ states that
too little or too much water hinders, while darkness and contact
accelerate. Snow in an extensive investigation on the causes of
their development finds that they are accelerated by a retardation
of growth, by mechanical means, or substratum resistance, espe-
cially if the roots of such are allowed to grow in a moist atmosphere.
She finds that they are retarded by a saturated atmosphere at high
temperatures, by a lack of oxygen, and by a saturated soil; light and
darkness, however, have no material effect.

Observations and Experiments with Root Hairs

It was noted that if seeds were planted in a heavy soil and germ-
ination was retarded by lack of moisture, the hypocotyl would some-
times swell considerably and give off adventitious roots which
branched profusely. By drawing the soil away, after the radicle
had grown an inch or two, until the lower end of the hypocotyl was
well exposed, the upper part of the side, and of the adventitious
roots with their hairs could be kept growing in saturated air. This
gave a good opportunity to determine the effect of sunlight on the
growth of hairs, as compared with those on a few plants whose roots
were kept in the dark, but also exposed to the air. The foliage of
both sets of plants had the same leaf exposure, so that the activity
and growth were as near the same in all as possible. In all cases,
both in light and in darkness, the rosette hairs were found at the base
of the side rootlets and there was no marked difference in their size
or abundance. This corresponds to the results of Snow, except
that in her observations she noted a slightly longer growth in the

WALDRON— THE PEANUT 323

dark. This was possibly due to a greater drying out in light in
her experiments. No normally produced tip hairs were observed.

Temperature. Plants with roots exposed were kept at 90° F.
to 100° F. in a moist chamber. Others were kept at 60° F. to 70° F.,
the other conditions being the same. Rosette hairs appeared on
all the plants, but at the higher temperature they were much more
luxuriant and abundant. Tip hairs were found near the tips of
several roots on two of the plants growing at the high temperature.
These two plants producing both types of hairs were the most vig-
orous of the set of twenty in the experiment. These results do not
correspond with those of Snow, where a high temperature and hu-
midity retarded their growth. This can possibly be explained by
the fact that the peanut requires a higher temperature for optimum
growth than those plants with which she experimented.

Soil. Plants in loose sandy soil composed of one-half light loam
and one-half sand grew vigorously. After germination they pro-
duced numerous long and almost pure white roots, a few of which,
after reaching the side of the pot or box, bore a limited number of
tip hairs. No rosette hairs appeared until the plants were one to
two weeks old and quite well established. Tip hairs appeared on
one plant only, of those that were older than three weeks. Plants
in light loam, without any mixture of sand, grew slowly and the
rosette hairs were the first and only type observed, after the side
roots had developed. They appeared on all the plants examined
after from one to two weeks' growth. Tip hairs were observed on
a two months' old plant which had been retarded in its growth,
and, when repotted, a few delicate roots appeared which bore a
very few scattered tip hairs.* Rosette hairs could be found on any
plant of any age, except the young vigorous specimens of less than
two or three weeks' growth.

On seedlings of various ages no hairs of either sort were ever ob-
served on the primary root. Many seeds germinated in sandy soil,
and on sterilized wet cotton in test tubes, did not produce them even
under optimum conditions of heat, air and moisture. With sufficient
moisture, however, rosettes always appeared on the bases of the
side roots. Those in test tubes grew slowly because of lack of

*0f about fifteen such plants examined this one was the only one in which
they were observed.

324 WALDRON— THE PEANUT

water, and as a result rosette hairs were the only type observed.
These were absent on the roots furthest from the moist cotton.

Concerning the function of root hairs on the radicles of seedlings,
Haberlandt and others state that one reason for their early forma-
tion on these, is, that the main root may have sufficient anchorage
in order to rapidly penetrate the soil for its immediate needs. Evi-
dence presented by the peanut thus entirely contradicts such a
view, indicating that they are not necessary for this purpose. Such
seeds as those of the peanut can alone supply adequate food material
for germination if water is present. The radicle elongates and pene-
trates a light soil with great rapidity without them. This may be
one reason why a light sandy soil is best for this species. This
thought also suggests an interesting relation to the hypogeal fruit
production. The plant always maturing its seed under ground
would not need such anchorage for its first root growth,
even though it were a desirable feature. The fruit wall acts also
as an aid. The writer feels, however, that the reason why hairs
are not present here is that the outer layers, in stripping off (see
page 314), do not allow their development. The fact that side
roots, which do not show this peeling, will form them under proper
stimuli is evidence of this. Even these do not often produce them,
probably because of their very loose structure (see page 314). The
cells which are best adapted to respond are at the base of side roots,
thus rosettes are formed.

To summarize these observations, the results of the root hair
experiments on the peanut indicate that (i) light and darkness have
no effect on their production; (2) high temperature, with sufficient
moisture and air, accelerate the growth and production of the rosette
type; (3) loose sandy soil, with root aeration, stimulates hair growth
on the tips of young plants, and possibly also on old plants, if a
period of retardation is followed by a suddenly renewed root vigor.
The rosettes of hairs may appear on any plant of more than one to
two weeks' growth. The tip hairs are found only on young roots
that are undergoing a vigorous elongation in a natural or artificial
moist air space at a high temperature. Low temperature, lack of
oxygen and wet heavy soil prevent the normal tip type from appear-
ing and retard the rosette form. None appear on the radicle. The
cause for the production of the rosette hairs at and on the base of
the side roots is hard to explain. It may be that, at the time the
side root is penetrating the cortex and epidermis of the main root.

WALDRON— THE PEANUT 325

the root tip tissue here is more active and vitalized than later, and
so is more sensitive to external stimuli. It was noted that when tip
hairs were produced usually no rosettes were present. Possibly
these offset the need for tip hairs, or at least utilize energy which is
lacking for their later formation.

Absorption

From the foregong observations on the roots and root hairs of
the peanut, it is evident that the young plant absorbs its water and
mineral foods by any one, two, or all of the following means, depend-
ing on environment and growth conditions: (i) Through the epi-
dermis of young roots the thin cuticle of which is mucilaginized; (2)
By means of normal tip hairs on vigorous growing roots; (3) By
means of the rosetted, basal hairs. The first and last are undoubtedly
the most important of these. As soon as the fruit stalks appear and
reach the soil, hairs are at once formed from the epidermal cells of
these. The older plants then have a much more extensive absorbing
surface from the formation of a considerable number ot gynophores.

Proof that these hairs do supply water to the plant is indicated
by the fact that, according to Pettit, when the roots of such are
severed the plant continues active and apparently uninjured for
some time. The xylem of the gynophore bundles, although not very
large, is sufficient to carry a considerable quantity of material. The
nearly mature fruit must also absorb some water as indicated by
the presence of delicate absorbing hairs. How important this is it
is difficult to say. That they are not absolutely essential is evident
from the fact that the fruit continues to swell somewhat if trans-
ferred from the soil to the air, causing the hairs to dry up. The
absorbing fruit stalk undoubtedly takes the place of roots to a cer-
tain extent. Root tubercles, which also appear at about the same
time, should be noted too in this relation, since some other nodule-
producing members of the Leguminosae, as is well known, seem to
have a reduction of root hairs. The absorbing surface of roots alone
on these older plants is comparatively small.

Development

Germination. The writer, in attempting to raise plants for study
in the greenhouse, had some difficulty in keeping insects and mice
away. Some of these seemed to have no difficulty in locating the
seed even before germination. A wire cage was finally built which

326 WALDRON— THE PEANUT

controlled these pests. Further trouble was experienced, however,
unless great care was used in planting and watering. It was found
that unless they were brought forward in loose material like sphag-
num or well aerated sandy soil, they rotted in one to two days. If
pure sand were used, they would rot if kept even moderately wet.
The method finally used which succeeded was to plant several to-
gether, allowing a considerable degree of aeration. When these
produced an inch or two of radicle, they were separated and trans-
ferred to individual pots with the cotyledons half exposed. Seeds
planted in the shell succeeded well, as this seemed to allow also for
free aeration. The ease with which the seeds rot is likely due to its
weak protection by the testa. This thin papery coat is easily rup-
tured and the embryo, rich in food material, seems to be very sus-
ceptible to infection by molds and decay-bacteria. If growth is
rapid, however, the increased oxidation gives it vitality to resist.
Those who raise peanuts say that good drainage in a loose soil is
absolutely essential for success. The plant must start quickly and
be kept growing. If planted in the uninjured shell, which is sterile
within, there is less likelihood of infection before it gets well under
way. A comparison was made by Bennett^^ of growing peanuts
from shelled nuts, nuts broken into two parts, dry and unshelled
nuts, and unshelled nuts which had been soaked in water for 12
hours and buried in the earth below the frost line for different per-
iods. The most perfect stand was obtained from nuts planted in
broken pods. The results seemed to indicate that when nuts had
been thoroughly wet and moist for a short time they would produce
a good stand, and save the expense of shelling. This corroborates
the author's thought that the seed benefits by free oxygen and pro-
tection furnished by the shell in order to start its growth success-
fully, at least in anything but an extremely loose soil.

Later Growth. After the radicle has reached two or three inches
the cotyledons are pushed about 2 cm. into the air by the elongat-
ing hypocotyl. The hypocotyl often becomes thick and fleshy in
its cortex. This is more marked when growth is retarded from some
cause, and then the lower end becomes tuberous from a deposition
of sugar. The roots of such are not able to utilize the food as fast
as it is supplied from the seed. When the soil is light and the tem-
perature optimum, the formation of an extensive root system re-
sults (see page 323). Under these conditions while the food of the
cotyledons is still available, the plant grows rapidly for about

WALDRON— THE PEANUT 327

two weeks, followed by a period of very moderate development.
Audouard"^ states that the plant grows slowly during the first half
of its existence, and that the most rapid growth takes place after
about ten weeks. This indicates a relation to the formation of the
gynophore and root tubercles again, both of which appear later.
This makes possible a greater activity in growth of the plant. One
feature somewhat difficult to understand is the presence of the num-
erous stomata on both epidermal layers. How is the water balance
kept with such a reduced hair surface, especially on the roots of
plants not yet producing fruit?

If the seed is deep, the hypocotyl elongates accordingly. If the
seed is planted in the shell, it will push up through three or four
inches of soil. It is often much curved and twisted in its efforts
to extricate the cotyledons from the shell. These remain green for
two or three weeks, when they wither and drop. Concerning the
presence of food materials during growth Audouard states (i) that
there is sugar in all parts of the plant, which decreases in amount
during fruit maturation; (2) That starch in the root and stem in-
creases from the beginning to the end of vegetation; (3) that fats
increase for six to nine weeks, that is, until the fruiting period, when
they suddenly decrease in the vegetative organs; (4) that proteins
decrease in roots and stems at flowering time and increase in the
fruit. The writer has observed that the gynophore is well stored
with starch until the ovary begins to grow, when apparently much of
it is carried as sugar to the inner fruit tissue, forming there the
broad, delicate-walled sugary endocarp (see page 319), — thus the
reason why immature fruits are sweeter. Some of this sugar at
least is apparently gradually transferred to the testa and there
stored temporarily as starch. Later, both this and that from other
sources (gynophore and stem) are transferred to the cotyledons and
largely stored as oil. Since the carbohydrates are early furnished
and carried to the gynophore and stem, the thought arises as to the
possibility of the fruit maturation being largely independent of the
roots of the plant. The other food materials can be obtained from
the soil by the fruit, and proteins can be formed in darkness, so there
is no known reason why this should not occur.

Biological Considerations Concerning the Fruit and Gynophore.

Observations on the Gynophore. In his work on "The Movements
of Plants,'" Darwin"^ says that while apheliotropism may act in

328 WALDRON— THE PEANUT

some slight measure on the downward growth of this organ,geotrop-
ism is unquestionably the exciting cause. The writer proved this
by inverting two plants which had produced several gynophores
whose tips were about to pierce the soil. As seen in Fig. ii the
tips turned away and became reversed in position. This not only
proved the effect of gravity, but also that the hydrotropic reaction
of the organ was weak or lacking. They acted in a similar way even
if the soil was saturated. When tips were allowed to penetrate
wet sphagnum, and then the plants reversed, they would recurve
downward and grow out of it. When the plants were righted again
the tips also turned back thus forming an S curve (Plate LXXX,Figs.
II and 12). The presence of definite granules in the lumen of each
of the epidermal cells of the gynophore (see page 318) at the tip, and
their absence anywhere else, suggests the possibility of such being
the structures by which this organ perceives when it is out of line
with gravity. This has been discussed by others in connection with
the presence of starch grains in root tips. The writer found that
by cutting off the tip of the gynophore, growth continued, but there
was no reaction to gravity when the plant was inverted. The ap-
parent homology between the behavior of the root apex and of the
gynophore apex is highly suggestive,

Darwin refers to the means by which this organ penetrates through
the soil. He says, "the sharp smooth point of the gynophore en-
ables it to penetrate the ground by mere force of growth, but its
action is aided by a circumnutating movement." The anatomy of the
organ is also suggestive. The patches of hard bast give strength,
while their separation, even though in a close ring, gives pliability.
Pettit states that the hairs produced at the tip are also an aid in
holding it firmly. Although this happens to be of some assistance,
the writer would question to what extent, since the relation of the
radicle to soil penetration puts a new light on this matter. Hair
experiments, similar to those of Pettit, were made with plants bear-
ing young gynophores, which had not yet reached the soil. Pettit
found that by putting these in a moist chamber a narrow zone,
averaging 3 mm. in length, always appears one to eight millimeters
from the tip. The writer found by repeated experiments that this
zone might be as much as 5 cm in length (Plate LXXX, Fig. 13).

In discussing these in relation to those of roots, Pettit says: "In
comparing the growth of gynophore hairs with that of root hairs it
must be remembered that the growing point of the gynophore cor-

WALDRON— THE PEANUT 329

responding to the punctum vegetationis of the root lies just below
the ovary which occupies the extreme tip of this organ. The ovary,
however, is almost microscopically small and remains so during the
growth of the gynophore. To illustrate the extremely small space
occupied by it, the hairs which were not more than one millimeter
from the tips of the gynophores as mentioned above were still below
the growing point under the ovary. While this difference in the
position of the growing point exists between root and gynophore,
the difference which it makes in estimating the relative distances of
the hairs from the tips is practically nothing.

"The resemblance between these hairs and those ot roots was fur-
ther tested by repeated experiments in pulling young gynophores
carefully from the soil. The minute portions of earth clung to the
hairs and refused to be separated from them in the same manner as
in the case of root hairs. In several instances these hairs were tested
for acids and were found to respond readily to the litmus paper test.

"Still another experiment was made which furnishes strong evi-
dence that one function of the gynophore hairs corresponds to the
chief function of those of the root. A large, well developed, thriftily
growing plant was cut in such a manner as to separate the whole
root system from the stems, but the latter were still connected with
the ground by numerous well grown gynophores. The result was
that the plant so treated after two weeks still presented nothing to
a superficial inspection to distinguish it from others in its vicinity
whose roots were left intact. Closer examination showed that some
branches were dead; but the majority were putting out new leaves
which appeared quite as strong and healthy as any of those on sim-
ilar plants in the vicinity which were supported by roots. Un-
fortunately these experiments were begun late in the season, and the
appearance of the frost prevented their continuance."

Fruit Maturation. Other writers state that all attempts to make
the gynophore produce aerial fruits by digging away the soil as the
gynophore elongates fail. It is also well known that any such stem
dries up unless it reaches the soil by the time it is two to three inches
long. The length to which it grows before drying varies with the
humidity of the air. Experiments were thus made in an attempt to
determine what caused the ovary to enlarge, and what would pre-
vent it. Plants with gynophores of various lengths were put in a
saturated atmosphere and not allowed to penetrate any substratum.
In all cases the gynophores continued to elongate and become green

330 WALDRON— THE PEANUT

for about a month. They usually attained five or six inches in
length and then wilted just back of the ovary. The longest gyno-
phore produced in this way was seven and one half inches in length
— nearly twice as long as any seen by the writer in the soil. Hairs
developed which gradually died away until there were but a few near
the tip. Other gynophores were allowed to grow in test tubes of
tap water, some kept in the dark, others in the light. These pro-
duced no hairs. Two of those in darkness, after eight weeks, pro-
duced a small one seeded fruit. The remainder produced none.
Others were allowed to grow into sphagnum and pure sand and re-
sulted in fruit formation in both cases. Gynophores which had pen-
etrated soil and whose ovary had begun to swell were exposed to a
saturated atmosphere and to ordinary greenhouse conditions. In
the former case, the fruit turned green and continued to grow
slightly, while the latter turned green and remained small.

The results of these trials, although not at all conclusive as to
evidence offered, indicate that (i) a thigmotropic, hydrotropic or ap-
oheliotropic stimulus, or a combination of these, is necessary for the
ovary to begin maturation, but (2) that a continuation of such is
not necessary, since the ovary continues to develop somewhat if
removed from the soil after its growth has begun. This develop-
ment, however, is somewhat abnormal. All successful experiments
were produced in complete or partial darkness, although those
in the moist dark chambers failed. This suggests the necessity
of the first two factors, i. e., water and contact. More research is
necessary in order to clinch this point and determine which of the
three is the most important.

From the writer's observations and from the varying results of
the above attempts at fruit production, he feels that two things
should be kept in mind — (i) the condition and activity of plant
growth, (2) the ability of the plant to possibly supply the necessary
substances to the fruit in two ways — (a) by direct absorption of
some of them from the soil with the carbohydrate supply from the
plant; (b) by transfer of all necessary materials from the vegetative
organs into the fruit. A number of plants that became pot-bound,
when several gynophores were being formed on them produced only
one or two small fruits. Similar plants that had abundant pot-
room produced several. Weak plants may have been experimented
with. In the field, many poor fruits without seeds called "pops"
are found — due possibly in part to this same cause.

WALDROX— THE PEANUT 331

The chlorophyll formed in the fruits exposed to light was found
to be exterior to and around the bundles that form the pericarp re-
ticulations. The testa and cotyledons also became green. This is
an interesting point since it indicates the retention through long
millenia of the factors necessary for chlorophyll formation.

Watt has observed in India that red ants are frequently found
working harmlessly around growing fruits in the soil. This would
be of benefit to the fruit since aeration of the surrounding soil would
allow for greater absorptive hair development. What benefit the
ants might obtain is hard to tell.

Fruit Hairs. It should be kept in mind that the fruit hairs are
different in origin from any found elsewhere on the plant (see page 3 19).
Points of evidence indicating this are — (i) that the epidermal and
subjacent layer of cells is seen to be thrown off (see pagejig and Fig.
6). (2) That the hairs are different from any found elsewhere
on the plant, all of which are truly epidermal. (3) That this differ-
ence, that is the bifurcations of the hairs, indicates the irregularity
of growth of the mesophyll of leaves. (4) That no hairs appear on
the fruit until it is well grown and the two outer layers have been
discarded.

Other queries raised here are: why is the second layer of cells
discarded as well as the epidermis? and why doesn't the periderm-
like tissue begin its formation by divisions taking place in this layer,
as is the case with the hypogeal gynophore, instead of in a deeper
layer? One possible answer is, that this corresponds to that layer
of water storage cells next to the lower epidermis of the leaves of
Arachis. This layer still persists in the carpel, and, being large-
celled and less able to divide, is thrown off with the epidermis.

Conclusions from Physiological Studies

It remains to be considered how far the facts ascertained in this
study contribute to the knowledge of hypogeal fruit production.
The fact that the fruit of so many plants of varied families seeks the
ground must be regarded as significant. Tschirch,^^ in a paper on
Leguminosae, says that one group of nitrogenous compounds pro-
duced by the Leguminosae can be formed only in darkness, and sug-
gests this reason for such a habit. It has also been suggested that
the fruit is thus protected from animals.

Concerning the present studies, the following new facts stand
out quite prominently — (i) The tendency to fruit formation in

332 WALDRON— THE PEANUT

moisture and darkness. (2) The formation of periderm-like tissue
on the hypogeal gynophore and fruit. (3) The formation of hairs
on the gynophore-epidermis and pseudoepidermis of the fruit. That
water is absorbed by these hairs is undoubtedly true. The most
puzzling new structural feature is the second, where thin-walled cells,
not suberized, are laid down by a late-formed cambium layer. Such
is rare in leaf tissue. The bud scales of Horse Chestnut and other
trees produce a limited amount of suberized tissue in this way. Two
to three layers are always produced on the gynophore when this
organ is subjected to a saturated atmosphere. This suggests a
possible water storage tissue, or it may be a result of pressure from
within when more water is absorbed. Cells are possibly necessarily
cut off to allow for this expansion.

Finally, it is suggested that if the plant does not require more
nourishment from the soil than might be supplied by root hairs, and
yet forms such hairs on the gynophore and nearly mature fruit, it
may be more advantageous to take some of its food by this special
method. Perhaps certain desirable changes are made possible by
such foods always being in darkness. Is it then darkness, or extra
water supply that the fruit seeks? Whatever the reason, the re-
sulting advantage is full maturation and selective survival of the
seed, which is highly concentrated in its food constituents.

Uses of the Peanut

To most people outside the peanut growing sections of the coun-
try, the peanut suggests only an unessential food article, — a delicacy
to many, — in the form of the roasted or salted nut, peanut con-
fectionery or peanut butter. During recent years, however, and
especially in the past year or two it has become of utmost importance
as a staple article of diet and otherwise. In the cotton growing
states it is saving the day for many farmers who have failed with
cotton growing because of the new insect pests or other reasons.

Uses as Human Food. The following from Beattie^^ in this con-
nection is worth quoting: "The use of the peanut for eating from
the shell is most important and popular, but the quantity of shelled
peas that are first roasted and salted and sold by the pound is con-
stantly increasing. Some of the better grades are first shelled, then
roasted after which the halves are broken apart and the germ re-
moved giving the meats a blanched appearance rendering them very
desirable for table use. Great quantities of shelled peas are used

WALDRON— THE PE.ANUT 333

every year in the manufacture of peanut candies and brittle, both
alone and in combination with other nuts, pop corn or puffed rice."
During recent years great quanxj,pes of shelled peanuts, especially
of the Spanish variety, have been employed for the manufacture of
peanut butter. It is used in the preparation of vegetarian meats
after a portion of the oil has been pressed from the nuts. This
extra oil and that pressed from nuts grown for the purpose is used
in thinning peanut butter and is as good for every purpose as is
that of the olive. It is one of the sweetest vegetable oils. Articles
fried in it keep well lor a longer time than in olive oil and have an
agreeable odor and flavor. It is mixed with cotton-seed oil to im-
prove it for salad purposes.

Peanut meal or flour of finely ground nuts is used in confections,
cakes and bread making. It is used as a substitute for rice and other
flours. Watt reports that in India the unripe nuts are sweeter (as
indicated in the author's discussion on the physiology of the fruit)
and, being more easily digested, are given women whose milk supply
is insufficient for their children. These unripe fruits, when fresh,
make an agreeable boiled dish. The very tender leaves of the plant
are sometimes cooked with ground coconut-
Concerning the food value and change of the peanut from the
category of a luxury to that of a more staple item of diet for man,
the following is taken from "The Literary Digest" for April 13, 191 8.
"The peanut enters into the preparation of most of the vegetable
'meat substitutes' long warmly advocated by the vegetarians and
now made more conspicuous by the governmental admonition to
'eat less meat;' and peanut 'butters' or 'pastes' are widely used. To-
day the value of the peanut crop, which is divided between the pro-
duction of the promising peanut-oil, peanut-cake for animal fodder
and roasted peanuts for human food, has begun to total many mil-
lions of dollars. At the University of Wisconsin, Daniels and Lough-
lin have demonstrated by feeding-experiments on animals that the
peanut can supply adequate protein ... in sufficient pro-
portions for growth and reproduction. It can also furnish an abund-
ance of the water-soluble vitamin. The food as used in the human
dietary does not, however, yield the growth-promoting fat-soluble
vitamin, which has come to be recognized as a remarkable consti-
tuent of butter fat and egg fat; nor are the inorganic constituents
adequate in quality to supply sufficient calcium and certain elements.
Of course, the peanut is not used as a sole source of nutrients for

334 WALDRON— THE PEANUT

man; nevertheless, the delineation of its physiologic value enables
one to define more intelligently the place which it can take in the
ration. Daniels and Loughlin foresee an increasing usefulness for
the peanut, now that its real value has been scientifically established.
When we consider the broad areas, they say, which may be adapted
for growing the crop, and the fact that our food supply tends toward
a wider use of the seeds of plants, it seems appropriate to expect
that the peanut, when rightly supplemented, will form a staple ar-
ticle of the human dietary. Like the soy-bean, which has lately
come into prominence in American homes, the peanut needs only
to have added suitable inorganic salts and the fat-soluble accessory
to make it a complete food."

Uses as Food for Live-Stock. Beattie is quoted in this connection
as follows: "In the factories where peanuts are cleaned, shelled,
and graded for the market there is always a certain percentage of
cleanings and inferior stock that can readily be turned into stock
foods. The outside shell, or hull, of the peanut, is rich in food ma-
terials, but is extremely diflicult to reduce to a condition in which
it can be fed. In large cleaning factories the shells are generally
used as fuel, and the ash resulting therefrom is valuable as a fertil-
izer, often containing as high as 3 per cent of phosphoric acid, 9
per cent of potash and 6 per cent of lime.

"The thin brown covering of the peas has a feeding value almost
equal to that of wheat bran. These hulls are especially desirable
for mixing with the smaller particles of broken peas for stock feed-
ing. In large factories where peanuts are prepared for the manu-
facture of peanut butter and similar preparations the waste in the
form of small particles of the meats and the germs is considerable
and this is sold to farmers for feeding purposes. In some cases the
waste is mixed with a portion of the hulls and finely ground or chop-
ped before leaving the factory. Peanut hulls make an excellent
bedding for use in stables, and by using them in this manner and
hauling the manure upon the land their full value can be obtained,

" Broken peas and germs are used largely as a food for hogs, but
both should be fed in moderation and in combination with some
grain, as the peanut fed by itself will produce a hog having soft fat
and inferior meat. The famous Smithfield hams and bacon come
from hogs that are fed partly on peanuts, the practice being to turn
the hogs into the peanut fields after the crop has been gathered and
allow them to glean the pods that were lost in harvesting. The

WALDRON— THE PEANUT 335

principal objection to the use of peanut by-products as stock feed
is their tendency to become rancid very quickly. The germs, which
are hicih in nitrogen content, become rancid and bitter in a short
while and should not be kept on hand for a greater period than fifty
or sixty days."

Peanut cake is a stock feed composed of the remains of seeds when
expressed for oil and is extremely rich. As hay, peanut tops are
worth just as much as alfalfa, pound for pound. Even the entire
plant is used, and often chopped fine for this purpose. It forms
a well balanced ration for dairy cows.

Use as a Soil Renovator. Here again the peanut is rapidly be-
coming a crop of much importance. Peanuts are valuable as a
substitute for cowpeas, especially in certain soils that are not adapted
to the growing of the cowpea. In many sections where the clovers
and other soil-renovating crops will not withstand the heat and
drought of the summer months the peanut will thrive and make an
excellent growth. A crop of peanuts for forage can often be grown
after the removal of oats or some other spring crop, and although
they may be badly overgrown by crab-grass, the tops may be mown
with the grass for hay, and the hogs turned in to root out the peas.

Miscellaneous Uses. The oil, beside its use as a food, is valuable
"in soap making, in lubrication and for illumination in some countries.
The shell is often ground into a fine powder for polishing tin plate.
It is said that tin plate manufacturers cannot get enough since this
and middlings are the only two things that will put that mirror-
like polish on tinware and not leave a scratch on the surface.

Summary of Results

The results of these investigations concerning the histology and
physiology of Arachis present marked features which are summariz-
ed as follows:

I. It was found that root hairs were present on the plant, although
reported as absent by two previous workers. These were usually
arranged in rosettes at and on the base of side roots. Their growth
is stimulated by a high temperature and humidity. The normally
produced tip hairs appeared on very young plants whose roots grew
rapidly and were exposed to moist air conditions. Later they never
formed unless the plant showed a sudden renewal of growth vigor.
Saturated and heavy soil conditions retarded the growth of the
rosette type and inhibited the appearance of the tip hairs.

336 WALDRON— THE PEANUT

2. The hypocotyl shows a tendency to enlarge and become tuber-
ous unless growth conditions are ideal. This is apparently due to
a deposition of sugar from the stored food of the cotyledon which
is unable to be cared for by the root as fast as it is supplied.

3. The stem is quite normal. Its epidermis, however, has crystal
cells in groups of two to four. The pith breaks down causing the
stem to become more or less hollow.

4. The leaf, with numerous stomata on both the upper and lower
surfaces, has also, in both epidermal layers, small cells with a single
contained crystal in each when young. Later these cells become
fused into an "epidermal vessel," containing two to thirty crystals
arranged in irregular rows or groups.

5. The fruit stalks or gynophores have been shown to be geotropic
in reaction, and the epidermal cells of the carpellary tips are marked-
ly granular, suggesting a possible perceptive relation in this regard.
These organs are very weakly hydrotropic and do not react to light
or darkness. The epidermis of the epigeal portion has crystal cells
like those of the stem. The epidermis of the hypogeal part becomes
elongated in the cell walls to form absorptive hairs. The second
layer becomes cambioid forming a phellogen-like hypodermis, and
this, by cambioid activity, may divide into two or three layers. The
bundles are separate and highly lignified in the outer phloem and in
the xylem. This gives mechanical strength for soil penetration.

6. The young fruit, as it begins to swell, bursts the epidermal and
subjacent layers of the ovary, throwing them off. The next or third
layer, now the pseudo-epidermis, forms irregular, more or less branch-
ing absorptive pseudo-hairs. These are different in nature from
any of the other hairs formed on the plant, and are indicative of the
irregular growth of the spongy mesophyll of leaves. Beneath this
layer are developed several zones of cells similar to, and continuous
with, the pseudo-periderm of the gynophore. This is a marked
peculiarity in leaf tissue formation.

7. Attempts to produce peanuts in the air by various means
failed to give definite results. Two succeeded by allowing the
gynophore to grow into water; several when grown on sphagnum
and also pure sand. None succeeded where the ovary was exposed
to light. The results indicate that water is an important factor,
but that contact, or darkness, or both may also be necessary. Young
fruits, if previously in contact with soil, and then exposed to the air,
continued to develop to a certain extent and turned green. The

WALDRON— THE PEANUT 337

re-formation and presence of chlorophyll in these indicate the re-
tention through long periods of time, of factors for its formation in
such a leaf structure.

8. The possible benefits derived from underground fruit formation
may be (a) protection from grazing animals, (b) formation of cer-
tain proteins possible only in darkness, (c) more rapid and greater
development in fruit size and number.

LITERATURE CITED

1. Acosta. Hist. Nat. Ind., 1598.

2. Monardes. Simpl. Med. ex Occidentali India, 1580.

3. Marcgraf and Piso. Brazil, edit., 1648.

4. Dubard. Une etude sur I'origine de I'arachide in Bull. Museum d'histoire nat-

urelle, n. 5, 1905.

5. Parkinson. Theatrum Botanicum, 1640.

6. Linnaeus. Species Plantarum.

7. Watt. Agl. Ledger, India No. 15, 1895.

8. Forskal. Fl. Aegypt, 1775.

9. Delile. Fl. d.Egypt, 1824.

10. De Candolle. Origin of Cult. Plants, 1885.

11. Bretschneider. Study and Value of Chinese Botanical Works, 18.

12. Sloane. Jamaica, 1696.

13. Rumphius. Herb. Amb., V, 1755.

14. Bentham. In Flora Braziliensis, Papil. 1859.

15. Petti t. Arachis hypogaea, Nfem. Torrey Bot. Club, 4, 1895.

16. Winton. Anatomy of the Peanut, Conn. Rept. pt. 2, 1904.

17. Adam. L'Arachide, 1908.

18. Richter. Beitrage zur Biologie der Arachis hypogaea, Inaug. Diss. 1899.

19. Solereder. Anat. of Dicotyledons, V. i, 1908.

20. Strasburger. Textbook of Botany, 1898.

21. Jost. Plant Physiology, 1907

22. Haberlandt. Physiological Plant Anatomy, 1914.

23. Pfeffer. Plant Physiology, 1900.

24. Snow. Dev. of Root Hairs. Bot. Gaz. 40, 1905.

25. Bennett. Planting Peanuts. Ark. Bull. 58, 1899.

26. Audouard. Development de I'arachide. Comp. Rend. 117, 5, 1893.

27. Darwin. The Movements of Plants, 1865.

28. Tschirch. Berichte der Deutchen Bot. Ges. 1887.

29. Beattie. The Peanut, Farmer's Bull 431, U. S. D. A. 1917-

Some Recent Agricultural Publications Pertaining to the Peanut

Batten. Peanut Culture. Va. Agr. Exp. Sta. Bull. 218, 191 8.

Spencer and Jenkins. Peanuts for Oil Production, U. of Fla., Bull. 12, 1918

338 WALDRON— THE PEANUT

Thompson. Peanut Growing in the Cotton Belt, S. R. S. Doc. 45, U. S. D. A.,

Ext. S., 1917.
Duggar, Cauthen, Williamson, Sellers. Peanuts, Ala. Agri. Exp. Sta. Bull.

193, 1917-
Spencer and Brown. Peanuts in Florida. Univ. of Fla. Bull. 6, 1916.
Wolf. Leaf Spot and Some Fruit Rots of Peanut. Ala. Agr. Exp. Sta., Bull.

180, 1914.
Kilgore and Brown. Peanut Culture, U. Car. Dept. Agr., Vol. 30, Bull. 3, 1909.
Jones. The Peanut Plant, 1907.

EXPLANATION OF PLATES
Plate LXXIX.

Fig. 4. Root hairs (rosette type) as they appear on and at the base of side roots.
Fig. 5. A portion of the upper epidermis of a leaf showing numerous stomata and

crystals in "epidermal vessels."
Fig. 6. A section through the outer layers of a peanut fruit which was about three-
fourths grown.

a — A portion of the epidermis and subjacent layer of the ovary

still adhering.
b — More or less branched fruit hairs developing from the former

ovarian third layer, now exposed,
c — Hypodermal tissue.
Fig. 7. A section through the outer layers of the hypogeal part of a mature gyno-
phore showing:
a — Hairs formed from the epidermal layers.
b — ^Two layers of the periderm-like tissue (hypoderm).
Fig. 8. A longitudinal section through a young ovary just before the blossoming,
a— Base of style,
b — Pollen chamber.

c — Enlarged and indurated cells which become the apex in Fig. 9.
d — Ovules in ovarian cavity,
e — Fibro-vascular bundles.
Fig. 9. A longitudinal section through a gynophore tip, the ovary of which is three
weeks older than that of Fig. 5, and drawn to the same scale. Note the
very slight difference in size, the style scar at a now pushed to one
side by the enlarged epidermal tip cells which are shown at c in Fig. 8.

Plate LXXX.

Fig. 10. Photo of a Spanish variety peanut plant showing flowers, flower buds and
gynophores clustered together at its base as is typical of the Jastigiata
subspecies.

Fig. II. Photo of an inverted plant showing gynophores recurving in reaction to
gravity.

Fig. 12. Photo of plant which was righted again, shown in Fig. 11. Note the change
of the gynophore tips again.

Fig. 13. Photo of a plant which, after forming gynophores, was put into a moist cham-
ber. Note the profuse development of hairs resulting.

Bol. Contrih. Univ. Penn.

Vol. IV, Plate LXXIX

Waldron on Peanut

Bot. Ccuitrib. Univ. Penn.

Vol. IV, Plate LXXX.

Fig. 13

Waldron on Peanut

THE COMPARATIVE MORPHOLOGY, TAXONOMY AND

DISTRIBUTION OF THE MYRICACEAE OF

THE EASTERN UNITED STATES

BY

Heber Wilkinson Younglcen, A.M., M.S., Ph.D.

With Plates LXXXI-XC

(Thesis presented to the Faculty of the Graduate School
of the University of Pennsylvania, May 8, 1915,
in partial fulfilment of the requirements
for the degree of Doctor of Philosophy).

I. Introduction

The writer, during the past years of his activities as a botanical
teacher, examined minutely numerous drug plants along lines already
mapped out by Tschirch, Oesterle and other leading pharmacognosists.
In continuation of these studies and upon the kind suggestion of Professor
John M. Macfarlane, of the University of Pennsylvania, he undertook
the investigation of those species of the Myricaceae or Bayberry Family
that are indigenous to the eastern United States.

This work was started during the winter of 1912 and has been carried
on continuously up to the present date. During this period extensive
studies have been made by the writer in the field, in the laboratory
and in the herbarium of the University of Pennsylvania, also in the
Academy of Natural Sciences of Philadelphia. The results already
secured are such that the writer trusts to gradually undertake a mono-
graphic account of the family, which presents many features of great
botanical interest.

The writer desires to express his grateful appreciation to Professor
Macfarlane for his many valuable suggestions and untiring aid in the
direction and completion of the work. He would also thank his former
teacher and present esteemed colleague. Professor George H. Meeker, of
the Medico-Chirurgical College, for the opportunities which led to its
undertaking.

340 YOUNGKEN— ON THE MYRICACEAE

Table of Contents

Historical Review ^^

Comparative Morphology of Seedlings 346

Comparative Morphology of Roots 352

Comparative Morpholog>' of Stems 357

Comparative Morphology of Root Tubercles 368

Comparative Morphology of Leaves 375

Comparative Morphologj' of Inflorescences and Flowers 384

Comparative Morphology of Fruits. The development of the fruits of M. cerifera

M. Carolincnsis and M. Macfarlanei 386

Taxonomy 389

Comparative Distribution and Summary 392

Synonyms 395

Supplementary BibUography and List of Illustrations 397

Historical Review

The most ancient document containing any reference to the plants
of the Myricaceae is the Rhya, a Chinese compilation, written during
the Chou dynasty of Tsz Hia, a disciple of Confucius.

To Valerius Cordus (1515-1544), a commentator of Dioscorides, is
attributed the first, mention in Europe of a plant of this group, made
in a simple enumeration of the names of plants in a posthumous publica-
tion of Conrad Gesner in 1561 (Gesner " Catalogus Plantarum, " p. 31).

Lobel reported in his "Elaeagnos" a plant which he described at
length and which is none other than Myrica Gale. In the same work
Valerius Cordus created the name Myrica, but employed it for the desig-
nation of Tamarix, (p. 68, 1. 12-14).

About the same period Chytraeus, enumerating the plants which
grew amongst the heaths of Mecklenburg, cited Myrica Gale, which he
named Teutona Myrtus. The document containing this information
is a piece of poetry entitled "Botanoscopium" recently found and pub-
lished by E. H. L. Krause. (Eine botanische Excursion in d. Rostocker
Heide vor 300 Jahr. in Arch. Ver. de Freunde d. Naturg, in Mecklenburg,
1879, p. 329.)

In 1576 Lobel described and figured for the first time the male and
female plants of Myrica Gale which he named Gagel Germanorum. He
classed this species in the group of trees and shrubs alongside Vaccinium
and Oxycoccos and stated that it was called by the pharmacists Myrtus
Brabantia, the Germans naming it Gagel and the English, Goul or Goele.
He compared its resinous odor with the perfume of clover, indicated its
distribution over heath and forest areas and mentioned that it flourished

OF THE EASTERN UNITED STATES 341

in England in June and July. (Lobel, Stirpium Observationes, 1576,
p. 547.)

In 1586, Dalechamps, in his "Historia generalis Plantarum," p. 110,
described and gave illustrations of the same plant which he called Pseu-
domyrsine. He observed it in swampy soil bearing fruits in July and
August in the neighborhood of Rouen and indicated that it was known
by the name of Royal Pimenta.

In 1616, Rambertus Dodonaeus, in his "Stirpium historia Pemp-
tades, " named it Chamelaeagnus and classified it under the group of
Arbustes non epineux (trees not thorny) between Rhus Cotinus and
Viburnum Lantana. He noticed small drops of resin on the fruit and
stated that it was found in Grande-Bretagne, Brabant, Flanders and
N. W. France along the Loire.

In 1640, Parkinson ("Theatrum botanicum," p. 1451) grouped it
with the Sumacs and named it Myrtus brabantica atit anglica. He ob-
served the plant in Sussex, Hertfordshire and Cornwall, England. To
the synonyms already mentioned he added those of Sweet Gale and
Sweet Willow.

In 1650, Jean Bauhin ("Historia Plantarum," p. 503) wrote that the
people of France called it Gale after the name given it by Dr. Peter
Turnerus.

About the same time, Quatramius observed it in swampy land around
Paris and called it Pigmen, probably after Pigmentarii.

In 1691, Plukenet in his " Phy tographia " illustrated and briefly
described one form, the figure of which, in some respect, resembles the
fructiferous branch of Myrica cerifera. The following is his description
"Myrtus Brabanticae accedens Africana, baccis carens Conifera. Arbor
Conifera odorata foliis Salicis, rigidis, leviter serratis Ray Hist, an Laur,
serrata odora Promont. Bon. Spei Bod a Stapel in Theophr. p. 333, ex
America quoque nobis allata est et a Nostratib; Insulam Bermudensem
vigentibus, Laurus odorata vulgo nominatur, ut ab honestiss viro Jacobo
Harlow didicimus. "

A branch of another plant somewhat resembling Myrica carolinensis
he described as Myrtus Brabanticae similis, Carolinensis, baccifera,
fructu racemoso, sessile, monopyrene, fructum fert veluti Saccharo
candidiss incrustatum, Cujus ramulus cum fructu ut per Microscopium
apparuit adpingitur. Hujus pulchri Sane, perravi ei prorsus novi Ar-
busculi Americani generis, munificentia illustris viri D. Giilielmi Courtene
compos factus sum.

342 YOUNGKEN— ON THE MYRICACEAE

In 1693, Plumier ("Description des Plantes de I'Amerique") called
a species which resembles Myrica cerifera, "Arbor carolinensis or Ligus-
triim americanum lauri folio. "

In 1695, Petiver in his "Musci Petiveriani" mentions two species,
Gale mariana asplenii folio now known as Comptonia asplenifolia, Ait.
and Gale capensis now called Myrica cordifolia L.

In 1696, Plukenet in his "Almagestum Botanicum" (p. 250) refers
to the plant now known by the name, Comptonia asplenifolia, (L.) Ait.
as Myrtiis Brabanticac affinis Americana foliorum laciniis Asplenii modo
divisis, Julifera simul et fructum ferens. "

In 1733, Philip Miller, in Vol. 1, ed. 1 of his Gardener's Dictionary,
placed the plants now known as Myrica Gale L, Myrica cerifera L, and
Myrica Carolinensis Mill, under the genus Gale. He called M. Gale of
Linnaeus the Sweet Gale, Sweet Willow, or Dutch Myrtle and referred to
it as growing upon bogs in several parts of England and casting its leaves
in winter. He designated M. cerifera L as the Candleberry Tree and
M. carolinensis Mill, as the Dwarf broad-leaved Candleberry Tree. Both
of these plants, he said, were evergreen and indigenous to America.

In his Systema Naturae of 1737, Linnaeus placed the genus Myrica
in class X, Dioecia and in section IX, Tetrandria, of his sexual system.
In applying the term Myrica to the plants now recognized as belonging
to the Comptonia and Myrica genera, he revived a name employed a long
time before him by Valerius Cordus for the Tamarix.

In 1753, in his "Species Plantarum" ed. 1, p. 1024, he named the tree
form with lanceolate shaped leaves Myrica cerifera. On different
pages of the same volume he gave the names of Liquidambar peregrina
and Myrica asplenifolia to Plukenet's Myrtiis Brabanticac afinis America-
na fohorum laciniis modo divisis.

In his "Spieces Plantarum" 2nd edition of 1763, he classified his
earUer Myrica asplenifolia under the group Monoecia polyandria and
called the plant Liquidambar asplenifolium, giving its habitat as North
America. In the same volume he classified Myrica Gale and Myrica
cerifera under the group, Dioecia Tetrandria. He described M. Gale
as having a suffruticose stem with leaves lanceolate in outline and having
a sub-serrate margin. He further referred to Peter Kalm's itinerary Vol.
2, p. 312 in which M. cerifera is described as a tree with lanceolate, sub-
serrate leaves.

In 1768, Philip Miller named a plant smaller than M. cerifera and
deciduous in nature, Myrica Carolinensis. (Mill. Gard. Diet. ed. 8 n. 3.)

OF THE EASTERN UNITED STATES 343

In 1769, Laurent de Jussieu placed these plants in his natural family,
Amentiferae, not far from Bctnla and Castanea.

In 1771, Catesby ("Natural History of Carolina, Florida and
the Bahama Islands," Vol. 1, p. 69.) designated as "Narrow leaved
Candleberry Myrtle" what Plukenet in his " Phy tographia " and "Al-
magestum" called Myrtns Brabanticae similis, Carolinensis, baccifera,
fructu racemoso sessili monopyreno. He described the plant as a small
tree or shrub about 12 ft. high with crooked stems branching forth near
the ground irregularly, having long, narrow and sharp pointed leaves,
some trees having most of their leaves serrated, others not. He also
briefly described the condition of the flowers in May, as well as the
appearance of the fruits later and stated a method for making candles
from the wa.x. His illustrations of flowering and fruiting branches re-
semble to a great extent those of Myrica cerifera L.

On page 13 of the same work he discussed another plant which he
called, "Myrtus Brabanticae similis, Carolinensis, humilior, foliis
latioribus et magis serrata, or Broad-leaved Candleberry Myrtle. " He
further mentions that this plant usually grows not above 3 ft. in height,
differing principally from the tall Candleberry Myrtle by its broader
leaves. His illustration resembles the fruiting branch of Myrica Caro-
linensis Mill.

In 1778, Lamarck designated as Myrica palustris, Myrica Gale, of
Linnaeus ("Fl. France" 11 p. 236).

In 1789, Alton, in "Hortus Kewensis" p. oZA divided M. cerifera L.
into two species, viz.: M. cerifera angustifolia and M. cerifera latifolia.
The latter corresponds to Miller's Myrica Carolinensis. He named
Liquidambar asplenifoUnm of Linnaeus, Comptonia asplenifolia.

In 1794, William Bartram, in the second edition of his "Travels"
p. 403, mentioned his excursion to Taensa, Alabama in 1778, where he
observed a plant which he named, Myrica inodora. He described it as
an evergreen shrub, growing in wet sandy soil about the edges of swamps,
rising to the height of 9-10 ft., dividing itself into a multitude of nearly
erect branches which are garnished with deep green, entire, lanceolate
leaves, the branches producing large, round berries covered with a scale
or coat of white wax. He further stated that it possessed no fragrance,
but was in high esteem among the French inhabitants of this region,
who named it the Wax tree on account of its yielding abundant wax for
their candles.

In 1802,Loiseleur-Deslongchamps in "Nouv. Duhamel" 2 p. 190, t55
gave the name of Myrica pensylvanica to the same species which Miller
called Myrica Carolinensis.

344 YOUNGKEN— ON THE MYRICACEAE

In 1803, Michaux ("Fl. Am. Bor. 2 p. 228") described a small leaved
variety of M. cerifera L. which he named var. pumila.

In Miller's "Gardener's Dictionary" appearing in 1807, Vol. 2, part
1, nine species of plants are enumerated under the genus Myrica. Of
those indigenous to the Eastern United States, Myrica cerifera is des-
cribed as rising to a height of 8 to 10 ft. with stiff leaves nearly three
inches long and one inch broad in the middle, smooth, entire, of a yellow-
ish-lucid green on the upper side, but paler on the under, alternate,
subserrate; male and female flowers on different plants; male catkins
about an inch long, erect; female flowers coming out on the side of the
branches in long bunches, and succeeded by small rounded berries covered
with a sort of meal. Myrica CaroUnensis is said not to rise as high as
M. cerifera L. to have branches not so strong and with grayish bark.
Its leaves are said to be shorter, broader and serrate, in other respects
like those of M. cerifera L.

In his "Introduction to Botany" published in 1836, Lindley named
the family Myricaceae.

In 1864, Casmir De Candolle (Prodromus Vol. 16, p. 147) grouped
the genera Myrica and Leitneria under the family Myricaceae and divided
the genus Myrica into four sections, viz.: Gale, Comptonia, Faya and
Subfaya.

Engler in 1894 (Pflanzenfamilien 3 pt. 1, 27) divided the genus Myrica
into three sections, viz.: Morella, Gale and Comptonia. He stated the
characters of each section as follows:

1. Morella: — Flowers dioecious or monoecious, staminate flowers
subtended by 2 to 4 or numerous scale like bractlets or ebracteolate ;
pistillate flowers solitary or in from 2 to 4 flowered clusters, subtended
by two bractlets persisting under the fruit. Pericarp, papillose, covered
with a warty secretion, or rarely succulent and fleshy, leaves serrate or
rarely entire.

2. Gale: — Flowers dioecious; pistillate flowers subtended by two
accrescent bractlets, forming lateral wings on the fruit, pericarp smooth
and resinous, leaves serrate.

3. Comptonia: — Flowers usually monoecious; pistillate flowers sur-
rounded by 8 linear subulate bractlets, accrescent and forming a spiny
involucre in the fruit. Pericarp smooth, resinous and lustrous. Leaves
pinnatifid.

Engler joined the family Myricaceae to the Leiineriaceae to form the
alliance of Myricales which he placed after the Piperales and Salicales
and before the Balanopsidales and Juglandales.

OF THE EASTERN UNITED STATES

345

In 1898, Solereder stated that the family Myricaceae included the
single genus, Myrica. He mentioned the following characters as pecuHar
to the plants of this genus: Large peltate glands; vertical transcurrence
of the smaller veins of the leaf; absence of a characteristic stomatal
apparatus; narrow medullary-rays in the wood; tendency to form scalari-
form perforations in the vessels which never have especially wide lumina;
wood prosenchyme with bordered pits; tendency to form a composite
and continuous ring of sclerenchyme in the pericycle; superficial forma-
tion of cork; oxalate of Ume in clustered and soUtary crystals; unicellular
hairs. (Systematic Anatomy of the Dicotyledons Vol. 2, p. 785.)

The same year, August Chevalier discussed the vegetative organs of
the family. ("L'apparail vegetatif des Myricacees," in C. R. Assoc,
franc, pour I'avanc. d. Sc, Congres Nantes, p. 457.)

By far the greatest work ever done on the Myricaceae was carried out
by August Chevalier from 1897 to 1902, whose observations are recorded
in a monograph appearing in Memoires de la Societe Nationale des
Sciences Naturelles et Mathematiques de Cherbourg, Vol. 32, p. 85-340,
and entitled, "Monographie des Myricacees, Anatomic et Histologie,
Organographie, Classification, et Description des Especes, Distribution
Geographique." Chevalier's monograph comprises two parts. Part
one deals with the general characters of the family. It includes three
chapters in which the vegetative apparatus, the root tubercles, and the
reproductive apparatus are respectively discussed. Part two embraces
the characters of the genera and species, the geographical distribution
of the species, and a s>ti thesis of his results.

In this work he divided the Myricaceae into three genera, viz:—
Gale, Comptonia and Myrica, the differential characters of which I
translate from the French in the following table:

Gale
Leaves thin, caducous,
entire or feebly denta-
ted.

Without stipules
Catkins inserted on the
deciduous branches.
Dioecious plants.

Ordinarily 4 stamens
Very smooth, guarded
by 2 entire bracteoles
forming aerial buoys.

Comptonia
Leaves thin, caducous,
pinnatifid.

With stipules.
Catkins inserted on the
deciduous branches.
Dioecious plants.

Ordinarily 4 stamens.
Ovary smooth, guarded
by 2 laciniate brac-
teoles, provided with
emergences at the base
and developing itself in-
to a cupule.

Myrica
Leaves coriaceous, ordin-
arily persistent, entire,
dentated or rarely mcised.
Without stipules.
Catkins inserted on the
growing branches.
Dioecious or monoecious
plants.

From 4-20 stamens
Ovary covered with wax
bearing or fleshy emer-
gences, either no brac-
teoles or bracteoles non
accrescent.

346 YOUNGKEN— ON THE MYRICACEAE

He remarked that the Myricaceae displayed many characters similar
to several genera of Piperales and Urticales.

Further reference to Chevalier's monograph, as well as to those of
other writers, will be made in the body of this thesis.

Comparative Morphology of Seedlings

The seedlings studied were in part gathered, and in part grown from
seeds planted in the greenhouses of the University of Pennsylvania. On
June 20th, 1913, we gathered 28 seedhngs of M. Carolinensis and 17 of
M. cerifera at Wildwood, N. J. These varied in age from a few months
up to 3 or 4 years of growth. On July 26th, 1914, we collected 15 seed-
lings of M. Carolinensis and 10 of the hybrid M. cerifera x M. Carolinensis,
these ranging from 1 to 4 years of age. On March 21st, 1915, we gathered
15 seedlings of M. hybrid at Palermo, N. J. Seeds of M. Gale, M. cerifera,
M. Carolinensis, and M. cerifera x M. Carolinensis {M. Macfarlanei)
were planted in the University of Pennsylvania greenhouses, but up
to the present writing only those of M. cerifera have germinated. All
of the seedlings collected possessed coralloid clusters of tubercles attached
for the greater part to the rootlets coming off from the primary root, but
occasionally attached directly to the primary root axis. Examination
of 1 1 seedling root systems of M. cerifera grown from seed under artificial
conditions revealed the presence of tubercles on all but one.

Gross Structure of Seedlings

Myrica cerifera, L

(Plate 81. Fig. 1)

Primary root, long, flexuose and tapering with numerous slender
lateral rootlets, many of which are furnished with tubercles.

Hypocotyl, terete, glabrous, red, somewhat curved in its lower portion
and becoming woody very early.

Cotyledons, sessile, glabrous, obovate, opposite, slightly emarginate
at apex, clasping at the base, 5.5 mm. long x 4 mm. wide, with a distinct
mid-rib, venation pinnate-reticulate.

Epicotyl, erect, terete excepting toward the apex where it becomes
triangular, brownish-red to green yellow and red, hairy, glandular, soon
becoming woody; first internode 6 mm. long; second 2 to 2.5 mm.; third
6 mm.

Leaves, simple, cauline or ramal, first and second pairs opposite, the
rest alternate, petiolate, alternately pinnately nerved, the lower ones

OF THE EASTERN UNITED STATES 347

thin hairy on upper surface and margin, the upper devoid of simple hairs
with the exception of the margin and petiole and mid-rib portions which
possess even fewer thin simple hairs than Myrica Carolinensis, yellow
and reddish, glandular on both surfaces, petiole flat on upper surface
and somewhat dilated at its junction with the stem, arrangement 2/5
spiral.

First pair of foliage leaves: Opposite, obovate, cuneate, tricleft,
apex mucronate, base oblique, margin entire in lower 2/3, serrate in
upper 1/3, venation alternately pinnate-reticulate, secondary nerves
coming off at angle of 45°, short petiolate.

Second pair of foliage leaves: Opposite, obovate-cuneate, tricleft,
apex mucronate, with terminal lobe entire or showing 1-2 serrations,
lateral lobes beginning to show 1 or 2 teeth, petiole very slightly longer
than that of first pair.

Third leaf: Similar to second leaf, but petiole sHghtly longer and
lateral lobes distinctly 2 toothed.

As we pass upward along the seedling axis we find the leaf margin
becoming more and more serrated until the highest leaves exhibit serra-
tions in their upper half, rarely in their upper 2/3. Occasionally we have
found one inecjuilateral leaf among the upper group.

Myrica Carolinensis, Miller
(Plate 81. Fig. 2.)

Primary root and hypoctyl similar to those of Myrica cerifera seedling.
Cotyledons, sessile, glabrous, opposite, obtuse and slightly emarginate
at the apex, somewhat clasping at the base, 5 mm. long x 3 mm. wide;
venation pinnate-reticulate with a distinct mid-rib.

Epicotyl, erect, terete, excepting toward apex where it gradually
becomes triangular, brownish-green to green, hairy, yellow glandular,
soon becoming woody; first internode 6 1/2 mm. long, second 3 1/2 mm.;
third 2 mm.

Leaves, simple, cauline or ramal, first pair opposite, the rest alternate,
petiolate, alternately pinnately nerved, slightly thinner than those of
M. cerifera, thin hairy on both upper and lower surface and margin,
yellow glandular on both surfaces, the glands on the lower surface
being more numerous than on the upper, petiole flat on upper surface,
slightly dilated at its insertion on the stem. Arrangement 2/5 spiral.

First pair of foliage leaves: Opposite, obovate, tricleft, apex mucron-
ate, base obHque, margin entire in lower 2/3, serrate in upper 1/3,
venation alternately pinnate-reticulate, secondary veins coming off at
angle of 45°, short petiolate.

348 YOUNGKEN— ON THE MYRICACEAE

Second foliage leaf: Alternate with first pair and third leaf, pinnately
tricleft with terminal lobe entire or showing one serration, lateral lobes
beginning to show two teeth.

Third foliage leaf: Alternate with second and fourth, petiole slightly
longer than second, lateral lobes distinctly two-toothed.

As we pass upward along the seedling axis we find the leaf margin
gradually becoming more and more serrate until the highest leaves are
generally serrate in their upper 1/2. We also find that the number of
glandular hairs on the upper surface gradually diminishes until some
of the highermost inserted leaves are entirely devoid of them on this
area.

Myrica Macfarlanei, (Youngken) : (M. cerifera x M. Carolinensis)

(Plate 81. Fig. 3)

Primary root and hypocofeyl similar to those of M. cerifera and M.
Carolinensis. Epicotyl, erect, terete excepting toward apex, where it
becomes triangular, brownish-green to green, hairy, yellow and red
glandular soon becoming woody and branched; first internode 5 mm. long,
second 3 1/2 mm.; third 1.75 mm.

Leaves simple, cauline or ramal, 1st pair opposite, the rest alternate,
venation pinnate-reticulate, secondary veins coming off from mid-rib
at angle of 45°, about the same texture as those of M. cerifera, petiolate,
thin hairy on margin and upper surface over veins, yellow to red glandu-
lar, but glandular hairs more frequent on lower than upper surface,
and yellow glands alone present on upper surface, petiole fiat on upper
surface and showing numerous hairs, somewhat dilated at its insertion
on the stem. Arrangement 2/5 spiral. While the leaves of the seedling
hybrid remain green during the summer and early autumn, they gradually
become discolored during the late autumn and winter, most of them re-
maining on the stem and branches until the new leaves bud forth in the
spring.

Of all the hybrid seedlings examined at Palermo, March 21st, 1915,
we found most of the leaves present and having a reddish brown color;
a few were only partly discolored.

First pair of foliage leaves: Opposite, obovate-cuneate to rarely
orbicular in outline, tricleft, apex mucronate, base more or less obHque
to seldom rounded, margin entire in lower ^, serrate in upper 3^, vena-
tion alternately pinnate-reticulate, extremely short petiolate.

Second foliage leaf: Alternate with first pair and third leaf, obovate-
cuneate, apex either 3-5 toothed or 3 lobed, each lobe being mucronate

OF THE EASTERN UNITED STATES 349

at its apex, base oblique, margin below apical region entire and hairy,
venation similar to first pair of leaves, somewhat longer petioled than
first pair. Upper surface with simple hairs often losing these in winter,
few or no \-ellow glandular hairs. Lower surface hairy along mid-rib
few or no simple hairs elsewhere. Both yellow and red colored glandular
hairs present on this surface.

Third foliage leaf: Alternate with second and fourth leaf, obovate-
cuneate, tricleft, apex mucronate, lateral lobes bi-dentate or one tooth
appearing below each lobe, larger petioled than second leaf, in other
respects resembling it. The leaves become more and more serrate in
ascending the axis until the higher ones are reached which are serrate
in their upper halves.

Histology of Seedlings
Root
The roots of primary growth of M. cerifera, M. Carolinensis and M.
Macfarlanei are very similar in their minute structure. In transverse
section passing from periphery toward the centre, one observes the follow-
ing structures:

1. An epidermis of brick shaped cells whose outer walls are cutinized
and whose contents are protoplasmic. Numerous thin walled root
hairs arise as outgrowths of this tissue for a considerable distance above
the short root cap.

2. Cortex of about 5 layers of radially arranged rounded cells showing
large intercellular-air-spaces with the exception of the layer just beneath
the epidermis. Calcium oxalate crystals of the rosette type are found
in a few of the cells.

3. Endodermis of a single layer of t>Tpical endodermal cells, the walls
of which become thickened very early and whose contents then consist
largely of a brownish substance which Chevalier calls " lignine-gom-
meuse."

4. Pericambium of 1-2 layers of large thin walled cells rich in proto-
plasm.

5. Radial fibro- vascular bundle which is tetrarch.

6. Pith in centre very small and composed of mostly polygonal cells
with punctations on their radial walls. Between the various pith cells
are very small intercellular-air-spaces.

Epicotyl
In all three of these seedlings a transverse section made midway
between the cotyledons and first pair of foliage leaves shows the following
structural detail:

350 YOUNGKEN— ON THE MYRICACEAE

1. Epidermis of polygonal cells highly cutinized on their outer face
and giving rise to short and long, frequently bent, unicellular trichomes
and short stalked glandular trichomes with balloon-shaped heads (M.
Carolinensis and M. Macfarlanei) and both bowl- and balloon-shaped
heads {M. cerifera).

2. Cortex of 6-8 layers of radially arranged polygonal cells, larger in
the meso- and endo- than in the exocortex region. Many of these cells
contain tannin, some starch grains, others rosette and monoclinic crys-
tals, a few gummy lignin.

3. Endodermis not distinguishable from the other cells of the endo-
cortex.

4. Pericambium showing a discontinuous ring of sclerenchyma ele-
ments.

5. Open collateral fib ro- vascular bundles whose phloem region is
very narrow and composed mostly of sieve tubes and phloem cells. The
phloem cells are loaded with tannin.

6. A rectangular pith zone of polyhedral parenchyma cells whose
walls are thickened. Many of these cells contain starch and a few gummy
hgnin.

First pair of fohage leaves of M. cerifera

The upper epidermis is composed of cells with slightly wavy vertical
walls having a thicker cuticle than those of the lower epidermis and
showing no stomata. In surface view these cells have a mean measure-
ment of 33.99 M X 26.22 m.

The cells of the lower epidermis have more prominently wavy vertical
walls with a mean measurement of 36.42 /x X 24.28 fx. The cells over the
veins possess rectilinear walls and are somewhat elongated. Among
the lower epidermal cells are to be noted many stomata which are slightly
elevated and surrounded by 3-8 neighboring cells.

The mesophyll shows dorso-ventral differentiation into palisade and
spongy parenchyma regions. The palisade zone consists of two layers.
The upper one of distinctly columnar shaped cells, the lower of both
columnar and cuboidal cells which for the most part are more loosely
arranged and smaller than those of the upper layer. The spongy paren-
chyma region consists of small rounded to irregular shaped cells which
surround large intercellular air spaces. A few of these cells are filled
with gummy lignin, probably due to the invasion of the endophyte which
will be discussed later.

The mid rib and lateral veins are found coursing through this region.
The mid rib shows 1 or 2 layers of schlerenchyma elements bordering its

OF THE EASTERN UNITED STATES 351

lower face. Many of the mesophyll cells bordering the mid rib and
lateral veins show rosettes of calcium oxalate. Both surfaces show
numerous crypts containing glandular peltate hairs of two kinds, viz.:
golden yellow and reddish. Of these, the golden yellow variety are by
far more numerous. They consist of a pedicel or stalk two layers thick
and 3-4 cells high with a balloon-shaped multi-cellular head containing
oil and resin. The reddish variety are by far fewer and possess a stalk
portion four layers thick and two cells high with a bowl-shaped multi-
cellular head containing a small mass of resin. Simple unicellular tri-
chomes are found on both lower and upper epidermis.

First pair of foliage leaves of M. CaroUnensis
The upper epidermis is composed of cells whose vertical walls appear
more wa\y in surface view than those of M. cerifera. The outer walls
of these cells have a mean measurement of 63.34 fx x 33.46 ix. The cells
directly over the veins show rectilinear walls and are slightly elongated
in the direction of the vein.

The lower epidermis possesses cells whose vertical walls are indeed
more wavy than those of the upper epidermis and whose outer walls
show a mean measurement of 54.87 n x 23.30 n. Numerous stomata are
present among the epidermal cells of this surface. These are surrounded
by 5-8 neighboring cells.

The mesophyll resembles that seen in the first pair of foliage leaves of
M. cerifera with the exception that it is only about half as wide, shows
more loosely arranged paHsade cells, and has fewer scelerenchyma ele-
ments below the mid rib and nerves.

The upper surface shows fewer crypts than that of M. cerifera and
both surfaces are devoid of reddish glandular hairs with bowl-shaped
heads. Both surfaces however possess the golden-yellow glandular
hairs with balloon-shaped heads. The unicellular simple hairs are more
numerous than on either M. cerifera or the hybrid {M. cerifera xM.
CaroUnensis).

First pair of foliage leaves of M. Macfarlanei {M. cerifera x M. Caro-
Unensis
The first pair of foliage leaves of the hybrid shows histological pecu-
liarities which in several respects are intermediate with those of both
of its parents— M. cerifera and M. CaroUnensis. For instance, the cells
of the upper epidermis in surface view have their vertical walls more
wavy than those of M. cerifera, less wavy than those of M. CaroUnensis,
Their outer walls have a mean measurement of 41.27 /x x 26.95 /x, while

352 YOUNGKEN— ON THE MYRICACEAE

those of M. cerifera are 33.99 m x 26.22 n and of M. Carolinensis, 63.34 n x
33.46 At; the mesophyll region is narrower than that of M. cerifera,
broader than that of M. Carolinensis; the lower epidermis has cells
whose outer walls in surface view show a mean measurement of 37.87 /x x
21.85 }x, while those of M. cerifera are 36.42 x 24.28 and M. Carolinensis
54.87 MX 23.30m; fewer unicellular simple trichomes on upper epidermis
than those of M. Carolinensis, more than those of M. cerifera.

They resemble those of M. cerifera in that both red and golden-
yellow headed glandular hairs are present on the lower surface and also
by the fact that their upper epidermis has a cuticle of about equal thick-
ness.

They resemble M. Carolinensis in that they have few crypts on the
upper epidermis containing golden yellow glandular hairs. The mid
rib and stronger veins are completely surrounded by a zone of scleren-
chyme and connected with the upper epidermis by collenchymatic cells.

Comparative Morphology of the Roots

The roots of M. cerifera, M. Carolinensis, M. Macfarlanei, M. Gale
and Comptonia asplenifolia show the general structure of Dicotyledons.
All of the younger ones which the writer has examined possessed a root
cap, above which were noted numerous root hairs with the exception
of those of M. Gale which possessed relatively few.

The subapical region showed three groups of generative tissues which
form the calyptra and epidermis, cortex, and central cylinder.

The radicle grows downward in a rather tortuous manner and forms
the fibrous tap root, which in the above named species is soon exceeded
by the lateral roots borne on the hypocotyl axis.

Primary Structure
The piliferous axis is composed of large clear pavement like cells,
which frequently elongate into root hairs. The cortex is composed
of 5-12 layers of cells, the most external layer of which is deprived of
intercellular air spaces and collenchymatic. The other layers are com-
posed of rounded to polygonal cells arranged in somewhat loose radial
rows extending to the endodermis. The air spaces in M. Gale and M.
Carolinensis growing in humid earth are very large. Some of the cells
of this region contain rosettes of calcium oxalate, others tannin, occa-
sionally gummy Hgnin.

OF THE EASTERN UNITED STATES 353

The endodermis consists of tj'pical endodermal cells. It dies very
earlv and becomes filled up with a brownish substance which Chevaher
calls " ligneux-gommeuse " (gummy lignin).

The pericambium is composed of one or two layers of large meriste-
matic cells which frequently give rise to side rootlets. The radial bun-
dle, according to Chevalier, (7, p. 97) possesses 4-8 tracheal poles, the
number varying in the same species. The writer has found the bundle
to be tetrarch in M. cerifera, M. Carolinensis, M. Macfarlanei and Com-
tonia asplenifolia.

The pith region is composed of polyhedral cells with simple puncta-
tions on their transverse walls, and possessing small air spaces. Transi-
tion : — The external layer of the pericambium very early becomes a cork
cambium laying down cork on its outer face, thus cutting off the nutri-
ent supply and causing the tissues beyond e.g., epidermis, primary
cortex and endodermis to sluff off. On its inner face the cork cambium
cuts off secondary cortex. Secondary meristem develops on the inner
face of the phloem patches forming intrafascicular cambium as well as
on the outer face of the xylem strands forming there interfascicular
cambium and so a continuous cambium zone gradually arises which
cuts off secondary phloem on its outer face, thereby adding meta-phloem
and cuts off secondary xylem on its inner face thus augmenting the
total mass of xylem tissue. Secondary medullary rays are also cut off by
the cambium. The original radial structure of the stele changes to the
open collateral type.

The primary phloem soon becomes inactive; the cells have their walls
thickened and frequently appear coUenchymatic.

Secondary Structure

Myrica cerifera, Linnaeus. (Plate 86. Fig. 16.)
In passing from periphery toward the centre the following pecuHari-
ties are to be observed : —

1. Cork of several layers of irregular brick-shaped cells which vary
in their staining capacity. Many of the cells have highly suberized
walls which stain green with extract of chlorophyll. Some, however,
have all of their walls lignified, while others show lignification only on
their inner walls. The lignification of these was readily determined
through the use of phloroglucin and hydrochloric acid which imparted
the characteristic pink coloration.

2. Phellogen of meristematic cells in a rapid state of division.

354 YOUNGKEN— ON THE MYRICACEAE

3. Secondary cortex whose cells are tangentially elongated and small-
est in the outermost portion, becoming larger as one passes toward the
phloem. Usually three or four of the outer layers of the exocortex
region are devoid of intercellular air spaces. The remaining layers
show numerous air spaces which are for the greater part small and
angular. Most of the cortical parenchyme cells contain spheroidal
starch grains with a central cleft hilum, either simple or 3-compound.
Some, however, contain rosette aggregates and monoclinic prisms of
calcium oxalate while others have their walls thickened and are filled
with a yellowish-brown substance which presents the following reactions:
It is insoluble in cold or boiling water, cold concentrated potash solution,
ammonia, alcohol or xylol. It is soluble in boiling nitric acid, boiling
concentrated potash solution and hot solution of sodium hypochlorite.
This substance has been termed " lignine-gommeuse " by Tison (8) and
Chevalier, (7 p. 133).

Sclerenchyme fibres which are quite narrow are few and usually
isolated singly or in small groups of 2, 3 or rarely 4 amongst the cells
of the cortex.

4. Endodermis whose cells do not differ materially from those of
the adjacent cortex.

5. Fibro-vascular bundles of the open-collateral type, consisting of
phloem, cambium and xylem regions separated by medullary-rays.
The phloem region is comparatively wide. Many of the medullary-rays
in this region are broadened out in fan-shaped fashion. The bast fibres
are arranged singly or in groups forming interrupted circles. They are
more numerous in the metaphloem than in the protophloem region.
Crystal-fibres containing monoclinic prisms of calcium oxalate accom-
pany the bast fibres. Starch grains and tannin are found both in the
phloem cells and the phloem medullary rays. Many of the phloem cells
and air spaces contain monoclinic prisms of calcium oxalate. Fre-
quently several crystals are present in one cell or space.

The xylem is porous and traversed by numerous medullary-rays,
continuous with those of the phloem. The medullary-ray cells of this re-
gion are more radially elongated and thickened than those of the phloem,
but resemble the latter in their content.

The tracheae are more or less polygonal in transverse sections and
show barred septa. Some of them contain Actinomyces which has
found its way into them from the tubercles. This organism is always
associated with debris, presumably due to its eroding action upon the
tracheal walls. Tyloses are frequent. The walls of the tracheae are

OF THE EASTERN UNITED STATES 355

pitted. Solereder errs in stating that the maximum diameter of the
vessels varies between .02 and .05 mm. (9) In this root, the writer has
found it to be .096 mm. The woody fibres are extremely elongated and
taper-ended. Their walls are considerably thickened with deposits of
hgnin and show oblique pores. The wood parenchyma cells are relative-
ly few and contain starch and frequently monoclinic prisms of calcium
oxalate. The primary medullary rays are mostly 1-4 cells, rarely
1-5 cells, broad. The secondary medullary rays are 1-2 cells broad.

Myrica Carolinensis, Miller. (Plate 86. Fig. 14.)

The root of M. Carolinensis resembles that of M. cerifera in respect
to tissue arrangement, character of cork, cells containing " lignine-gom-
meuse" and tracheae, some of which contain Actinomyces. It differs
from M. cerifera root in the following particulars: The cortex shows
larger intercellular air-spaces with more crystals of calcium oxalate.

The sclerenchyme elements are more numerous and usually in
groups of several, forming small islets.

The phloem contains fewer bast fibres and for the most part broader
medullary rays. Many of the latter are 1-5 or 1-6 rows of cells in width.

The tracheae are less numerous while the woody fibres are more abun-
dant. The tracheae are to some extent narrower and show both pitted
and reticulate markings on their walls.

The primary medullary-rays are wider and more numerous. They
range from 1-5 to 1-6 rows of cells broad. The cells of both the primary
and secondary medullary rays are somewhat broader and shorter.

Myrica Macfarlanei, Youngken. {M. cerifera x M. Carolinensis)

The root of this hybrid between M. cerifera and M. Carolinensis
shows characters some of which resemble both parents, some, one parent
and others which are peculiar to itself.

It resembles both parents in respect to its cork, its cells containing
gummy lignin, Actinomyces in ducts, barred septa, and crystal fibres
accompanying the hard bast.

It resembles M. Carolinensis in the number of primary medullary-rays
and in having comparatively broad medullary-ray cells.

It differs from both parents by having larger intercellular air spaces
in the cortex, more crystals of calcium oxalate in both the cortex and
phloem, broader and more numerous sclerenchyme elements in the cortex,
and narrower tracheae. (Plate 86. Fig. 15.)

The primary medullary-rays are 1-6, occasionally 1-7 rows of cells
in width.

356 YOUNGKEN— ON THE MYRICACEAE

Myrica Gale, L. (Plate 86, Fig. 17.)
In passing from periphery toward the centre, the following structures
are to be observed.

1. Cork of several layers of elongate brick-shaped cells which are
unevenly stained and whose walls are partially lignified, in part suberized.

2. Cork cambium of tangentially elongate meristematic cells.

3. Cortex of a few layers of tangentially elongated parenchyma
cells with small angular intercellular air-spaces. Most of the cells of
this region contain starch and tannin, many possess conglomerate and
monoclinic prisms of calcium oxalate, while a few have their cell walls
thickened and contain gummy lignin. No sclerenchyme elements are
present.

4. Endodermis whose cells do not differ materially from those of
the adjacent cortex.

5. Phloem which is comparatively narrow and composed solely of
sieve tubes, companion cells and phloem parenchyme. The primary
medullary-rays in this region are broadened out in fan-shaped manner
and show a width of 1-5 rows of cells.

6. Cambium of thin walled tangentially elongated cells.

7. Xylem which is very compact in nature showing fewer tracheae
than in any other Myrica root found in the Northeastern states. Most
of the tracheae are distinctly polygonal and form interrupted rings in
the spring wood. (Plate 86. Fig. 17.) Their walls show transverse
pits which are usually arranged in longitudinal rows. The greater
part of the xylem consists of tracheids which are square in transverse
section. The primary medullary-rays are 1-4 rows of cells wide and
are always four in number. They are arranged in a cruciform manner.
The secondary medullary-rays are 1-2 rows of cells wide. Both kinds
of medullary-rays contain numerous spheroidal starch grains. Tannin
is found in considerable quantity in the cortex, phloem and medullary
rays.

Comptonia asplenifolia, (L.) Alton. (Plate 86. Fig. 18.)

This root presents the following characteristics: —

A cork of several layers of reddish-brown, brick-shaped cells whose
walls are for the greater part suberized. Beneath the cork is found a
phellogen zone of somewhat tangentially elongated cells which cut off
cork in an uneven manner.

The cortex is comparatively narrow and is composed of tangentially
elongated parenchyma cells, most of which contain starch, some rosettes

OF THE EASTERN UNITED STATES 357

and monoclinic prisms of calcium oxalate, while a very few contain
gummy lignin. Sclerenchyme fibres are rare in this region.

The phloem in relation to the cortex is comparatively broad. It
contains, besides soft bast, numerous islets of bast figures which
are accompanied by crystal fibres. The crystal fibres are com-
posed of rows of superimposed thin walled parenchyme cells, some of
which contain rosettes, others, monoclinic prisms of calcium oxalate.
The primary medullary-rays in this region are numerous and are 1-5
rows of cells wide. They broaden out toward the cortex in fan-shaped
fashion.

The cambium zone is distinctly wavy in character.

The xylem is traversed by numerous primary and secondary medullary
rays which are continuous with those of the phloem region. The cells
of these contain numerous starch grains and considerable tannin. The
tracheae are very numerous, occupying most of the wood area. Their
walls are pitted. No barred septa have been observed. The woody
fibres are generally quite narrow, taper-ended and extremely Hgnified.

The secondary medullary-rays are from 1-2 to 1-3 cells wide, while
the primary rays vary from 1-4 to 1-5 cells in width.

Tannin is present in cortex, phloem and medullary-rays.

Comparative Morphology of Stems

Gross Structure
Myrica cerifera, L.
A shrub or tree from 1 to 12 m. high, erect to ascending in habit and
having a tall crooked trunk 20 to 30 cm. in diameter, and upright or
spreading crooked branches which form a round-topped head. The
recent shoots are greenish to reddish brown, twigy-elongate at the ends
of branches or branchlets and bearing numerous leaf buds above and
flower buds below during the season The upper portions of these
shoots are densely covered with yellowish- and orange-red glands and a
sparse number of whitish non-glandular hairs. The lower portion has
fewer glands and hairs. Many reddish-white to white, raised, oval,
spotlike lenticels are scattered over the surface. These are slightly
elongated longitudinally and show a length of 5 mm. The branchlets
of second year's growth are reddish-brown and show a few glands but no
non-glandular hairs. The lenticels are larger, arranged longitudinally
and vary from .5 to 1.5 mm. in length. The branchlets in their lower

358 YOUNGKEN— ON THE MYRICACEAE

2/3 bear either male or female catkins. The older branches are grayish
in color. The bark of the trunk is about 6 mm. thick and has a smooth
close light gray surface. The plant is frequently a shrub from .5 to 3
m. high sending up numerous crooked branches from near the ground.
The underground branches grow horizontally for long distances through
the soil and give off numerous tufts of suckers which arch upward.

Myrica cerifera var. pumila, Michaux
A low much branched shrub 2-6 dm. tall, having erect or ascending
stems which are frequently tufted.^" The branchlets of first year's
growth are occasionally extremely hairy, frequently polished and shining.
Other features of the branches are similar to those of M. cerifera.

Myrica Carolinensis, Miller

A shrub attaining a height of 2 to 3 m. with crooked stems which
branch frequently. The recent shoots are greenish to dull reddish-brown
and are covered with yellow glands and shaggy white non-glandular
hairs, the former being more numerous toward the summit. The non-
glandular hairs are far more numerous than on similar shoots of M. ceri-
fera. Both leaf and flower buds are borne on these shoots, the former
always above the latter. The twigs are stiff at the ends of the shoots
of last year's growth.

The branchlets of second year's growth are purplish-brown and show
numerous small whitish to grayish lenticels which vary from .5 to 1
mm. in length. These branchlets bear either staminate or pistillate
catkins in their lower 2/3. The fruit bearing axes persist on the stem
of the third year's growth. On January 31, 1915, the writer found them
also present on the stems of the 4th, 5th and 6th year of one plant growing
in the open. The older branches are ash-brown to ash-gray in color,
reticulately wrinkled and show numerous transversely elongated len-
ticels The underground branches form crooked scaly rhizomes which
creep through the soil in a horizontal manner and give off numerous lateral
branches. They send up tufts of spreading suckers at different nodes
which frequently serve to propagate the species. These suckers grow
downward and then upward in arcuate fashion. They are of a whitish
to reddish-white aspect when recently dug up.

Myrica Macfarlanei, Youngken. (M. cerifera x M. Carolinensis)
A shrub rising to a height of 2 to 2.5 m. with very crooked branches,
the younger of which frequently appear stunted in habit. The shoots

OF THE EASTERN UNITED STATES 359

of the first year's growth are somewhat intermediate in nature with
those of M. cerifera and M. Carolinensis. They are of a greenish to
reddish-brown color, more thin hairy than those of M. cerifera, somewhat
less hairy than those of M. Carolinensis, but showing numerous yellow
and a few orange-red glands. Upon them appear the buds of next sea-
son's leaves and flowers. The branchlets of two years' growth are dull
reddish-brown, and devoid of thin hairs but possess a few golden-yellow
glands. Lenticels are present which vary from .5 to .7 mm. in length
and are arranged longitudinally. Alike with similar branchlets of M.
cerifera and M. Carolinensis, staminate or pistillate flowers are borne
on special catkin axes which, in the case of the pistillate, persist for a
long time after the fruits have fallen. The writer observed these also
present on stems of the 3d, 4th and 5th year's growth at Wildwood,
January 31, 1915 (Plate 85, Fig. 13). The older stems are brownish
gray to ash gray in color and have numerous circular to oval somewhat
raised lenticels, arranged both longitudinally and transversely and vary-
ing from .5 to 1.5 mm. in length. The underground stem creeps through
the soil for long distances and gives off numerous branches which are
similar in their habit to those of M. cerifera and M. Carolinensis.

Myrica Gale, L.
A dark looking shrub rising to the height of from .4 to 2 m. and
dividing into several slender branches which in turn divide and redivade
in furchate manner. The recent shoots are of a rich dark purple color,
polished and shining and bear numerous thin white hairs and yellow
glands which are most numerous toward their summit. The lenticels
on these shoots are relatively few, of a white color, and circular m shape
havmg a diameter of . 25 mm. The branchlets of second year's growth are
also dark purple, showing polished and shining circular to oval, whitish
lenticels, which vary in diameter or length from .35 to .5 mm. The older
branchlets and branches are reddish brown in color and usually polished.
They show elevated, transversely-elongated lenticels which vary from
.5 to 2 mm. in length. On the main and lowest aerial stem the outer
bark cracks and rolls horizontally becoming rough, but still is somewhat
shiny. Special branches are developed on the young shoots which bear
the catkins. These die after the descent of the pollen. The rhizome
creeps through the soil for long distances and bears numerous lateral
branches and suckers. The latter are reddish to reddish-purple in color
and bear thick pink to pinkish-white scales.

360 YOUNGKEN— ON THE MYRICACEAE

Myrica inodora, Bartram
A shrub or small tree attaining a height of 6 mm., with a maximum
trunk diameter of nearly 9 cm(lO). The first branches are ascending,
often straight with a white bark. The branchlets of recent growth are
smooth and reddish brown with a few longitudinally elongated lenticels,
5 mm. long. Those of the second year's growth are ash brown in color
and show lenticels 5 to 7 mm. long.

Compfonia asplenifolia, L. Alton

A round-headed low shrub growing to the height of 1 to 1.25 m.
The main aerial stem is ascending in habit and grows to the length of
about 3 dm. It frequently bifurcates into two crooked branches which
bear numerous branchlets toward their summit. The recent shoots
are greenish yellow to reddish brown in color and covered with long
white hairs and minute yellow shining glands which are always more
numerous toward the ends of the branches. Usually one or two thin
special branches are borne at the ends of these which bear catkins. These
die after the descent of the pollen but often remain on the shoots until
the next season, especially in localities where the plant is well protected
from winds. The branchlets of the 2d and 3d year's growth are of a
steel to reddish-brown color, somewhat pohshed and show scattered
transversely-elongated lenticels which vary from .5 to 1 mm. in length.
The older branches become darker in color but retain their reddish aspect.
These are usually quite crooked and show narrower transversely elon-
gated lenticels from 1-3 mm. in length. The main stem shows a dark
and considerably cracked cork. The underground stem is extremely
crooked and creeps through the soil for a long distance giving rise to
numerous branches and suckers at its various nodes. The suckers branch
downward then upward forming inverted arches (Plate 82, Fig. 4). It
possesses relatively few lenticels which are not raised like those of the
aerial stems. Some of the underground stems bear tubercles, others
thin fibrous roots upon which the tubercles are clustered.

Histology

Under this sub-caption the microscopic peculiarities of stems of
primary growth, showing primary structure, will first be considered from
a general standpoint. Those of aerial and subterranean branches of
secondary growth showing secondary structure will be specifically treated
in the following order: (a) M. cerifera, (b) M. Carolinensis, (c) M.
Macfarlanei, (d) M. Gale, (e) Comptonia asplenifolia.

OF THE EASTERN UNITED STATES 361

Primary Striuture

(M. cerifera, M. Carolinensis, M. Macfarlanei, M. Gale, and

Comptonia asplenifolia)

The epidermis of the young stems is covered by a cuticle more or less
thickened according to the species. The stomata are comparatively few
and frequently elevated. Hairs analagous to those of the leaves are
found in large numbers and are especially numerous on the distal extremi-
ties {vide supra).

The cortical parenchyme is composed externally of cells containing
chloroplasts. Internal to these are cells which during periods of rest
contain many starch grains. Certain thin walled cells are very rich
in tannin content, which may be readily determined by the use of ferric
chloride solution which imparts a bluish-black color to this substance.
Other cells of this region contain crystals of calcium oxalate in the form
of rosettes, tabular monoclinic prisms, or crystal sand. Hambright and
Moore have found in the cortex of Myrica species traces of gum, resin,
volatile oil, palmitic and myristic acids. Beringer has observed and
figured "secretory resin cells" in the cortex of Comptonia asplenifolia.
Chevalier (7, p. 103) and the writer have found these "secretory resin
cells" of Beringer to be nothing other than dead cells filled with gummy

lignin.

The pericambium consists of polyhedral cells of uniform size and
formed opposite each of the fibro-vascular patches. Very early many
cells of this region become thickened in their walls through the deposition
of lignin. The primary conducting system is composed of 10 cauline
fibro-vascular bundles which form in the internode a lobed gamostelic
ring. These bundles break up in the vicinity of the nodes to furnish 3
veins running into each leaf. In the crown one is able to distinguish up
to 19 apparent conducting masses separated by very large medullary-
rays. At each node, the crown bundles split up in the region of leaf
insertion. At first a large cord of wood and bast becomes de-
tached, then, a Httle higher two smaller cords separate, which take up
their position, respectively, to the right and to the left of the first. The
gamostele closes up after first filling in the gap left by the departure of
the lateral bundles.

In each procambial cord the first phloem elements consisting of sieve
tubes, companion cells and phloem parenchyme cells may be distin-
guished in contact with the pericambial arc. The primitive tracheae
appear later in contact with the pith parenchyme. The woody fibre
differentiation of the procambial parenchyme, corresponding to the bast

362 YOUNGKEN— ON THE MYRICACEAE

fibre elements takes place subsequently (7 p. 100). The external layer
of the cortex gives rise to the cork cambium.

Secondary Structure

Aerial Stem of M. cerifera, L.

(Plate 87, Fig. 19)

The cork consists of several layers of brick-shaped cells, whose walls
are for the greater part suberized. The contents of these cells stain
variously with safranin and methyl-green.

Separating the cork from the cortex, one notes a phellogen layer of
tangentially elongated cells. The cortex is usually narrow and shows
comparatively small angular intercellular air spaces. Its cells are
tangentially elongated. Some of them contain gummy lignin and tannin,
while many are filled with starch grains, or crystals of calcium oxalate
which may be in the form of monocUnic prisms or rosette aggregates.
Occasionally these crystals may also be seen in the air-spaces. Stone
cells of irregular form and size, arranged singly or in small groups, are
scattered throughout this region. The outer 3 or 4 layers of exocortex
are composed of collenchymatic cells and show no air-spaces. A con-
tinuous zone of sclerenchyme fibres and cells is present in the pericycle.
The sclerenchyme cells are of different sizes and shapes and show varying
degrees of lignification. Frequently cells containing crystals of calcium
oxalate are found adhering to them. Some of the stone cells contain
gummy lignin. The phloem region is traversed by numerous primary
and secondary medullary-rays which generally broaden out toward the
the cortex, arching in many phloem masses. The primary medullary-
rays are usually 1-3 rows of cells in width in this region. The secondary
medullary-rays of the phloem vary from 1-2 rows of cells in breadth.
The sieve tubes show obHque sieve plates. Their walls show numerous
transverse pits. Tannin and crystals are quite abundant in the phloem
parenchyma. Both stone cells and narrow bast fibres are sparsely
scattered either singly or in small groups amongst the other phloem
elements. The bast fibres are accompanied by crystal fibres. Each cell
of the crystal fibres contains a monoclinic prism of calcium oxalate. The
cambium zone is distinct and somewhat wavy. The xylem region is
traversed by numerous primary and secondary medullary-rays which
separate this zone into many narrow wood wedges. The primary medul-
lary-rays of this region are mostly 1-2, rarely 1-3 rows of cells wide, the
secondary, mostly 1 cell, rarely 2 rows wide. The walls of the medullary- .
ray cells are lignified. The tracheae are pitted and appear circular to

OF THE EASTERN UNITED STATES 363

very slightly angular in transverse section. Some of them show the
presence of the Actinomyces organism. Alike with those of the root, they
e.xhibit barred septa. Their distribution in the spring, summer and
autumn wood is strikingly uniform. The numerous wood fibres are
narrow, elongate and taper ended. Their walls show many oblique pits.
The wood parench>Tna cells are elongated in the direction of the long
axis of the stem and show slightly thickened walls.

The pith region is composed of rounded to polygonal parenchyma
cells, whose walls become lignified very early. Some of them contain
a brownish-yellow substance (gummy lignin) while others contain starch
grains. Protoplasmic processes extending from cell to cell are quite
prominent.

Subterranean Branch of Myrica cerifera, L.
The rhizome of Myrica cerifera differs histologically from the aerial
stem in the following particulars: — There are fewer layers of cork and
more layers of cortical parench\Tne cells. In transverse section the
cortex cells appear more rounded, less tangentially elongated. The
intercellular air-spaces are somewhat larger. No crystals of calcium
oxalate have been found either in the cortex or pith. There are generally
more cells containing gummy lignin. Sclerenchyme elements are
entirely wanting in the cortex. A discontinuous zone of sclerenchyme
is present in the pericycle, composed of numerous small islets of peri-
cambial fibres. The phloem region shows fewer bast fibres. The
medullary-ray cells are broader and have thinner walls. The tracheae of
the xylem are fewer and for the most part about twice as broad. The
largest have a maximum breadth of 57.6 n, while the largest of the aerial
stem are 38.4 fx. In some of these the writer has observed colonies of
coccus-like structures which probably represent an involution form of the
Actinomyces previously discussed. The pith cells are about twice as
large, contain more starch and have considerably thinner walls.

Aerial stem of Myrica Carolinensis, Miller

Passing from periphery toward the centre the following structures
can be observed: —

1. Cork of several layers of brick-shaped cells, whose cell walls are
either suberized or lignified or in part suberized and in part lignified.
There is a tendency for more lignification to occur in this region than in
that of the aerial stem of M. cerifera.

2. Phellogen of a single layer of thin-walled tangentially elongated
cells.

364 YOUNGKEN— ON THE MYRICACEAE

3. Cortex generally broader than corresponding region of M. cerifera
of same age. Its cells are tangentially elongate and contain either starch
grains, gummy lignin or crystals of calcium oxalate either in the form of
rosette crystals or monoclinic prisms. The intercellular air-spaces are
generally somewhat larger than those of the cortex of M. cerifera aerial
stem and frequently contain several crystals lying in a row one above the
other. Stone cells are quite numerous and scattered either singly or in
groups throughout the cortical parench}Tne.

4. The pericycle shows a broken ring of sclerenchyme elements.
These are for the most part considerably narrower than those in the
stem of M. cerifera.

5. The phloem is uniformly narrower than the similar region of
M. cerifera. Through it run numerous medullary-rays. The phloem
elements are smaller than those of M. cerifera. Many of the phloem
cells contain rosettes or monoclinic prisms of calcium oxalate. Tannin
is also present in large amounts Few bast fibres occur in this region.

6. The cambium line is much less wavy than that of M. cerifera.

7. The xylem consists of numerous tracheae containing Actinomyces
forms intermingled with many woody fibres and comparatively fewer
wood parenchyma elements The tracheae are generally narrower than
those of M. cerifera aerial stem. The largest have a measurement of
38.4 jjL across, similar to the largest in the aerial stem of M. cerifera.
The autumn wood is almost wholly composed of greatly thickened woody
fibres which collectively form a band in each annual ring fully twice as
wide as that observed in M. cerifera. (Plate 87, Fig. 21.)

The medullary-rays are thinner than those of M. cerifera. This is
not due to fewer rows of cells in these structures but rather to the fact
that the cells are smaller and narrower. Like those of M. cerifera the
medullary-ray cells of the xylem have their walls more or less lignified
and serve for the storage of starch. The primary medullary-rays are
mostly 1-5 cells wide, the secondary 1-2 and 1-3 cells wide.

8. The pith is composed of active polygonal to rounded cells, whose
walls are lignified. They serve for the storage of starch. Some fre-
quently contain gummy-lignin.

Subterranean branch of Myrica Carolinensis, Miller

There are fewer layers of cork formed than in the aerial stem. The

cells of the cortex are somewhat larger, more rounded and contain more

starch. The intercellular air-spaces between cortical parenchyma cells

are larger and devoid of calcium oxalate crystals. The sclerenchyme

OF THE EASTERN UNITED STATES 365

elements in the cortex are smaller and fewer than in the above ground
stem. The pericycle differs from that of the aerial stem in having widely
separated small groups of narrow sclerenchyma elements.

The phloem region is composed of wider elements than exist in the
similar region of the above ground stem. It is traversed by broader
medullary-rays which are from 1-4 to 1-5, rarely 1-6, rows of cells wide
(primary medullary-rays) or from 1-2 to 1-3 rows of cells wide (secondary
medullary-rays). Bast fibres are present in both proto- and meta-
phloem regions, in small groups, which are surrounded by crystal fibres
containing monocUnic prisms of calcium oxalate. The pitted tracheae
of the xylem are fewer than in the aerial stem and about twice as broad.
The largest have a maximum breadth of 81.56 /x. The xylem medullary-
rays are numerous and have cells considerably broader than those of the
aerial stem. The primary medullary-rays are mostly 1-5, rarely 1-6,
rows of cells wide. The secondary medullary-rays are from 1-2 to 1-3
rows of cells in breadth. The autumn wood is similar to that of the
aerial stem. The pith region is uniformly wider, its cells somewhat
smaller, and with thinner walls than the corresponding area of the aerial
stem. iSIany of its cells, like those of the cortex, contain starch and
gummy lignin.

Aerial stem of Myrica Macfarlanei, Youngken
(M. cerifera x M. Carolinensis)
The aerial stem of the hybrid resembles that of M. cerifera in the
following structural details: — (a) a continuous sclerenchyme zone in the
pericycle, (b) the tendency of the phloem masses to become arched in
their outer portions due to the broadening out of the medullary rays at
their extremities, (c) the presence of many bast fibres accompanied by
crystal fibres as well as stone cells in the phloem, (d) the uniformity
in distribution of the tracheae. It resembles the aerial stem of M.
Carolinensis in the size and relative number of tracheae found in the
protoxylem.

It differs from both parents as follows — The autumnal wood is
intermedate in thickness; the pitted tracheae are fewer than in M.
cerifera, more numerous than in M. Carolinensis; the mean diameter of
the tracheae is intermediate between that of both parents (Plate 87, Fig.
20). The primary medullary rays are 1-3, rarely 1-4, rows of cells wide;
the secondary, 1-2 rows in width. Other stem structures common to
both parents are likewise seen in the hybrid.

366 YOUNGKEN— ON THE MYRICACEAE

Subterranean branch of Myrica Macfarlanei, Youngken
The rhizome of the hybrid between M. cerifera and M. Carolinensis
shows the following histological peculiarities: — Compared with the aerial
stem of the same age, it shows fewer cork layers and a broader cortex.
The cells of the cortex are generally larger, less angular, and contain
more starch grains. The bast fibres of the phloem are about twice as
numerous and in larger groups. The tracheae are fewer and generally
broader, while the woody fibres are more numerous and more lignified.
The pith is narrower; its elements smaller and more lignified. The pri-
mary medullary-rays are broader and frequently show 1-5, less often 1-6,
rows of cells in width. The secondary medullary-rays vary in width
from 1-2 to 1-3 rows of cells. Cortical parenchyme, phloem, and tracheal
elements with a gummy lignin content are more frequent than in the
above ground stem. Compared with rhizomes of its parents, it shows
intermediate characters in respect to the number and mean diameter
of the tracheae and the woody fibres.

Aerial Stem of Myrica Gale, L
(Plate 87. Fig. 23.)
In passing from periphery toward the centre, the following histologi-
cal peculiarities are seen: —

1. A cork zone of unevenly stained brick-shaped cells whose walls
are partly suberized and partly lignified. The lignification occurs on
the inner tangential walls.

2. A phellogen of thin walled tangentially elongated cells.

3. A comparatively narrow cortex composed of cells some of which
are thick walled and contain gummy lignin, some thin walled, containing
starch while others are rich in rosette, monoclinic prism, or crystals and
forms of calcium oxalate. The intercellular air-spaces of this region
are small to medium and angular. Many contain several crystals of
calcium oxalate. The cortex is entirely devoid of sclerenchyme elements.

4. A pericycle which contains an interrupted zone of sclerenchyme
elements.

5. A very narrow phloem, even in old stems, which is devoid of bast
fibres but contains numerous sieve tubes and phloem parenchyme cells.
Through this region course numerous medullary-rays which broaden
out toward their extremities and arch in many phloem masses.

6. A cambium of one layer of meristematic cells with thin walls.

7. A xylem composed of numerous narrow compactly arranged wood
wedges separated by medullary-rays which are mostly 1 cell wide (second-

OF THE EASTERN UNITED STATES 367

ary med. rays); a few (primary med. rays) vary from 1-2 to 1-4 rows
of cells in width. The cells of these medullary-rays exhibit marked
protoplasm connections which are best seen in a radial longitudinal
section. Their walls are lignified. The wood wedges are composed
mostly of tracheids with bordered pits. The tracheae are generally
pitted and mostly found in the spring wood. Like those of the other
stems treated, they show barred septa. They are arranged radially
and are more or less polygonal in transverse section. Some of them
show a bacterioid content. Spiral tracheae, as in the other stems, are
found in the protoxylem.

8. The pith, as noted by Gris,^^ is composed of thick walled cells
which are mostly functionally active. Some of them contain gummy
hgnin.

Subterranean Branch of Myrica Gale, L

The subterranean branch of this plant differs microscopically from
the above ground stem in the following particulars: — The cork zone is
narrower. The cells of the cortex are somewhat larger and show fewer
crystals of calcium oxalate. There are more tracheids and fewer spiral
ducts in the protoxylem. The pitted tracheae with bacterioid contents
are more numerous. The tracheae tend to become larger.

Aerial Stem of Comptonia Asplenifolia, (L.) Alton

Passing from exterior toward the centre one notes the following
structures: —

1. A cork consisting of several layers of irregularly brick-shaped
cells which stain unevenly with safranin and methyl green. Many of
them have their inner tangential walls more or less lignified.

2. A cork cambium (phellogen) of thin walled meristematic cells.

3. A cortex of tangentially elongated parenchyme cells, some of
which contain starch, others gummy Hgnin, tannin or crystals of calcium
oxalate of rosette or tabular form.

4. A pericycle displaying an almost continuous zone of sclerenchyme
elements.

5. A phloem containing numerous more or less oval shaped groups
of bast fibres arranged in interrupted circles and separated from one
another by the soft bast elements. (Plate 87. Fig. 22.) Each group
of bast fibres is surrounded by a number of crystal fibres whose cells
contain monoclinic prisms of calcium oxalate. Considerable tannin
is evidenced by the use of ferric chloride test solution. The phloem
medullary rays are 1-5 rows of cells wide.

368 YOUNGKEN— ON THE MYRICACEAE

6. A cambium forming a very wavy zone of tangentially dividing
meristematic cells.

7. A xylem which is quite porous and composed of numerous wood
wedges separated by medullary-rays, for the most part 1-2 rows of cells
wide. The tracheae of the metaxylem show broad transverse pits in
their walls as well as barred septa and Actinomyces forms. The woody
fibres are extremely lignified especially in the autumn wood. The
medullary-ray cells show many prominent protoplasmic connections.

8. The pith, like that of M. Gale, is irregularly lobed. It is composed
of parenchyma cells which become lignified very early. These cells
are of variable size and appear mostly rounded in transverse section.
Many of them contain gummy lignin, while some possess rosette crystals
of calcium oxalate.

Subterranean Branch of Comptonia Asplenifolia, (L.) Alton
The microscopical characteristics of an underground branch of this
plant are for the most part similar to those of an aerial stem. The
important differences are a diminution in the number of cork layers,
the tendency of the medullary ray cells to become broader, the presence
of more cells containing gummy lignin and a slight increase in the diame-
ter of the tracheae.

Comparative Morphology of the Root Tubercles

Within the past thirty years various investigations have been carried
on by Brunchorst, Moller, Shibata, Chevalier, Harshberger and Arzber-
ger in respect to an endophytic organism living in the tissues of the
Myricas and forming tubercles.

Brunchorst" was the first to mention the tubercles on Myrica Gale
and named the fungus producing them — Frankia Subtilis — because
he considered this organism similar to that in the tubercles of Alnus.

Moller" later found the organism to differ considerably from that
infesting Alnus and named it — Frankia Brunchorstii.

Shibata^^ investigated the tubercles found on Myrica rubra, his ob-
servations being on both fresh and preserved material. He described
the morphology of the tubercle, showing that the fungus confines itself
to a ring of from 1-3 layers of cells beneath the cork thus differing from
the condition found in Alnus. He also pointed out that infection takes
place acropetally by means of fungal threads. He traced these threads
into the already differentiated meristematic cells where they grew rapidly

OF THE EASTERN UNITED STATES 369

to form a dense thready reticulum, then branched into radiate threads
whose free ends became swollen in clavate fashion. He assigned to
the fungus a position in the genus Actinomyces.

Chevalier, (7 p. 124-139) examined the tubercles on the roots of
Myrica Gale {Gale palustris), Myrica cerifera, Myrica Carolinensis
{M. Pennsylvanica), and Myrica sapida var longijolia. He found them
on main roots, adventitious roots of subterranean branches, and on
subterranean stems. He described at length their general gross structure
and histology, the occurrence of gummy lignin in the cells attacked, and
called the infesting organism, Frankia Brunchorstii, previously observed
by Moller.

Harshberger^^ observed the tubercles on the adventitious roots of
Myrica cerifera. He studied the structure of the mature tubercles from
dry material only which had been boiled in water and afterwards treated
with alcohol. He called the tubercles "mycodomatia" and claimed for
the infesting fungus a position closely related to the Oomycetes.

Arzberger" investigated the root tubercles of Myrica cerifera, Myrica
Gale, and Comptonia asplenifolia {Myrica asplenifolia) and stated like
Harshberger that these structures appear on adventitious roots which
grow out from the lower part of the stem or from branches or stems which
have been covered over with leaf mold or soil for several years. He
described the morphology and cytology of the tubercles but his illustra-
tions do not show the true nature of the radiating clavate branches. He
favored the opinion of Shibata in placing the fungus in the genus Acti-
nomyces. .

During the last three years the writer has collected and examined
abundant tubercle material of M. cerifera, M. Carolinensis, M. Mac-
farlanei, Youngken, {M. cerifera, M. Carolinensis) Youngken and Comp-
tonia asplenifolia at Palermo and Tuckahoe, N. J. ; of M. Carolinensis and
Comptonia asplenifolia at Clementon and Albion, N. J. ; of Comptonia
asplenifolia near Mainville, Pa.; and of M. cerifera, M. Carolinensis,
and M. Macfarlanei at Wildwood and Rio Grande, N. J. He has raised
M. cerifera seedlings bearing tubercles from seeds which he planted in
sandy soil in the University of Pennsylvania greenhouse. He has
furthermore examined tubercle material of M. Gale collected by Dr.
John M. Macfarlane along Trefethan Bay in Chebeague Island of Casco
Bay and on the southeastern part of Peak's Island in Casco Bay, Maine.

370 YOUNGKEN— ON THE MYRICACEAE

Methods

Material from many plants of each species was macroscopically and
microscopically examined both in its fresh and preserved condition. All
of the preserved material was fixed in weak, medium and strong Flem-
ming's on the ground immediately after the position and nature of the
tubercles on the plants had been ascertained. Samples of each lot were
then dehydrated in gradually increasing strengths of alcohol, cleared in
cedar oil and xylol and imbedded in parafi&ne. Transverse, tangential,
longitudinal radial, and longitudinal tangential sections were then cut
6-10 microns thick and subsequently stained in several ways. The best
results were obtained with the Methylene Blue and Acid Fuchsin com-
bination, although satisfactory results were also obtained with a com-
bination of Safranin and Gentian Violet.

The writer employed the following technique in isolating the endo-
phyte which produces the tubercles on M. cerijera, M. Carolinensis, M.
Macfarlanei, M. Gale, Comptonia asplenifoUa and probably most, if not
all, of these lesions on other plants of the Myricaceae.

A tubercle cluster from a root of one of the seedHngs grown in the
University of Pennsylvania greenhouse was washed thoroughly with clean
water to remove all traces of adhering soil. It was then introduced into
a test tube containing 1:1000 corrosive sublimate solution for 20 seconds
in order to destroy any surface organisms. From this it was transferred
with sterile forceps to a test tube containing distilled water which had
previously been sterilized in the autoclave. Into this was introduced a
sterile scalpel and two of the tubercles were cut into small fragments.
These fragments were next transferred to 5 tubes of sterile slant agar by
means of a sterile platinum loop. The tubes containing the culture were
then stored in a dark closet at ordinary room temperature for several
weeks. All 5 cultures when examined revealed the presence of Acti-
nomyces rosettes, non septate thin filaments, and rods of different sizes
as well as coccus forms, all of which stained well by Gram's method.
The coccus forms are probably for the most part products of the degenera-
tion of the above filament.* Jordan supports this view in regard to
similar forms of Actinomyces found in cattle, sheep, hogs and man.
The Actinomyces rosettes were found to be present in the depth of the
agar. This shows the anaerobic nature of the organism.

*From two of the above cultures, the writer has recently successfully grown pure
sub-cultures on coagulated horse serum in sealed tubes, kept at the temperature of
37.5'' C.

OF THE EASTERN UNITED STATES 371

Five seedlings of M. cerifera, which the writer had previously grown
from seed in the University of Pennsylvania greenhouse, were then
removed from the soil and their root systems loosened from adhering
sand by gently washing in clean water. With the aid of Mr. Lambert,
of the University Gardens, who, Uke the writer took special care to insure
against sources of infection by other organisms, the root systems were
one by one quickly dipped into 1:1000 corrosive sublimate solution and
then washed in sterile distilled water. While Mr. Lambert, with sterile
hands, held each seedling so treated, the writer, by means of a long needle,
previously sterilized by passing through the bunsen flame, removed a
small portion of the Actinomyces culture from one of the tubes and pricked
it into the root of 4 seedlings, marking the place of inoculation by tying
a sterilized piece of cord just above the puncture. The last seedling was
treated similarly -to the first four, with the exception that it was merely
pricked with a sterile needle. This served as a control. Each seedling
was then planted in a sterile pot containing sterile sand. Both pot and
sand were previously sterilized in the hot air oven at a temperature of
210° C. for 8 hours. The potted seedlings were then placed in a special
case in the greenhouse and daily watered with sterile Knop's solution.
At the expiration of 9 weeks the seedlings were carefully removed, washed
in clean water and their roots examined for the presence of tubercles.
These were found in a primitive state at the points of inoculation on all
but two including the control, which was pricked with a sterile needle
only Thin hand sections of one of the tubercles revealed the presence
of Actinomyces in the same condition as observed in the cells of the
tubercles of the M. cerifera seedling, as well as of the tubercles on the
other species above noted. The appearance of the infesting Actinomyces
within the cells of the host plants will be treated under the caption dealing
with the histology of the tubercles.

Gross Structure of Tubercles

The writer has found tubercles on the M. cerifera, M. Carolinensis,
and M. Macfarlanei seedling primary roots of 5 to 6 months' growth, and
from thence onward on the secondary roots inserted on the hypocotyl
axis, on nearly all the adventitious roots of subterranean branches and
on the subterranean branches of M. cerifera, M. Carolinensis, M. Gale,
M. Macfarlanei, and Coniptonia asplenifolia.

The tubercles occur either singly, as is frequently the case on sub-
terranean branches, in small groups the size of a pea, or in larger coralloid
loose or compact clusters which frequently attain the size of a large

372 YOUNGKEN— ON THE MYRICACEAE

black walnut. Each tubercle is a short cyhndrical blunt-ended root
hke structure which branches di- or trichotomously after attaining a
certain length. The branches frequently rebranch at their tips, which
grow out into long thread-like structures from 1-3 cm. in length that
may also branch and become entwined about the roots of other plants.
The maximum length of a tubercle is 5 mm. The average length of the
branches being from 2-3 mm. The color of the youngest tubercles is a
pinkish gray-brown. As the tubercles become older their color changes
to brown, dark-brown and even black.

Histology

The tubercles when studied microscopically exhibit the following
structural detail:

A cork, constituted of from 2 to 4 layers of suberous cells, whose outer
ones are dead, filled with gummy lignin, and in the process of exfoliation,
forms the external bounding layer. The cork tissue is derived from the
outer layer of pericambium of the host root which functions as a phellogen
during the development of the tubercle. Beneath the cork lies a very
broad cortex, which, instead of being formed as in normal roots of 5-12
layers of cells separated by large intercellular air spaces, is constituted
of 15-24 layers of very closely united parenchyma cells. The outer 3-5
layers of this region are composed of rounded to tangentially elongated
cells, some of which contain starch grains, others tannin, a few gummy
lignin. Underneath this lies a zone usually 2 to 3 cells broad of radially
elongated cells and a few smaller rounded cells which are separated by
small air-spaces. The radially elongated cells are hypertrophied and
contain the Actinomyces parasite (Plate 88, Fig. 24), which may or may
not be enveloped by gummy lignin. Many of the abutting smaller
cells are rich in tannin and show no evidence of the parasite. Beneath
this zone of infested cells is found a usually broader zone of smaller
isodiametric cells intermingled with a few oblong cells. In M. Carolinen-
sis as noted by Chevalier, in M. Macfarlanei, M. Gale and Comptonia
asplenifolia as noted by the writer, it frequently happens that other
cells scattered without order throughout the cortex are also infested by
Actinomyces. These hke those of the infested radially elongated zone
are also hypertrophied. All of the infested cells are united by means of
Actinomyces threads which run through the cell walls from cell to cell
as well as the intercellular air-spaces. The endodermis or innermost
layer of the cortex is composed of small oval thick-walled cells which
contain a yellowish-brown substance (gummy lignin). The walls of

OF THE EASTERN UNITED STATES 373

these cells become suberized very early. Underneath the endodermis
is found the vascular cyUnder, which is quite reduced in size as compared
with that of the normal root. In the young tubercle it is constituted of a
a radial tetrarch fibrovascular bundle which surrounds a small pith. The
phloem elements of the bundle become inactive very early. The xylem
is composed mostly of wood fibres intermingled with a few tracheae.
Secondary development is of very short duration.

In the younger tubercles the vascular cylinder extends only part way
into the apex, while in older ones the cylinder, with some cortical paren-
chyma cells surrounding it, grows out into a slender thread from which
lateral branches are then sent off.

Actinomyces living in the tubercles is best observed in its various
relations, in a radial longitudinal section. There the youngest stages
may be observed in the meristematic region of the apex, while the older
stages may be traced back toward the base of the tubercle. As ob-
served by Shibata in M. rubra tubercles, the writer has likewise noted in
the case of the tubercles of M. cerifera, M. Carolinensis, M. Macfarlanei,
M. Gale and Comptonia asplenifoUa that the differentiation of the ring
of infested cells starts in the meristem near the growing apex, thus indi-
cating that infection takes place acropetally. Since tubercles are found
on the seedling roots of 5-6 months' growth, it would indicate that
infection takes place very early in the life of the young seedUngs.

These infested meristematic cells become radially elongated and are
found to contain several extremely fine non septate and branched thread-
Uke structures. The Actinomyces (Streptothrix) threads extend through
the transverse and longitudinal walls of these cells aided evidently by
the secretion of a ferment which dissolves the cell wall in the line of the
organism's progress. They then invade neighboring cells where they
run between the starch grains toward the nucleus around which the
organism seems to derive its greatest benefit. Shortly after the appear-
ance of parasite threads within the invaded cells, the starch grains become
dissolved and in this form are appropriated as food by the organism.
The nucleus becomes hypertrophied and finally perishes. The endophyte
by this time has grown very rapidly into a dense thready reticulum.
Gummy lignin appears at first of a clear yellow color but later becoming
yellowish brown. The dense thready web of the organism sends out
clusters of radial thread branches forming an Actinomyces rosette, which
in many cells completely fills up the whole cell lumen. These threads
frequently become club shaped at their extremities. Some of the
threads after piercing through the wall of a cell develop clavate ends.

374 YOUNGKEN— ON THE MYRICACEAE

In due course of time, as is evidenced in many older infested cells, the
fungal threads become shrunken together and impregnated with gummy
lignin forming a good sized lump of degenerating material within the
cell, which remains connected with similar lumps, or mycelial webs of
adjacent cells by means of threads which penetrate the cell wall. Some
of the cells containing the rosettes also show coccus-Uke forms, while
other cells, especially in the older basal portion of the tubercle, are almost
completely filled with these. These cocci are probably products of the
disintegration of the filament. They may be involution forms of the
Actinomyces organism which appear in cells whose contents are poorly
adapted to the trophic needs of the endophyte. Their presence in such
numbers on artificial culture media would support this hypothesis.
While Actinomyces is the primary infecting agent responsible for the
tubercles on Comptonia asplenifolia, there frequently later appears in
the cells and intercellular air-spaces of some of the tubercles a mycelium
producing fungus with unseptate hyphae belonging probably to the
Oomycetes as Harshberger suggested. The h>T3hae of this fungus are
several times as thick as those of Actinomyces. They penetrate through
the cell walls of the tubercle, passing from cell to cell and often coil up
into a mycelial mass in many of the cells invaded (Plate 88, Fig. 27).

Since Actinomyces is frequently a virulent pathogenic organism in
cattle and other domestic animals up to man, because the swellings it
produces on plants are analagous to those on animals, since the forms of
the organism as shown by Jordan^** in the infested lesions of animals are
similar to those which the writer has described in the lesions of Myrica,
and since the cultural characteristics of the organism isolated from the
lesions of animals by Wright,^' Wolff and IsraeP^ are in many respects
similar to those isolated from the Myricas and described by the writer,
he would regard the organism as a parasite and suggest its possible
pathogenic relation to such animals.

The Actinomyces not only confines itself to the cortex of the tubercular
roots, it later works its way into the tracheae of these structures, passes
into the pitted vessels of the main roots, thence into those of the stems,
and, conveyed by the transpiration stream gradually upward is carried
through the axes of catkins so as finally to reach the flowers, bracts and
fruits. In these it confines its existence to the parts corresponding to the
mediocortex of the root tubercles, namely the mesophyll and outer
mesocarp regions respectively.

The writer having isolated the organism in pure culture from the
lesion produced by it on the seedUng tubercles hereby assigns to it the
name Actinomyces myricarum.

OF THE EASTERN UNITED STATES 375

Comparative Morphology of the Leaves

Gross Structure
Myrtca cerifera, L.
The leaves of this plant are simple, alternate, exstipulate, inserted in
a 2/5 spiral, pinnate-reticulate in venation, fragrant with a balsamic
resinous odor, coriaceous, evergreen and appearing toward the ends of
the branches which bear the flower buds of the following season about
the middle of May and persisting until the flower buds begin to open,
when they gradually fall, as the new leaf buds expand without assuming
a copper-red color. They are lanceolate-cuneate or oblong lanceolate,
(Plate 83, Fig. 6) varying from 30-100 mm. in length and 5-15 mm. in
width, often draw^n out toward both extremities, acute, mucronate, or
less often obtuse, or sharply notched at the apex, long cuneiform at the
base and decurrent on the short stout petiole. Their margin is thickened,
incurved beneath, very slightly eroded, entire or showing one to several
teeth in the anterior half or third. The upper surface of the lamina is
of a shining dark-green color and exhibits numerous crowded pits, some
of which contain orange-red and others golden-yellow glandular hairs.
Simple hairs are also present on the depressed midrib and along the mar-
gin in the young condition of the leaves, but disappear as the leaves be-
come older These are fewer than on M. Carolinensis or M. Macfarlanei,
leaves. The lower surface of the lamina is of a yellowish-green color
showing fewer simple hairs than M. Carolinensis in the young condition
of the leaves. The simple hairs are entirely absent or very rare on this
surface of older ones. Orange-red and golden-yellow glandular hairs
are very numerous as on upper epidermis. Both midrib and lateral
veins, as well as many branches of the lateral veins are very prominent,
the lateral veins being inserted on the midrib at an angle of 45-60°.

Myrica Carolinensis, Miller
Leaves simple, alternate, exstipulate, inserted in a 2/5 spiral, pinnate-
reticulate in venation, fragrant, membranous, deciduous, beginning
to appear toward the ends of the branches which bear the flower buds
of the following season the latter part of April (the writer has observed
many leaves expanded at Clementon, N. J. on April 23, 1915) and in
full foliage by the first or second week in May. They remain green all
summer, assume a greenish-brown hue on an extensive scale in October
and November and usually have completely fallen by the middle of

376 YOUNGKEN— ON THE MYRICACEAE

December. An exception to this rule was noted by the writer at Clemen-
ton, N. J. where on February 14, 1915, in a valley protected by tall pines
and harboring abundant underbush, the young plants retained many
of their leaves, but the majority were partly discolored. They are usually
elliptic-obovate, varying from 30 to 105 mm. in length and from 18 to
45 mm. in width, the maximum size being attained on the sapling shoots,
rounded at the summit and shortly apiculated, feebly attenuated at
the base; the most with margin entire, very hairy, feebly incurved below,
the others presenting in their anterior half small to large thickened,
rounded crenations, each terminated ordinarily by a very small point.
The petiole shows a few golden-yellow glandular hairs but covered with
white simple hairs on both faces. The upper surface of the lamina
is of a bright green color, more or less shining, covered in the adult state
with short simple hairs amongst which are scattered golden-yellow glan-
dular hairs, never very numerous. The lower surface is dull green and
shining in color, finely reticulated, showing numerous simple white
hairs and golden-yellow glandular hairs. The mid-rib and lateral
veins while visible on the upper surface, are more prominent on this sur-
face.

Myrica Macfarlanei Youngken. (ilf . cerifera x M. Carolinensis)
The leaves of this hybrid between M. cerifera and M. Carolinensis
show several striking macroscopic characters, which are intermediate
between those of its parents. For instance, they vary from lanceolate-
cuneate to elliptic-obovate in shape, many of them being a mean between
these two forms (Plate 85. Fig. 13). In duration they are semi-evergreen,
and usually fall during February-March, by which time they have
often assumed a sUght coppery tint. In size, they vary from 25-58 mm.
in length and from 8-20 mm. in width. They have numerous orange-red
and golden-yellow glands on their lower surface with merely a few of
each of these on their upper surface. Simple hairs are found on both
surfaces and margin, as on the leaves of Myrica Carolinensis, but relatively
fewer in number. The texture of the leaves is sub-coriaceous. In
color the leaves have blended the color characteristics of both parents.
The margin is more incurved beneath than M. Carolinensis, less than
M. cerifera.

Myrica Gale, L

The leaves of this plant are simple, alternate, exstipulate, inserted
in a 2/5 spiral, pinnately and reticulately veined, subcoriaceous, de-

OF THE EASTERN UNITED STATES 377

ciduous, varying in length from 25 to 90 mm., in width from 8-30 mm.,
oblong, lanceolate or oboval cuneiform in shape, longly attenuated,
cuneiform at the base; margin entire along the 2/3 or 3/4 of their length,
somewhat thickened and hairy, slightly incurved below, showing ordinar-
ily toward the summit from 3 to 5 pairs of small serrations, acute mu-
cronate, or obtuse at the ape.x. (Plate 84. Fig. 8 and 9). The petiole
is 1-3 mm. long with its superior surface plane and thin as well as glandu-
lar hairy. The upper surface of the lamina is dark green, slightly thin
hairy, wrinkled, covered with depressed reticulations and showing a
scattering of golden yellow glands which are fewer than on the lower
surface; mid-rib depressed, secondary nerves non apparent. The lower
surface is of a pale green color. It is covered with thin hairs and nu-
merous golden-yellow glands. The mid-rib on this surface is projecting
and tomentose; the secondary nerves are inserted on it at an angle of
45-60°.

Myrica inodora, Bartram

Leaves simple, alternate, e.xstipulate, inserted in a 2/5 spiral, pinnate-
ly and reticulately veined, coriaceous, evergreen, oblong-obovate or
occasionally ovate, obtuse or sometimes pointed and occasionally apicu-
late at the apex, narrowed at the base, decurrent on the short stout
petioles, the margin thickened, entire or rarely obscurely toothed toward
the apex, varying in length from 30-100 mm. and in width from
17-38 mm. When they unfold they are covered with pale yellow
glands. The upper surface is glabrous; lustrous dark green, showing
a somewhat depressed glandular mid-rib and lateral veins. • The lower
surface shows its mid-rib veins more prominent. The mid-rib is
often slightly puberulous. Sargent states that the leaves of this plant
begin to fall in May and disappear from the branches before mid-sum-
mer. They differ from the fragrant leaves of the other eastern species
in being inodorous.

Myrica cerifera var. pumila, Michaux

Leaves simple, alternate, exstipulate, sessile or with a petiole 1-2
mm. long, inserted in a 2/5 spiral, pinnate-reticulate in venation, coria-
ceous, evergreen, oblanceolate, varying in length from 20-45 mm., and
in width from 5-10 mm., acute or mucronate at the apex, cuneately
narrowed at the base. The margin is revolute, slightly thickened
and incurved beneath, commonly saw-toothed in its upper 1/3. The
petiole when present is the shortest of any of the eastern species. In

378 YOUNGKEN— ON THE MYRICACEAE

Other respects the herbarium material studied closely resembles the
leaves of M. cerifera.

Comptonia asplenifolia, L, Aiton

The leaves of this plant are alternate, inserted in a 2/5 spiral, petio-
late, stipulate, pinnately-veined, distinctly pinnatifid, showing 4-15
pairs of sub-reniform pinnules measuring 2-14 mm. long and 2-10 mm.
wide, separated from each other by a sinus extending almost to the mid-
rib, membranous, deciduous, with membranous caducous stipules (Plate
82. Fig. 5). They vary in shape from lanceolate to elliptic-ovate and
in size from 10-138 mm. long to from 2-22 mm. wide. Each lobe, as
previously shown by Chevalier, receives ordinarily two secondary nerves
and two demi nerves arising from a thread corresponding to the indenta-
tion, which thread, arriving at the sinus, divides into two branches dis-
tributed to two adjacent lobes. More often the sinus extends to within

1 or 2 mm. from the mid-rib. The small secondary nerve which nor-
mally divides into two has not the time to bifurcate and is usually thrown
into the superior lobe. The result is that all the lobes have 3, 4 or 5
secondary nerves of approximately equal importance. Occasionally

2 lobes completely fuse to form an auricle with a maxunum length of
12 mm. This receives a double number of nerves. The stipules are
auriculate or semi saggitate, up to 13 mm. long, gibbous at their base,
acuminate at their summit, hairy around their margin. The petiole
is sub-cylindric, thin hairy and showing a sparse covering of glands
(contrary to Chevalier's statement that glands are absent). The
pinnules are entire or showing occasionally a tooth, with a thin hairy
margin, which is not incurved below. The upper surface is dark green
tomentose in young leaves, becoming nearly glabrous as the leaves attain
maturity.

The lower surface is light green, thin hairy in the young state of the
leaf but dropping as the leaf ages. Both surfaces show shining yellow
glands, not very numerous. These are somewhat more frequent on
the lower than on the upper surface.

Histology

Myrica cerifera, L

The upper epidermis consists of a single layer of cells whose outer

walls are strongly cuticularized and have rectilinear or very slightly

curvilinear vertical walls. Chevalier (7, p. 113) errs in stating that

Comptonia is the only species whose upper epidermis shows cells with

OF THE EASTERN UNITED STATES 379

walls slightly curvilinear. Over the veins the cells are rectangular
and elongated in the direction of the leaf bundles. The mean dimen-
sions of these cells are 23.20 nx 15.20 fx. The cuticle is punctated by
very small pearls within which, according to Chevalier, (7, p. 113) are
associated fine granules of wax. The summit of each epidermal cell
forms a rounded projection which together with others give a rough
aspect to the surface. The bottom and transverse walls remain thin.
The surface of this epidermis contains numerous cr^-pts not any larger
than those of the lower epidermis. Some of these crypts contain a
stalked golden-yellow balloon-shaped gland, while others possess a
stalked reddish-brown bowl-shaped gland which becomes black as the
leaf ages. The stalk of the balloon gland is 4-7 cells high and two rows
of cells in width. The gland is unicellular and contains oil in its young
state which resinifies as the gland becomes older. The stalk of the
bowl-shaped gland ordinarily varies from 7-8 cells in height and is 4
cells wide on the mature adult leaf. The head or gland is multicellular
and contains resin. Very few short unicellular simple trichomes are
found. These when present are sclerified at their bases. Scattered
here and there on the surface and in the middle of normal cells are isolated
refractile elements which represent the bases of these unicellular simple
hairs which have broken down on the level with the epidermis and are
completely sclerified. Stomata are absent on this epidermis. The
mesophyll shows palisade and spongy parenchyma differentiation.
The palisade zone is 4-5 cells broad and is composed of columnar elements
usually about twice as long as wide. Some of these elements contain
rosette crystals of calcium oxalate. The spongy parenchyma region
consists of chains of oval to irregular shaped cells surrounding large
intercellular air spaces. The fibro-vascular bundle of the mid-rib, as
well as smaller vein bundles, run through this region. Each is connected
with the upper and lower epidermis by means of hx-podermal cells, many
of which are enlarged and contain a rosette of calcium oxalate. The
fibro-vascular bundle of the mid-rib is reinforced above and below by
an arc of sclerenchyme. The xylem of this bundle is quite broad and shows
medullary rays 1-cell wide, which radiate like a fan. The ducts show
spiral thickenings. The phloem forms a thick semilunar band below
the xylem.

The lower epidermis consists of polygonal cells whose outer walls
are much less cuticularized than those of the upper epidermis. Like
those of the upper epidermis they contain a brownish content. Their
vertical walls are curvilinear. The mean dimensions of these cells are

380 YOUNGKEN— ON THE MYRICACEAE

24.28 jux 12.14 m. Stomata are scattered amongst the lower epidermal
cells without order. Each in the mature adult leaf is surrounded by
6-8 neighboring cells. Crypts are more numerous than on the upper
epidermis. These contain similar glands to those found on the upper
surface. Very few 1-celled trichomes or their bases are ordinarily found.

Myrica Carolinensis, Miller.
The upper epidermis of this leaf consists of polygonal cells, whose
outer walls are covered by a thin cuticle. The vertical walls are recti-
linear. Over the veins the cells are rectangular and elongated in the
direction of the bundles. The cells are filled with a brownish substance
as in M. cerifera and M. Macfarlanei. Their mean dimensions are 29. 13 /x
X 17 M- Scattered over this epidermis are numerous simple unicellular
trichomes with sclerotic bases, as well as the bases of sclerotic hairs,
each of which is surrounded by several rows of cells arranged in radial
fashion. Very few or no golden yellow glandular trichomes are found.
These are more frequently present than absent. Each consists of a
stalk and a gland or head. The stalk is 4-7 cells in length and two rows
of cells wide. The gland is balloon-shaped. The glandular trichomes
are inserted into the epidermis at the bottom of crypts which are less
numerous on the upper than on the lower surface.

The mesophyll, as in the leaves of M. cerifera, shows dorso-ventral
differentiation into palisade and spongy parenchyma regions. The
palisade zone consists of 2-3 layers of loosely arranged columnar shaped
cells, some of which contain rosettes or monoclinic prisms of calcium
oxalate. The spongy parench}Tiia region comprises a network of oval
to irregular shaped cells arranged in chain fashion and surrounding large
intercellular air spaces. Some of these cells contain crystals of calcium
oxalate. The fibro-vascular bundle of the mid-rib and smaller veins
course through this region. The stronger veins are united to the upper
epidermis by means of hypodermal elements which are coUenchymatic.
Other veins are united to both epidermises by hypodermal cells contain-
ing crystals of calcium oxalate. The fibro-vascular bundle of the mid-
rib is surrounded above and below by an arc of sclerenchyme, the upper
arc as in M. cerifera being the more strongly developed of the two.

The lower epidermis consists of wavy walled cells which have the mean
dimensions of 28.02 m x 14.4 m- Numerous stomata occur on this
epidermis. These are surrounded by 5-6-7 neighboring cells. Crypts
are numerous and scattered without order over the lower surface. Each
contains a golden-yellow balloon-shaped gland subtended by a stalk

OF THE EASTERN UNITED STATES 381

two rows of cells wide and 4-7 cells high. The base of each balloon-
shaped gland stalk is surrounded by 16 somewhat elongated cells ar-
ranged in radial fashion. Simple unicellular trichomes also occur in
large numbers on this surface and along the margin. The base of each
is sclerified and elevated in papillar fashion above the epidermis. No
bowl-shaped glands occur on either surface of the leaf. The average
width of the lamina outside of the mid-rib is 168.96 n.

Myrica Macfarlanei, Youngken.

The microscopic characteristics of the leaf of this hybrid are inter-
mediate between those of its parents. For instance: — the upper epider-
mis is composed of cells whose vertical walls are more curvilinear than
M. cerifera, less curvilinear than M. Carolinensis. Their mean dimen-
sions in surface view are 26.7 yu x 16.89 ^u. Their outer walls are less
projecting than those of M. cerifera, more so than M. Carolinensis.
The cuticle is thinner than that of M. cerifera, thicker than that of M.
Carolinensis. The palisade region of the mesophyll is 3-4 layers wide
and so is intermeditate between this region in M. cerifera and M. Carolin-
ensis {vide supra). The lower epidermis consists of cells whose vertical
walls are more curvilinear than those of M. cerifera, less curviUnear
than those of M. Carolinensis. The mean dimensions of these cells
in surface view are 26.46 /x xl3.59 n which is likewise an intermediate
character. The stomata are fewer in number than on M. Carolinensis,
more numerous than on M. cerifera. The upper surface shows a scatter-
ing of orange-red bowl-shaped glands and golden-yellow balloon-shaped
glands, or sometimes the latter only. The lower epidermis shows
numerous orange-red bowl-shaped and golden-yellow balloon-shaped
glands, both of which are often fewer than on the similar epidermis of
M. cerifera. The tendency for the head of the bowl-shaped gland in
the hybrid to become saucer-shaped is very striking. Simple unicellular
trichomes and the sclerotic bases of these are common on both lower and
upper epidermis, but intermediate in number between those on the
leaves of the parents. Finally, the average thickness of the lamina
of the hybrid outside of the mid-rib is 201.6 /x while that of M. cerifera
is 268.8 /i and of M. Carolinensis, 168.96 fx.

Myrica Gale, L
The upper epidermis is composed of a layer of polygonal cells with
rectilinear to very slightly curvilinear external walls. It shows here
and there stomata which are fewer than on the lower epidermis. The

\

382 YOUNGKEN— ON THE MYRICACEAE

cuticle of this epidermis is somewhat thicker than that of the lower
epidermis. Many of the epidermal elements contain brownish contents.
In some the contents appear granular. Numerous bases of sclerotic
unicellular hairs are scattered here and there over the surface. A few
sclerotic hairs are found and generally occur above the mid-rib and the
veins. Each of these is somewhat elevated and surrounded by usually
radially arranged cells. Crypts, containing stalked yellowish glandular
hairs are numerous. These are generally deeper than those of the lower
epidermis. The mesophyll shows a differentiation into a broad upper
paHsade zone, a middle spongy parenchyme region and a narrow lower
palisade layer. The upper palisade zone consists of three layers of colum-
nar shaped cells with variously stained contents. Some of them con-
tain crystals of calcium oxalate which may be in the form of rosettes,
monoclinic prisms or crystal sand. Others contain one to several
radiating structures which suggest the Actinomyces rosette. The spongy
parenchyme region is composed of small celled elements which are ir-
regularly shaped and surround intercellular air spaces. The lower
paHsade layer is incompletely differentiated and consists of columnar
as well as irregular shaped cells, which are loosely arranged. Some of
these contain crystals of calcium oxalate. The fibrovascular bundles
of the mid-rib and veins run through the spongy parenchyma zone.
Those of the stronger veins, as well as that of the mid-rib, are connected
with both lower and upper epidermis by hypodermal elements. Those
above the bundles are usually lignified. Many of the non-lignified
hypodermal elements contain rosettes of calcium oxalate. The vascular
bundle of the smaller veins is connected with both iht upper and lower
epidermis by special layers of elongated cells with wide lumina. The
lower epidermis is less cutinized than the upper and consists of cells
which for the most part have their outer walls arched in papillose fashion
giving to this epidermis a roughened aspect. Numerous crypts contain-
ing yellow stalked balloon-shaped glandular trichomes are found, but
these appear broader and shallower than those of the upper epidermis.
Stomata are also present but more numerous than on the upper surface.
They are covered in by neighboring cells. The yellow balloon-shaped
trichomes consist of a stalk or pedicel supporting a balloon-shaped
head. The stalk is usually 3 cells high, the basal layer alone consisting
of 2 cells. The head or gland contains oil which resinifies as the leaf
ages. Simple unicellular sclerotic trichomes as well as the bases of
these are also found, but in larger numbers than on the upper epidermis.
Occasionally a 3-4 celled uniseriate glandular hair is found which has

OF THE EASTERN UNITED STATES 383

its terminal cell or one of its middle cells specialized as a gland and
containing oil. Isolated epidermal cells of the ordinary kind often
become enlarged and filled with oil and so also function as glands.

Comptonia asplcnifolia, (L.) Alton
The upper epidermis of the leaf of this plant consists of a layer of
cells whose external walls are slightly arched outward and covered by a
granular cuticle. The vertical walls of these cells are curvilinear. Over
the veins the cells become rectilinear and elongated in the direction of
the vein bundles. The mesophyll shows dorso-ventral differentiation
into a broad palisade zone and a narrow spongy parenchyma region.
The palisade zone consists of usually two layers of somewhat loosely
arranged columnar shaped cells, the outer layer having its elements
much broader than the inner one. The spongy parenchyma region is
composed of irregular shaped loosely arranged cells surrounding inter-
cellular-air-spaces. Rosettes of calcium oxalate are common in many
cells of the mesophyll. The vascular systems of the mid-rib and smaller
veins run through this region. That of the mid-rib is supported above
and below by an arc of sclerenchyme. Those of the smaller veins are
connected with the upper and lower epidermis by means of elongated
hypodermal elements, some of which contain rosettes of calcium oxalate.
The lower epidermis consists of a layer of cells with a granular cuticle.
Their external walls are arched outward in a more prominent manner
than those of the upper epidermis. The cells of this epidermis conse-
quently are distinctly papillose in character. Their vertical walls are
somewhat more curvilinear than those of the upper epidermis. Numer-
ous stomata are scattered irregularly amongst the epidermal cells. These
are surrounded by neighboring cells. Both upper and lower epidermis
show four kinds of hairs. Of these, two are simple and two glandular.
One type of simple hairs and by far the more common is composed of
a single cell whose wall is sclerified throughout. At its base, as also
shown by Chevalier, (7, p. 118) sclerenchyme threads extend in between
surrounding cells in order to strengthen its insertion. The second type
of simple hair is furchate and really con'Sists of two hairs arising from
contiguous basal cells whose adjoining margins have become soldered
at the base. The first type of glandular hair is also seen in M. Gale
and consists of 3-4 superimposed cells whose terminal cell or one of its
median cells becomes enlarged and filled with oil. The second type
consists of a stalk of 3-4 cells bearing on its summit a several celled
spherical head which contains oil.

384 YOUNGKEN— ON THE MYRICACEAE

Comparative Morphology of the Inflorescences and Flowers.

The characteristic inflorescence of the plants of the Myricaceae is
a catkin. For M. cerifera, M. Carolinensis and M. Macfarlanei, the
catkins are partly formed the year before flowering below the leaves
on last season's growing branches. For M. Gale and Comptonia aspleni-
folia they are formed on special branchlets, which, after the development
of the stamens on the male individuals and the fruits on the female
individuals, cease to grow, and, upon the descent of the pollen tube to
the egg, they soon die up to the node whereon the most inferior catkin
is inserted and often up to 2 or 3 nodes lower. Chevalier (7, p. 143)
separated M. Gale {Gale palustris, Chevaher) from the Myrica genus
because of this character and placed it in the genus Gale. The writer
feels that this character is insufficient to warrant the establishing of
a new genus. The special branchlets after their death are separated
from the living part of the stem by several thickenings of cells filled with
gummy Hgnin. No disarticulation of the dead part takes place.

M. cerifera, M. Carolinensis, M. Macfarlanei and M. Gale have
staminate catkins which are borne on different plants from those which
bear the pistillate catkins. They are, therefore, dioecious. In regard
to Comptonia asplenifolia, this is frequently the case ; but it also frequently
happens that both staminate and pistillate catkins are found on the same
plant, even on the same branch, so that this species is both dioecious
and monoecious. On monoecious plants the pistillate catkins appear
below the staminate.

Structure of Catkins and Flowers of Myrica cerifera, L

The male catkins of M. cerifera are cylindric and unbranched and
vary from 6 to 18 mm. in length, and from 2.5 to 4 mm. in breadth. The
peduncle and rachis are yellowish-red and show a few thinly scattered
hairs as well as a scattering of orange-red and golden-yellow glands.
The staminate flowers are inserted in the axils of bracts which are ar-
ranged along the rachis in a 2/5 spiral. Each staminate flower consists
of 4-6 stamens. The filaments of these are coalescent in their lower
half, free above and bear extrorse anthers on their free ends. The
bracts are deciduous, reddish oval structures, rounded at the base and
sometimes attenuated. They are scarious, sHghtly thin hairy along
the margin and glandular hairy beneath. The bracteoles are caducous,
1 to 2 in number, sometimes absent, deciduous, and inserted on the
staminal column at different heights. The female catkins are short

OF THE EASTERN UNITED STATES 385

and oblong and attain a length of 5-10 mm. at the time of flowering.
They develop more slowly than the male. Each pistillate flower con-
sists of 2 carpels fused in their ovarian portion to form an ovate one-
chambered compound ovary, which is narrowed above into two elongated
spreading styles and stigmas. The pistillate flowers are arranged along
a rachis in a 2/5 spiral. Each is inserted in the axil of a deciduous bract
and accompanied by four deciduous bracteoles.

Myrica Carolinensis, Miller

The catkins of this plant possess a peduncle and rachis more hairy
than those of M. cerifera. Furthermore, they have but one kind of
glandular hair, the golden-yellow balloon type. The male catkins
(Plate 84. Fig. 11) are 6-8 mm. long and about 5 mm. wide. The writer
has found the staminate flowers of the male catkins with anthers in the
tetrad state at Noroton, Connecticut, April 27, 1914. He has also ob-
served them with mature pollen and ready to dehisce at Wildwood,
N. J., May 13, 1914. In material collected at Clementon, N. J., April
23, 1915, the writer has observed tetrads and young pollen. Each
staminate flower consists of 4-8 stamens whose filaments are welded
together similarly to those of M. cerifera in half of their length, thus
forming a staminal column which is inserted in the axil of a bract. The
bract is thin hairy along the margin and yellow glandular hairy beneath.
Bracteoles are either present or absent.

The female catkins (Plate 84. Fig. 10) are 8-10 mm. long and oblong
in shape. Each pistillate flower inserted thereon consists of 2 carpels
fused in their ovarian portion to form an ovate one chambered compound
ovary, which is extended above into 2 spreading styles and stigmas.
It is inserted in the axil of a greenish oval-lanceolate bract. The bract
is obtuse at its summit, slightly hairy along its margin, and conceals
two small deciduous bracteoles which are situated in its axil.

Myrica Macfarlanei, Youngken

The male and female catkins of the hybrid between M. cerifera and
M. Carolinensis resemble those of M. cerifera in possessing both orange-
red bowl-shaped and golden-yellow balloon-shaped glandular hairs
on both catkin axis and the bracts. Other characters are still under
investigation by the writer.

Myrica Gale, L

The male catkins of this plant are cylindrical (Plate 84. Fig. 9) 10-15
mm. long at the time of flowering. The staminate flowers each consist

386 YOUNGKEN— ON THE MYRICACEAE

of 3-6 stamens with distinct filaments bearing adnate yellow anthers.
The filaments are extremely short and are inserted in the axil of a red-
dish brown, broadly oval to sub cordate bract.

The female catkins (Plate 84. Fig. 8) are ovoid oblong, obtuse, from
10-20 mm. long and 6 mm. wide at the time of maturation. Each
female flower consists of 2 carpels fused in their ovarian portions, each
being terminated by a style and stigma. The ovary is flanked by two
lateral bracteoles which develop into aeriferous floats upon its maturation.

Comptonia asplenifolia, (L.) Alton

The male catkins are cylindrical and at the time of flowering attain
a length of 20-25 mm. They are clustered at the ends of special branches
in numbers ranging from 5-10 on each branch. Each staminate
flower usually consists of four stamens (occasionally 5) whose fila-
ments are distinct and short and bear reddish puberulent anthers.
It is inserted in the axil of a bract which is not accompanied by brac-
teoles. The bracts along the catkin axis are oval to oval-lanceolate,
generally 3 mm. long, terminated at the summit by a long point, brown
and scarious, showing long white hairs along their margin and golden-
yellow glands and thin simple hairs below.

The writer has observed the anthers in the pollen mother stage in
material collected at Clementon, N. J., November 10, 1913, in the tetrad
and mature pollen stage in material collected at the same place April 23,
1915.

The female catkins are ovoid and much smaller than the male,
attaining a maximum length of 5 mm. at the time of flowering. They
are grouped toward the end of special branches on pistillate plants or
are frequently found below the staminate catkins on monoecious plants.
Each pistillate flower consists of two carpels fused in their ovarian por-
tion, and each terminated by a long red style and stigma. The ovary
is fllanked laterally by two bracteoles which develop at their base and
later along their margin small laciniate outgrowths.

Comparative Morphology of the Fruits

M. cerifera, L
Fruits small, spherical, bluish-white, from 2-3 mm. in diamter, cov-
ered almost to their summit by fleshy knob shaped glands which are
devoid of a covering of wax in their young condition, but later are en-

OF THE EASTERN UNITED STATES 387

tirely covered over by that substance. Apex of mature fruits, aceri-
ferous. Two-hundred fruits which the writer gathered at Palermo,
N. J. November 7, 1914 weighed 2.86 gms. Fruits of one season past
alone adhere to the axes of the catkins.

M. Carolinensis, Miller
Fruits larger than those of M. cerifera, spherical , tomentose hairy
in their young condition, covered over early, as those of M. cerifera by a
thick exudation of wax, traversed at maturity by a portion of the per-
sistent hairs and having a diameter of 3. 5-4. 5mm. Apex of mature
fruits, ceriferous. Two hundred fruits which the writer gathered at
Palermo, N. J. November 7, 1914 weighed 9.27 Gms. The fruits of
two seasons past adhere to the persistent catkin axes.

M. Macfarlanei, Youngken

Fruits intermediate in size, apex, weight and duration between those
of M. cerifera and M. Carolinensis, having a diameter of 3-4 mm. Apex
of mature fruits pitted. Two hundred fruits which the writer gathered
at Palermo, N. J. November 7, 1914 weighed 6.7 Gms. The fruits
of one season past and several of the previous season adhere to the per-
sistent catkin axes.

Careful study has been made by the writer of stages in the develop-
ment of the fruits of the above types. These will be described as one
category seeing that they closely agree with each other.

In early June the maturing wall of each ovary has already developed
a series of striking and complex glandular hairs of the nature of tubercular
emergences. Each is a knob-like expansion of the fruit wall into which
a copious prolongation of subjacent mesocarp tissue has spread but
further, from an abundant and densely anastomosing complex of vascular
bundles that ramify through the outer half of the maturing mesocarp,
fine vascular diverticula composed of 2 or 3 spiral tracheae along with
delicate elongated sieve-like elements pass through the stalk of the
emergences and end in a slight swelling in its middle. The epidermis
at this time is comparatively shallow and thin walled, while from the
junction of the epidermis with the base of each emergence, elongated
unicellular hairs spring, which form a basket like system, round and upon
which copious wax exudations subsequently become aggregated.

At this time the mesophyll is divisible into two recognizable zones,
viz.: an outer irregular and large celled region consisting of about 12-13

388 YOUNGKEN— ON THE MYRICACEAE

layers and an inner smaller and more round celled tissue of more numer-
ous layers. Prolonged into the former from the point of attachment
of the fruit with the axis is a vascular tissue that on entering the outer
layer ramifies abundantly and as above stated gives off delicate diverti-
cula to the different emergences. At this time only slight indication is
observed of a difference in cell contents between the outer and the inner
zone. The endocarp is a shallow and delicate layer that from now on
becomes less and less conspicuous. By mid June or soon thereafter
striking changes begin to appear. The epidermis, as well as the meso-
phyll cells of the emergences, the general epidermal (epicarp) cells of
the fruit wall and the outer zone of the mesocarp have all enlarged
steadily and become filled with an abundant secretion which assumes a
bizarre coloration when stained with Safranin and Methyl Green. Tints
varying from neutral-gray through pink, crimson, crimson-green, green-
blue, yellow and brown are all present in distinct but neighboring cells
and though the writer has as yet been unable to apply abundant tests,
the above coloration suggests the formation and presence already of
the palmitic, myristic and stearic acids already tested for and recorded
by pharmacists and synopsized by Chevaher (7, p. 159).

At this time the inner layer of the mesocarp contracts conspicuously
with the last, its cells remaining small, thin walled, rounded and its
cavities filled with delicate protoplasm.

A further stage in the maturation of the fruits is noted by the middle
or latter part of July. Each knob-shaped emergence has developed
around itself an abundant waxy layer which, by hardening, gives a
bluish white color to the fruit surface. The mesophyll cells of these hairs,
as well as 3 to 6 of the innermost cell layers of the outer zone, become
loaded with cell contents that assume a uniform red or reddish-green
hue which even heightens the bizarre coloration noted above. By this
time also the cells of the inner zone have become largely thickened from
within-outward by lignified thickening and the cells themselves, having
increased greatly in size and become markedly sinuous in outline, assume
a bright red staining with Safranin.

The thickening process clearly proceeds in centrifugal fashion for
in mid or late July the inner walls, surrounding the ovarian cavity,
may be highly lignified and stained a bright red hue, while as yet the
external cells adjacent to the outer zone are little, if at all, altered in
shape or lignified. Progressive lignification of this area gives rise by
mid August to the extremely hard scleroid fruit layers that efficiently
protect the enclosed seed.

OF THE EASTERN UNITED STATES 389

Meanwhile steady excretion of wax takes place over the cells of each
emergence and these later have so grown together as to form a complete
coating around the fruit wall proper. So between the abundant wax
excretion and the close apposition of the wax secreting emergences, the
entire surface of each emergence assumes a uniform blue gray color
and is coated over by a rather brittle waxy investment that readily
crumples to pieces when slightly pressed between the fingers.

This investment forms an admirable defensive covering alike against
intense insulation, the attack of fungoid spores, and the destructive
action of caterpillars and other animal enemies.

Gross Structure of the Fruits of Comptonia asphnifolia, Alton

The fruit of Comptonia aspleuifolia is an akene which is subtended
by an involucral cupule consisting of 8 linear subulate bracteoles. The
akeneal portion is ovoid-oblong in shape, obtuse at the apex, light-brown
and shining. In association with other akenes and their bracteoles on
the catkin axis, it forms a bur-like structure.

Gross Structure of the Fruit of Myrica Gale, L
The fruit of .1/. Gale is a keeled nut consisting of the ripened ovary
accompanied by two accrescent bracteoles which have formed aeriferous
floats.

Prevalence of Actinomyces in the Fruits
Actinomyces myricarum Youngken has been observed by the writer
in its most luxuriant form in the cells of the middle fruit wall of the
various species studied. Here it can be recognized best in thin hand
sections stained with Safranin and Methyl Green in the form of rosettes
almost filling the cell lumina. When the fruits fall to the ground and
subsequently break open their walls, the organism probably makes its
way from the infected cells into the soil where it spreads through wide
areas infecting the roots and stems of other Myricas and producing char-
acteristic lesions.

Taxonomy of the Myricaceae of the Eastern United States

Myricaceae (Lindley) Bayberry Family

Myricaceae, Lindley, Nat. Syst., ed. 2, 1836, p. 179;

Bentham and Hooker, Genera Plantarum, vol. 3, p. 400;
Gas. de Candolle, Prodromus, vol. 16, 2d p., p. 147.
Engler, Pflanzenfamilien, t 3, p. 26.

390 YOUNGKEN— ON THE MYRICACEAE

Myriceae, L. C. Richard, Anal. Fr., 1811, p. 493;

H. Baillon Hist. Plant., t. 6, p. 245.
Galeaceae, Bubani, Fl. Pyr., 2, 1899, p. 49.

Dioecious or sometimes monoecious, aromatic, resinous shrubs or
trees with watery juice and possessing underground branches which
arch downward then upward producing many suckers. Roots fibrous
and bearing many short rootlets upon which are frequently found coral-
loid clusters of tubercles containing the Actinomyces myricarum, Young-
ken. Leaves alternate, revolute in vernation, serrate, irregularly dentate,
lobed or entire, rarely pinnatifid, pinnately and reticulately veined,
pellucid punctate, evergreen or deciduous, generally exstipulate, rarely
stipulate. Flowers naked, unisexual, monoecious or dioecious, in the
axils of unisexual or androgynous aments from scaly buds formed in
the summer in the axils of the leaves of the year, remaining covered
during the winter and opening in March or April before or with the
unfolding of the leaves of the year. Staminate flowers in elongated
catkins, each consisting of 2-8 stamens inserted on the torus like base
of the oval to oval-lanceolate bracts of the catkin, usually subtended by
2 or 4 or rarely by numerous bracteoles; filaments short or elongated,
filiform, free or connate at the base into a short stipe; anthers ovoid, erect,
2-celled, extrose, showing longitudinal dehiscence. Pistillate flowers
in ovoid or ovoid-globular catkins. Gynoecium of two united carpels on
a bract. Ovary sessile, unicellular, subtended by two lateral bracteoles
which persist under the fruit or by 8 linear subulate bracteoles, accrescent
and forming a laciniate involucre inclosing the fruit; styles short and
dividing into 2 elongated style arms which bear stigmatic surfaces on
their inner face; ovule orthotropous, solitary, with a basilar placenta
and superior micropyle. Fruit an akene or ceriferous nut. Pericarp
covered with glandular emergences which secrete wax or fleshy emer-
gences, smooth and lustrous or smooth, glandular. Seed erect, exal-
buminous, covered with a thin testa. Embryo straight, cotyledons
thick, plano-convex; radicle short, superior.

There are two distinct genera of this family, eg. Myrica and Comp-
tonia.

Characters of the Genera and Species

Genus: Myrica, L., Species Plantarum, 1024 (1753)

Mostly dioecious, evergreen, sub-evergreen or deciduous aromatic
shrubs or trees with entire, dentate, or lobed glandular-dotted, exstipu-
late leaves; scales surrounding the ovary generally 2-4 very short. Sta-
mens usually 2-6 in the axil of each bract. Ovary covered with ceriferous

OF THE EASTERN UNITED STATES 391

glands or fleshy non-ceriferous emergences with bracteoles adherent
or non-adherent. Fruit an akene or ceriferous nut.

M. Gale, L. — Leaves membranous, fragrant, deciduous, petiolate, ser-
rate near the summit; catkins borne on special deciduous branches; ovary
flanked by 2 entire bracteoles which develop into aeriferous floats;
wood consisting largely of tracheids.

M. cerifera, L.— Leaves coriaceous, fragrant, evergreen, petiolate,
generally lanceolate-cuneate, covered on both surfaces with numerous
orange-red and golden-yellow glands; fruits aceriferous at apex, 2-3
mm. in diameter.

.1/. Carolinensis, Miller— Leaves membranous, fragrant, deciduous,
petiolate, elliptic-obovate, bearing numerous golden-yellow glands on
both surfaces or glands absent on upper surface. Fruits entirely cer-
iferous at apex and 3.5 mm. — 4.5 mm. in diameter.

M. Macfarlanei, x Youngken.— Leaves subcoriaceous, subevergreen,
fragrant elliptic-obovate to lanceolate-cuneate, bearing a few orange-
red and golden-yellow glands on their upper surface, or golden yellow
glands only and numerous glands of both kinds on the lower surface.
Fruits somewhat punctate at the apex and 3-4 mm. in diameter.

M. inodora, Bartram.— Leaves coriaceous, evergreen, petiolate
generally oblong-obovate, inodorous. Rachis of the pistillate catkins
glabrous. Fruits 5-7 mm. in diameter.

M. cerifera var pumila, Michaux. — Leaves coriaceous, evergreen,
fragrant, obovate to linear-oblanceolate, sessile, or with a petiole
1-2 mm. long. Fruits subglobose, 3.5-4 mm. in diameter.

Genus: Comptonia, Banks, Gaertner. Fr. and Sam. 2:58 pi. 90

(1791).

Monoecious or dioecious aromatic shrubs with pinnatifid, stipulate,
membranous leaves. Stamens commonly 4, occasionally 3-5, filaments
free, short and not accompanied by bracteoles. Female flowers constitu-
ted by a bud carrying an ovary at its summit, flanked laterally by two
bracteoles which develop at the base of their ventral surface and later
on their margins small laciniate accrescent outgrowths. Fruit an
elongated akene, surrounded until it falls from the catkin by a laciniate
cupule formed by the development of two lateral bracteoles and the out-
growth which they have produced. Spikes of fruits globular, spiny
and bur like. Edpidermis of fruit composed of stone cells of a shining
black at maturity. Of this genus there is but one species, Comptonia
asplenijolia, (L) Alton.

392 YOUNGKEX— ON THE MYRICACEAE

For other and more detailed characters of the various species indi-
genous to the eastern United States see chapters on Comparative Mor-
phology.

Comparative Distribution

Myrica cerifera L.^ grows in brackish marshes, in sandy soil on the
border of brackish ponds, estuaries and near the sea from as far north
as the Tuckahoe River, N. J., through Southern Jersey, Maryland,
Virginia, as far south as Southern Florida, west through the Gulf States
to the shores of Arkansas Bay in Texas and northward in the region west
of the Mississippi to the valley of he Washita River in Arkansas. Dr.
Macfarlane and myself collected it along the banks of the Tuckahoe
River, around the marshes and ponds bordering Petersburg road, at
Palermo, Rio Grande, Wildwood, Cape May and Ocean City, N. J.

Myrica cerifera pumila (Michaux)^ is found in low, sandy Pine Barren
regions of South Carolina, Georgia and Florida, and on dry sand, and
hills in Northern Alabama, Eastern Texas, Northern Louisaina and
Southern Arkansas.

Myrica Carolinensis (Miller)^ thrives in sandy, or sterile soil chiefly
near the coast, but also in inland swamps and pine forests, from Labrador
to Florida and Louisiana, on the border of the Great Lakes and in Indiana.
We collected it at Noroton, Connecticut and at Tuckahoe, Clementon,
Albion, Palermo, Ocean City, Rio Grande, Wildwood, Wildwood Junc-
tion, and Petersburg in N. J.

Myrica Macfarlanei (Youngken) {M. cerifera x M. Carolinensis)
grows in the sandly soil of swamps, and on pine and holly forests near
the sea in Southern New Jersey. W^e have found it at Palermo, Rio
Grande, Wildwood, Tuckahoe, Cape May and Ocean City, wherever
both of its parents — M. cerifera L. and M. Carolinensis Miller — abound.

Myrica Gale (L)^ is widely distributed through northern regions, from
Labrador and Newfoundland as far south as Warren County, New
Jersey, from the Atlantic to the Pacific, and in eastern mountain regions
to Virginia.

It was collected in 1913 on the border of back waters occasionally
communicating with the sea at Peaks Island and Chebeague in Casco
Bay, Maine, and in 1899 on Martha's Vineyard, Massachusetts by Dr.
John M. Macfarlane.

Myrica inodora (W. Bartram)^ is found in deep swamps in the vicinity
of Mobile and Stockton, Alabama, near Poplarville, Mississippi in the
valley of the Pearl River, and near Appalachicola, Florida.

OF THE EASTERN UNITED STATES 393

Comptonia asplenifolia L. (Aiton)^ thrives in dry, sterile soil and is
widely distributed from Nova Scotia to Saskatchewan, and southward
to North Carolina and Tennessee. We have found it forming a large
part of the underbush associated with Pteris aquilina near Clementon,
Albion, Palermo and Wildwood Junction, N. J., along Mainville Road,
near Mainville, and at Strafford, Penna.

Summary

1. Along the eastern seaboard of the United States there occur five
good species of Myricaceae and a hybrid between two of these.

2. Myrica cerifera varies from a low shrub to a tree 12. m. high and
extends northward, contrary to past statements, as far as Tuckahoe,
New Jersey. The author finds this species to be evergreen and wholly
confined to coastal regions within sight of the sea.

3. Myrica Carolinensis, with which the previous named species has
often been confounded, is strictly deciduous except when strong basal
shoots are formed. The leaves on these may be sub-evergreen.

4. The lanceolate leaves of Myrica cerifera drop without assuming a
copper red color. The elliptic obovate leaves of Myrica Carolinensis
assume a greenish-brown hue on an extensive scale in October and
November previous to leaf fall. This species is of wide distribution
along the coastal plain and even ascends to 1200 feet at Mount Desert.

5. Hybrids between the two above species (Myrica Macfarlanei,
Youngken) are frequent where both parents abound. The leaf charac-
ters of this are averagely intermediate in duration, shape, thickness
and coloration between the parents, M. cerifera, L. and Myrica Carolinen-
sis, Miller.

6. Comptonia asplenifolia is distributed from Nova Scotia to Saskat-
chewan and southward to North Carolina and Tennessee. Its leaves are
strictly deciduous.

7. Myrica Gale is distributed through northern regions, from Labra-
dor and Newfoundland as far south as Warren County, New Jersey,
from the Atlantic to the Pacific and in eastern mountain regions to
Virginia. Its leaves are strictly deciduous

8. Myrica inodora, from the statements of authors, is evergreen.
In height and aspect it resembles M. cerifera.

9. Seedlings are here for the first time described and figured from the
cotyledonary stage onward for M. cerifera, M. Carolinensis and M. Mac-
farlanei. Their comparative morphology has been traced.

394 YOUNGKEN— ON THE MYRICACEAE

10. The author shows that from the seedhng primary root of five to
six months growth and from thence onward characteristic root tubercles
are formed in the above named two species and their hybrid. The or-
ganism is found in all these to be an Actinomyces (here first described
as Actinomyces myricarum of the author) that abundantly fills infested
cells in the cortex of the tubercles, which owe their origin as arrested
and modified roots to its irritant and invading action. As a result
of cultures made from tubercles, the author concludes that good cultures
of the organism can be secured on nutrient agar.

11. Since Actinomyces is frequently a virulent pathogenic organism
in cattle and other domestic animals up to man, the author suggests
the possible pathogenic relation of the Myrica organism to such animals.
He, therefore, would regard the infesting organism as parasitic in relation.

12. In Comptonia asplenifolia, a similar Actinomyces organism is the
primary infecting agent, but there often appears a mycelium producing
fungus probably belonging to the Oomycetes.

13. As for the roots so for the stems of M. cerifera, M. Carolinensis
and the hybrid, a careful comparative histological study has been made
and details recorded as to the resemblances and differences of the hybrid
and its parents. During this study obUque barred septa have been
discovered separating the pitted vessels from each other.

14. Coccus-hke forms that the writer believes to be involution forms
of the infesting Actinomyces have been discovered in the cavities of the
pitted vessels. Such seem to indicate the pathway of infestment taken
by the Actinomyces in order ultimately to reach the fruit wall.

15. In the study of the leaves new structural details have been
observed, but special interest attaches to the presence now recorded of
orange-red bowl to saucer-shaped glandular hairs specifically character-
istic of M. cerifera and in a reduced degree of M. Macfarlanei intermingled
with golden-yellow glandular hairs of Chevalier. The latter only are
present in M. Carolinensis. The hybrid, moreover, in general histology
is shown to be more or less intermediate between the parents.

16. The last named conclusion regarding the hybrid and its parentage
is further verified by comparative studies recorded for the stem and root.

17. Exact phytophenological records have been made as to the
maturation of the floral parts and the period of blossoming in April and
May.

Spore mother cell formation is completed by autumn of one year,
but formation of tetrads proceeds in the dififerent species studied from
mid April to mid May in the Philadelphia neighborhood.

OF THE EASTERN UNITED STATES 395

18. Successive stages in the formation of the fruits have been ob-
served and studied histologically. A highly interesting detail here is the
presence of the Actinomyces organism in beautiful radiate patches that
fill many of the cells of the mid-fruit layer.

19. The structure, mode of origin and wax secretion of the knob-
shaped glands covering the ovarian wall have been traced and the inter-
mediate character of the fruits of the hybrid, as compared with those
of the two parents above mentioned, has been demonstrated.

New and more exact diagnostic taxonomic descriptions than have
hitherto been submitted by authors are presented above and in particular,
the diagnostic characters of M. cerifera, M. CaroUnensis and their hybrid
have been fullv elucidated.

SYNONYMS

For the convenience of reference, the writer presents the following table of techni-
cal and common names of the species considered together with their bibliography.

Myrica cerifera, L. (Species Plantarum, 1753, p. 1024).

Ligustrum americanum, lauri folio, Plumier, Descrip. pi. Americ. 1693; Toume-
fort, Institut., p. 597.

Myrtiis hrahanticae similis carolinensis baccifera, fructu racemoso semili mono-
pyreno, Plukenet Almagestum, 1696, p. 250, t. 48, f. 9; Catesby, Carol 1, p. 69, t. 69.

Lychnochrodyopkoros, Plukenet, Phytogr, tabl. 48, fig. 9.

Myrica cerifera, IMiller, Gardener's Diet. Ed. 8; Sargent, Silva of N. Am., Vol. 9,
p. 87 and plate 459; Britton and Brown's Illus. Flora of the U. S. and Canada, Ed. 2,
vol. l,p. 585.

Myrica cerifera arhorescens, Castiglioni, Viag. negh Stati-Uniti 1790, vol. 2, p.
302; Michaux, FI. Bor. Am., vol. 2, p. 228.

Lacistema Berterianum, Schult., Mant., 1822, vol. 1, p. 66.

Lacistema alterum, Spreng. Syst., vol. 1, 1825, p. 124.

Myrica helerophylla, Rafinesque, Alsograph. Am. 9 (1838).

Ceropliora lanceolata, Rafinesque, Alsograph. Am. 11 (1838).

Myrica Carolinensis, A. Richard, Fl. Cub. vol. 3, p. 231 (1853).

Myrica microcarpa, Grisebach, Flor. Brit. W. Indies, p. 177 (1864).

Myrica altera, C. de Candolle, Prodr. 16, pt. 2, p. 595 (1868).

Morella cerifera, Small, Flora of S. E. U. S. ed. 2, p. 337 (1913).

Candleberry, Candel-wood, Candle Tree (Amer. sept., Plukenet 1. c, p. 250).

Bayberry, Wax Myrtle, (Beringer, Notes on the genus Myrica, Am. Jour. Phar.,
May 1894).

Waxberry, Tallow-bayberry, Wax-myrtle, Candleberry, Tallow-shurb, Sweet Oak,
Candleberry-lSIyrtle (Britton and Brown, Illus. Flora of U. S. and Canada, Ed. 2,
vol. 1, p. 585 (1913).

Myrica Carolinensis, Miller (Diet. Ed. 8, 1768, n 3).

Myrtiis brabanticae similis carolinensis humilior foliis latioribus magis serratis,

396 YOUNGKEN— ON THE MYRICACEAE

Catesby, Hist. Carol., p. 13.

Gale amerkana humilis fruclitr coriandri, Burman, Herb.

Myrica pensylvanica, Loiseleur Deslongchamps in Nouv. Duhamel, vol. 2, 1802.
p. 190, Chevalier Monograph Myricaceae, Mem. de la Soc. Nat. des. Sc. Nat. et. Math,
de Cherbourg, vol. 32, p. 266.

Myrica cerifera B. media, Chapman, Fl. p. 427.

Myrica sessilifolia, Rafinesque, Alsog., 1836, p. 10.

Morella Carolinensis, Small, Fl. S. E. U. S., ed. 2, p. 337 (1913).

Small \Vaxberr\-, Bayberry, Britton and Brown lUus. Flor. of North. States and
Canada. Ed. 2, vol. 1, p. 585.

Myrica cerifera var pumila, Michaux (Fl. Am. Bor., 1803, vol. 2, p. 228).

Willd. Sp. 4, p. 746.

M. sessifolia, Rafinesque, Alsograph. Am., 10 (1838).

M. pusilla, Rafinesque, 1. c, 10.

Morella pumila, Small, F. S. E. U. S. ed. 2, p. 337 (1913).

Myrica inodora, Bartram, Trav., ed. 2 (1794), p. 403.

Myrica obovaia, Chamisso ex H. Kew. in C. DC. 1. c, p. 150.

Myrica lanreola, Trecul ex H. Paris in C. DC, l.c ., p. 154.

Myrica Torreyana, Chapm ex H. Delessert.

Cerophora inodora, Rafinesque, Alsograph Am. 2 (1838).

Wax Myrtle, Sargent, Silva of N. A., Vol. 9, p. 91.

MyricaGalc, L. Spec. Plant, 1024 (1753).

Myrica pahistris, Lamk., Fl. Franc, vol. 2, p. 236 (1778).

Myrica brabantica, J. E. Gray, Nat. arr. Brit. PI. vol. 2, p. 249, (1821).

Gale bclgica, Dumort., Fl. Belg., p. 12 (1827).

Cerophora anguslifolia, Rafinesque., Alsog. Am. vol. 2, (1838).

Cerophora spicans, Rafinesque, 1. c, 12, (1838).

Gale uliginosa, Spach. Hist, veg., vol. 2, p. 259, (1824) Bubani, Fl. Pyr., 1, p. 49.

Gale palustris, Chevalier, Monograph des Myricacees, Mem. d. e. Soc. Nat. des
Sciences Nat. et Math, de Cherbourg, vol. 32, p. 177.

Gagel (of Old Germans).

Piemont, Piment royal (Centre de la France, Franchet).

Sweet Gale (Michaux).

Meadow Fern, Bog Myrtle, Dutch Myrtle, Willow Myrtle, Bay Bush (Beringer)
Fern or Scotch-Gale, Sweet Willow, Golden osier. Moor-, bog-, Dutch- or Burton-
myrtle (Britton and Brown, Illust. Flora of U. S. and Canada, vol. 1, p. 584).

Comptonia asplcnifolia, Alton, Hort. Kew. 334 (1789).

Myrica Comptonia, C. de Candolle, I.e., 151 (1864).

Myrica Gale Bentham, PI. Hartweg., 336 (1857).

Liqtiidambar asplenifolia, Linnaeus, Syst. ed. 10, 1273 (1759).

Liquidambar peregrina, Linnaeus. Spec. 999 (1753).

Comptonia peregrina, Coulter, Mem. Torr. Club 5: 127, 1894, Chevalier Monog.
des Myricacees, Mem. d. 1, Soc. Nat. des Sciences Nat. et Math, de Cherbourg., vol.
32, p. 193 (1902).

Myrica peregrina, O. Kze. Rev. Gen. PI. (1891), p. 638.

Comptonia Ceterach, Mirb. in Duham. Arb ed. nov. 2, t 2.

Gale mariana, asplenii folio, J. Petiver, Mus. Petiv. 1695), p. 773.

Myrti hrahanticae americana, foliorum laciniis asplenii modo divisis, Plukenet.

OF THE EASTERN UNITED STATES 397

Almagest (1696) p. 250, t 100, f. 6, 7.

Sweet Fern, Sweet Ferrj-, Sweet Bush, Fern Gale, Spleenwort Bush, Beringer, Am.
Jour. Pharm., May 1894, p. 222.

Meadow or Shrubby Fern, Canada Sweet Gale, Britton and Brown, lUust. Flora
of North. U. States and Canada, ed. vol. 1, p. 586 (1913).

SUPPLEMENT.VRY BIBLIOGRAPHY

In addition to the literature cited in the chapters on historical review and
synonyms, the following references are appended.

1. Linnaeus: Species Plantarum, p. 1024 (1753).

2. Michau.x: Fl. Bor. Am. 2, p. 228 (1803).

3. Miller: Gardener's Dictionary, vol. 2, pt. 1 (1807).

4. Linnaeus: Species Plantarum, 2: p. 1453 (1763).

5. W. Bartram: Travels, 2d ed., p. 403 (1794).

6. Aiton: Hortus Kewensis, p. 334 (1789).

7. Chevalier: Monograph, des Myricacees, Memoires de la Societe Nationale des
Sciences Naturelles et Mathematiques de Cherbourg 32 (1902).

8. Tison: Recherches sur la chute des feuilles chez les dicotyledones, Caen, p. 42
(1900).

9. Solereder's Systematic Anatomy of the Dicotyledons, vol. 2, p. 785 (1908).

10. Small: Flora of the Southeastern U. S., 2d ed., p. 337 (1913).

11. Beringer: Notes on the genus Myrica, Amer. Jour. Pharm, May 1894, p. 220.

12. Gris: Nouv. Arch. Mus. d'hist. nat., t. 6, p. 284 (1870).

13. Brunchorst: Die Structur d. Inhaltskorper in d. Zellen einiger Wurelangschwell-
ungen, Bergens Museums Aarber, p. 233 and plate 21.

14. Moeller: Beitrag zur Kenntnis der Frankia sublilis Brunchorst, Berichte der
deutschen Bot. Gesellschaft, p. 215-224 (1890).

15. Shibata: Die Wurzelanschwellungen von Alnus und Myrica in Cytologische Stu-
dien iiber die endotrophen Mykorrhizen, Jahrbiicher fur wissenchaft. Botanik, Bd.
37, p. 668-670.

16. Harshberger: The form and structure of the mycodomatia of M. cerifera, Proc.
Acad. Nat. Sci. of Phila., 55: 352-361.

17. Arzberger: The fungus root tubercles of Ceanolhus Americana, Elaeagnus argen'xa
and Myrica cerifera, Missouri Botanical Garden Report, p. 82 (1910).

18. Jordan: General Bacteriology, p. 404-413 (1908).

19. Wright: Pub. Mass. Genl. Hosp., Boston 1905; Journal Med. Res., 8, p. 349 (1905).

20. Wolff and Israel: Arch. f. path. Anat., 126, p. 11 (1891).

398 YOUNGKEN— ON THE MYRICACEAE

List of Illustrations

Plate LXXXI

Fig. 1 Seedlings of Myrica ccrijera, showing root tubercles, collected at Wildwood,

N. J., June 20, 1913.
Fig. 2 Seedlings of Myrica Caroltnensis, also possessing root tubercles and collected

on same date and at same place as above.
Fig. 3 Seedlings of Myrica Macfarlanei {M. cerifera x M. Carolinensis) gathered

at Palermo, N. J., March 21, 1915. Note the position and frequency of the

tubercles.

Plate LXXXII

Fig . 4 Entire plant of Comptonia asplenifolia, showing habit of growth. Collected

near Clementon, N. J., July 15, 1913.
Fig. 5 Leafy branch of Comptonia asplenifolia, showing nature of stipulate leaves

in mid-summer.

Plate LXXXIII

Fig. 6 Fructiferous branches of Myrica cerifera, showing evergreen leaves persisting
on the branches all winter. Collected at Tuckahoe, N. J., January 2, 1914.

Fig. 7 Fruiting branch of Myrica Carolinensis, collected as above, showing but one
leaf present and about ready to fall.

Plate LXXXIV

Fig. 8 Fructiferous branch of Myrica Gale, showing mature pistillate catkins and

nature of leaves. Collected by Dr. John M. Macfarlane at the south end

of Peak's Island in Casco Bay, Maine, September 1913.
Fig. 9 Staminate branch of the same plant, collected at the south end of Chebeague,

in Casco Bay, Maine, by Dr. J. M. Macfarlane, August 1913.
Fig. 10 Branches of pistillate plant of Myrica Carolinensis, growing near the coast

of Long Island at Noroton, Connecticut, April 27, 1913. Note total absence

of leaves.
Fig. 11 Branches of staminate plant about 2)4 feet high, growing near the sandy

beach of Long Island shore, Noroton, Conn., April 27, 1913. Note total

absence of leaves.

Plate LXXXV

Fig. 12 Staminate branch of Myrica Macfarlanei, collected at Wildwood, N. J.,

January 31, 1915. The leaves were nearly all discolored and somewhat

shrivelled.
Fig. 13 Pistillate branch of Myrica Macfarlanei, collected as above. Note that many

of the fruits are still adhering to the catkin axes on the second year's wood.

Many of the fruit bearing axes of several seasons past are still adhering to

the wood.

OF THE EASTERN UNITED STATES 399

Plate LXXXVI

Fig. 14 Transverse section of the root of Myrica Carolinensis x 25.
Pig 15 " " " " " " Myrica Macfarlanei.

Fig 16 " " " " " " -^/vnVa cerifera x 13.5.

Fig 17 " " " " " " Myrica Gale x 25.

Fig. 18 " .. » » y >' Complonia as pi eni folia.

Plate LXXXVI I
Fig. 19 Transverse section aerial stem of Myrica cerifera x 13.5.
Fig. 20 " " " " " " Myrica Macfarlanei x 13.5.

Fig. 21 " " " " " " Myrica Carolinensis x 13.5.

Fig. 22 Transverse section of aerial stem of Complonia asplenifolia x 13.5.
Fig. 23 " " " " " " Myica Gale x 25.8.

Plate LXXXVTTI

Fig. 24 Transverse section through a tubercular root of M. Carolinensis, showing

(a) the region of infestnient of Actinomyces Myricarum.
Fig. 25 Longitudinal section through a tubercular root of M. Carolinensis, showing

its nature of branching and the appearances of the h>-pertrophied cells

containing Actinomyces rosettes.
Fig. 26 Section through an old tubercular root of Complonia asplenifolia, showing

the presence of a mycelium producing fungus having non-septate h^-phae

and probably belonging to the Oomycetes x 160.
Fig. 27 A portion of the above section enlarged and showing the hyphae of the

fungus curled up in different cells, while other cells contain broken down

contents which probably supply the necessary food for the organism. X 600.

Plate LXXXIX

Fig. 28 Longitudinal section through a staminate catkin of M. Carolinensis. X 25 .

Fig. 29 Longitudinal section through a pistillate catkin of the same species. X 25-

Fig. 30. Longitudinal section through a staminate catkin of Complonia asplenifolia.
X 10.

Fig. 31 Longitudmal section through a portion of the catkin of Complonia aspleni-
folia, showing the anthers in the pollen mother stage.

400 YOUNGKEN— ON THE MYRICACEAE

Plate XC

Fig. 32 M. Macfarlanei. Entire plant, growing in moist sandy soil near Palermo
Station, Palermo, N. J. and collected March 21, 1915. The leaves present
had all assumed a copper red aspect, some had fallen, and the rest were
almost ready to drop.

Fig. 33 Tubercular clusters on underground stem and roots of M. Macfarlanei
observed by the author at North Wildwood, N. J., January 31, 1915.

Bol. C 0)1 /rib. Univ. Peiiii.

Vol. IV, Plate LXXXI

.h

f

• >' y

-•rn.1 r

Fig 1 (below), 2 (above)

mv:^mr

Fig. 3 .
YOUNGKEN ON MYRICACEAE

Bot. Contrib. Univ. Penn.

Vol. IV, Plate LXXXII

he

YOUNGKEN ON MYRICACEAE

Bot. Contrih. Univ. Penn.

Vol. IV, Plate LXXXTII

tc

■E

<

<
U
I— <
^^

z
o

o

&

o

Bot. Coutrib. Univ. Pcnn.

Fig. 8

\'ol. IV, Plate LXXXIV

Fig. 10

Fig. 11

Fig. 9

YOUNGKEN ON MYRICACEAE

Bot. Contrib. Univ. Penu.

Vol. IV, Plate LXXXV

M

w
<:

u
<
u

be

o

w

o

:?

o

Bot. Contrih. Univ. Peun.

Vol. IV, Plate LXXXVI

Fig. 17 Fig. 18

YOUNGKEN ON MYRICACEAE

Boi. Conlrib. Univ. Penn.

Vol. IV, Plate LXXXVII

Fu. 19

Fig. 20

Fig. 21

Fig 22

Fig. 23

YOUNGKEN ON MYRICACEAE

Bot. Contrib. Univ. Peitn.

Vol. IV, Plate LXXXVIII

Fig. 24

Fig. 25

L/r

Fig. 26

YOUNGKEN ON MYRICACEAE

Fig. 21

Bot. CorUrib. Univ. Pcnn.

Vol. IV, PlateLXXXIX

/gS^

Fig. 28

Fig. 30

*QI^-^'

Fig. 29

,.< ^ •' • ■ ,> / ■ >, . '. .S..«;*;i5i!R'„

'::-^:^^§^^m

Fig. 31

YOUNGKEN ON MYRICACEAE

Bot. Contrib. Univ. Penn.

Vol. TV, Plate XC

FiK. :->!

Fig. yo

YOUNGKEN ON MYRICACEAE

A COMPARATRT STUDY OF FLOERKEA PROSERFIXACOIDES

AND ALLIES

BY

Alice M. Russell, B.S., M.S.
With Plates XCI, XCII

CONTENTS

I. Introduction and Historical Review 401

II. Synopsis of Family Limnanthaceae 403

III. Distribution of Species 404

IV. Morphological and Anatomical Study of Floerkea and Limnanthes 406

V. Conclusions 413

VI. Summary 415

VII. Bibhography 416

VIII. Description of Plates 417

Introduction and Historical Re\^ew

The purpose of this paper is to verify the affinities of Floerkea proser-
pinacoides in relation to other plant types.

Floerkea proserpinacoides was first described by Willdenow in 1801
in the "Neue Schriften der Gesellschaft der naturforschenden Freunde"
(29). He did not give it a definite position in the natural orders of
plants. In Barton's Compendium Florae Philadelphicae (30), it is placed
in the division Hexandria Monogynia.

In 1833 Robert Brown described Limnanthes Douglasii before the
Linnaean Society in June of that year. The abstract of the paper,
quoted from the London, Edinburgh and Dublin Philosophical Maga-
zine of General Science for July-December 1833, (31) is as follows: —
"June 18, a paper was read, entitled, 'Characters and Descriptions of
Limnanthes, a new genus of plants allied to Floerkea, ' by Robert Brown,
Esq. V.P.L.S. For specimens of the plant described, the writer is in-
debted to the Horticultural Society, and to Mr. David Douglas, F.L.S.,
by whom it was recently discovered in California.

"Mr. Brown was led more particularly to examine Limnanthes, from
its resemblance to Floerkea of Willdenow, a genus which he had many

402 RUSSELL— A COMPARATIVE STUDY OF

years since investigated without being able to determine its place in the
natural system. Examination proved these two plants to be so nearly
akin, that they might perhaps be included in the same genus. They
are here, however, separated and the two genera are considered as form-
ing a family distinct from all those at present known.

"The place of the new family (Limnantheae) is not absolutely deter-
mined, but it is suggested that in two remarkable points of its structure,
namely, the presence of glands subtending the alternate filaments and
the existence of a gynobase it more nearly approaches to h^pog^nous
famiUes than to perigynous with which it has hitherto been associated.

"The following are the characters of the natural order, and of the two
genera forming it: —

Limnantheae

Flos completus, regularis. Calyx 3-5 partitus, aestivatione valvata
persistens. Petala 3-5, marcescentia. Stamina 6-10 insertione ambigua
(hypo-perigyna) marcescentia. Filamenta distincta, 3-5 sepaUs opposita
basi extus glandula munita. Ovaria 2-5, sepalis opposita, cum
stylo communi 2-5- fido mediante gynobasi connexa, monosperma,
ovulo erecto, nucleo inverso. Achenia subcarnosa. Semen exal-
buminosum. Embryo rectus; radicula infera. Herbae (Americae sep-
tentrionalis, paludosae) glaberrimae, alternifoliae exstipulatae, foliis
divisis; pedunculis unifloris ebracteatis, apice dilatato basin turbinatum
calycis simulante.

Limnanthes

Calyx 5 partitus. Petala 5, calyce longiora, aestivatione contorta.
Stamina 10, ovaria 5. Herba {Limnanthes Douglas! I Americae occi-
dental! boreahs.) foliis bipinnatifidis, pinnis sub-oppositis segmentis
alternis.

Floerkea, Willdenow

Calyx 3 partitus. Petala 3 calyce breviora. Stamina 6. Ovaria 2
(raro 3) Herba (Americae orientali boreaUs). Foliis pinnatifidis seg-
mentis indivisis. "

Since the time when the above was written, it has been variously
placed, some including it under the Geraniaceae as a sub -order Lim-
nantheae (1). Others directly absorb it into the Geraniaceae (5). Recent
authors accept Robert Brown's views, keeping it as a separate family (1).
Some authors consider the two genera to be so closely allied as to com-
pose one genus only; naming it from the first described genus of the
family Floerkea (5).

FLOERKEA PROSERPINACOIDES AND ALLIES 403

Synopsis of Family Limnanthaceae

The family Limnanthaceae, as recently described by Axel Rydberg
in North American Flora, Vol. 25, part 2, includes: —

I. Limnanthes

1. Limnanthes alba Hartw; Benth. PL Hartw. 301 (1848).

Floerkea alba Greene, Fl. Fran. 100 (1891).

2. Limnanthes floccosa Howell, Fl. N. W. Am. 1 :108 (1897).

3. Limnanthes rosea Hartw; Benth. PI. Hartw. 302 (1848).

L. pulchella Hartw. Jour. Hort. Soc. London 4:79 (1849)

as a synon}TTi.

F. rosea Greene Fl. Fr. 100, (1891).

4. Limnanthes Douglasii, R. Br. Phil. Mag. Ill 2:40, (1833).

F. Douglasii, Baillon, Adansonia 10, (1873).

5. Limnanthes sulphurea, Loud. Encyc. PI. (1843).

L. Douglasii Sweet, Br. Flower Gard. 2. (1838).

6. Limnanthes versicolor (Greene) Rydberg.

F. versicolor, Greene, Erythrea 3:62. (1895).

7. Limnanthes gracilis, Howell, Fl. N. W. Am. 1, 108, (1897).

8. Limnanthes pumila, Howell, Fl. N. W. Am. 1, (1897).

9. Limnanthes Macoiinii, Trelease Mem. Boston Soc. Nat. Hist.

4:85, (1887).

L. Douglasii Macoun Cat. Can. PI. 3.502, (1886.)

F. Macounii, Trelease in A. Gray Syn. Fl. 1, (1897).

II. Floerkea

1. Floerkea occidentalis, Rydb. Mem. N. Y. Bot. Gard. 1.268, (1900).

F. proserpinacoides of S. Wat. Bot. King's Expl. 50, (1871).

2. Floerkea proserpinacoides, Willd. Neue Schr. Ges. Nat. Freunde.

Berlin, 3.449, (1801).

F. lacustris, Pers. 3 yn. pi. 1, (1805).

Nectris pinnata, Pursch, Fl. Am. (1814).

F. uliginosa, Muhl. Cat. 36, (1813).

F. palustris, Nutt. Gen. 1, 229, (1818).

Cabomba pinnata R. & S. Syst. Veg. 7, (1830).

Thos. Howell (2) has the above species of Limnanthes recorded ex-
cepting L. sulphurea. He does not recognise F. occidentalis. Asa Gray
(5) accepts F. proserpinacoides. He places Limnanthes and Floerkea in
the same genus, Floerkea. The species accepted by him are F. Macounii

404 RUSSELL— A COMPARATIVE STUDY OF

F. Douglasii, F. rosea, and F. alba. L. sulphurea he considers synonymous
with L. Douglasii. According to him L. pumila and L. versicolor are
varietal forms of L. Douglasii.

The most stable species of the two genera composing this family
(Limnanthaceae), accepted generally by all, are as follows:

LlMNANTHACEAE

1 Floerkea

(a) Floerkea proserpinacoides.
var. F. occidentalis.

2 Limnanthes.

(a) Limnanthes Douglasii.

var. L. sulphurea.
var. L. pumila.
var. L. versicolor.

(b) Limnanthes rosea.

(c) Limanthes alba.

(d) Limnanthes Macounii.

Distribution or Species

It will be seen from Robert Brown's key to the family that the two
distinctions between the genera are that Limnanthes is a form with
pentamerous symmetry, and is western in distribution, while Floerkea
is a trimerous form, given as purely eastern in distribution. At that
time L. Douglasii was the only species of Limnanthes known. There
are, however, several other pentamerous forms, having a distribution
identical with that of L. Douglasii, which have been described more
recently. These pentamerous forms of Limnanthes have a purely western
distribution — viz. California, Oregon, Washington (1).

There is a form of Limnanthes showing tetramerous symmetry de-
scribed as Limnanthes Macounii (Trel. Mem. Boston Soc. Nat. Hist. 4,
85, 1887). Its distribution is restricted apparently to Vancouver Island,
British Columbia (5) (1). But the discovery of a tetramerous form of
Limnanthes is noteworthy, forming as it does a link connecting the pen-
tamerous western forms with the trimerous Floerkea.

There is a species of Floerkea described by Rydberg as Floerkea
occidentalis, resembling Floerkea proserpinacoides in all essentials, dif-
fering from it by a slight reduction in size of its parts as compared with
F. proserpinacoides. It is probably a starved, or feeble form of F. pro-

FLOERKEA PROSERPINACOIDES AND ALLIES 405

serpinacoides of varietal rank. Some authors consider it as identical
with F. proserpinacoides (12) (13).

Floerkea occidentalis is reported from Yellowstone Park, Utah, and
Washington, ascending to an altitude of 2,000 to 2,500 m. (16); Wyoming,
Colorado to California and Washington (1).

Those who consider F. occidentalis to be synon\Tnous with F. pro-
serpinacoides record F. proserpinacoides from: —

Washington, Ontario, south to California, Utah and Pennsylvania
(12); Canada, Oregon, south in the east to Pennsylvania and Illinois,
and in the west to California and Utah (5); Quebec to Ontario, Oregon,
Pennsylvania, Tennessee, Missouri to Utah, and California (13); Oregon,
California, Illinois, Canada and New England (2).

Those distinguishing F. proserpinacoides give its distribution as
follows: —

In Washington, North Utah (18); Quebec to New Jersey, Tennessee,
Missouri, Wisconsin (1).

From this it can be seen that F. occidentalis (F. proserpinacoides)
overlaps the distribution regions of the pentamerous forms of Limnanthes
and connects with the distributional area given for the tetramerous form
of Limnanthes (L. Macounii). Thus, these two forms, if they are to be
so considered, F. occidentalis and F. proserpinacoides, have a distribution
ranging through California, Washington, Oregon, extending eastward
through southern Canada to Ontario, Quebec, into New England and
the middle Atlantic States to Pennsylvania. Here they reach, under
the name of Floerkea proserpinacoides, the distribution accepted by the
older authors for it in the East.

The whole group seems to have originated as a pentamerous one which
still retains in L. Douglasii and L. rosea the attractive large flowers,
which, to quote a rather popular Western Flora, make the western
"meadows all a cream" in April. In its northern representatives a
tendency toward reduction in the size of the flowers and the number of
parts appears in the tetramerous species, Limnanthes Macounii. Over-
lapping the distribution of this tetramerous form is that of the trimerous
one, Floerkea occidentalis, a condensed form which is by Piper and others
regarded merely as a reduced type of Floerka proserpinacoides. This
then stretches across the continent from California to Oregon thence to
the Lakes, where a spur of distribution perhaps passed down tributaries
of the Mississippi to Missouri, Illinois, Ohio and Tennessee. In making
its way from Canada down through Quebec, Ontario, New England to
Pennsylvania, the distributional lines may have been along the foothills

406 RUSSELL— A COIMPARATIVE STUDY OF

of the Allegheny Mountains. The indications are that it is not found
on the more recent areas which have become the coastal plains. Stone's
Flora of New Jersey (17) records no locality for Floerkea proserpinacoides.

Morphological and Anatomical Study of Floerkea and Limnanthes
Besides the close relation of the two genera, Limnanthes and Floerkea,
as shown by their distribution, there is an even closer relation revealed
by their structural similarities. This will be shown immediately in the
discussion of the two forms Limnanthes Douglasii and Floerkea proser-
pinacoides, representative of the two genera.

Gennination

Floerkea proserpinacoides shows hypogeal cotyledons and Limnanthes
Douglasii shows epigeal cotyledons. The cotyledons are in both ovate
and thick. Those of Limnanthes Douglasii are peculiar in that they
possess at their tips a very noticeable water stomatic area as is the case
in the leaflets. The cotyledons are green, functioning as leaves, and the
presence of the water stomatic area on the leaf is not surprising since its
habitat is usually an extremely moist locality (Fig. 1).

In Floerkea, a longitudinal section of the cotyledon shows at the tip,
in the region corresponding to that of the water stomatic area in Limnan-
thes Douglasii, a distinct mass of epithem tissue (Fig. 2). This area
can scarcely be considered to function as a water stomatic area in such
colorless subterranean cotyledons. Rather it might be regarded as a
vestigial structure. Its presence, however, is valuable as an indication
of an ancestry of plants possessing epigeal cotyledons equipped with a
water stomatic area at the tip as already traced in Limnanthes Douglasii.

This distinction between Limnanthes and Floerkea may possibly be
broken down by the finding of a series ranging from epigeal through semi-
hypogeal to hypogeal in the relation of their cotyledons. It may be that
a reduced form, such as Limnanthes Macounii, might present this transi-
tion condition. No material for actual observation was obtainable and
no references relating to the cotyledons were found.

Floerkea proserpinacoides this year (1917) germinated in the regions
about Philadelphia in the week of March 21-28. By March 28, the plants
were showing their first leaf in most locaUties. By the foUowdng week,
the petiole of the first leaf had lengthened enormously. The second leaf
was beginning to unfold and the flower buds were visible in the axils of
the unfolding leaves. The first flowers appeared from April 9th on.
The flowers are produced through April, one arising from the axil of each
successive leaf. The present was an exceptional year for Floerkea.

FLOERKEA PROSERPIN ACOIDES AND ALLIES 407

From the end of April to the first week in May, it is usually dying out,
and the seeds are ripening. This spring has, however, been extremely
backward and now (May 25-29) the flowers are still appearing. The
fruits formed from the earlier flowers are still green. The plants are
beginning to show signs of decay, however. The leaves and stems are
becoming withered and yellow. A few days of warm weather will
probably end the unusually long period of growth of Floerkea proser-
pinacoides for this year.

The references made to the blooming season of Limnanthes indicate
that the height of the blossoming period corresponds to that of Floerkea.
April is the time for the appearance of the showy flowered Limnanthes
in the western states.

Root

The root system of Floerkea proserpinacoides is characteristic of a
shortlived reduced umbrophyte, that is, it consists of a few short fibrous
unbranched roots. Besides the normal roots, adventitious roots may
appear at the nodes on the stem, especially when the stem is procum-
bent upon the moist earth. These adventitious roots do not often appear
above the third node. Plants growing upright in habitats more umbro-
phytic than hydrophytic do not show adventitious roots.

Limnanthes Douglasii shows a similar adventitious root formation at
the first or second node.

In order to study microscopically the structure of the root in Floerkea
proserpinacoides and Limnanthes Douglasii, material was imbedded in
paraffin and sectioned. All sections were double stained with safranin
and Delafield's haematoxylin unless other%\'ise stated.

A transverse section of a mature root of Floerkea proserpinacoides
thus prepared (Fig. 3) showed : Externally a bounding layer of rounded
epidermal cells, cuticularized externally. Within the epidermis is a
homogeneous cortical region three to four layers in depth, composed of
large round cells with thin walls. The innermost cortical layer is dif-
ferentiated into a well-marked endodermis of elliptical cells thickened
upon their inner and lateral faces. Within the endodermis is the peri-
cambium, a single layer of large thin-walled cells This layer forms the
outermost part of the vascular strand. The stele consists of a simple
diarch bundle. The xylem plate is uniseriate, being formed of two to
three spiral tracheae at each end, and three to four spirally thickened
vessels centrally placed. The two phloem masses are of small thin- walled
cells, not separable into the various phloem elements.

408 RUSSELL— A CO^NIPARATIVE STUDY OF

In the transverse section of the mature root (Fig. 4) of L. Douglasii,
there appears: Externally an epidermis essentially as in F. proser-
pinacoides. The cortex region beneath is deeper, but is of cells of the
same character. The endodermis and pericambium are as in F. proser-
pinacoides. The vascular strand is diarch, as before, but the xylem
plate is not uniseriate. The spiral tracheae are three in number, at the
outer edges of the xylem plate, much as in F. proserpinacoides. The
spiral vessels occupying the central part of the plate are arranged radially
about a central cell. There are about ten of these vessels in L. Douglasii.

In a longitudinal section of the root tip region in Floerkea proser-
pinacoides the plerome cylinder is about 4 to 5 ceUs in width. Overhdng
the plerome is a regular periblem layer, which by its division gives rise
to the plerome and periblem. Over the periblem of the apex region of the
root is a crescentic group of cells which gives rise to both dermatogen and
root cap, and can therefore be called a cal}^trodermatogen layer. The
root cap arising from the division of this layer is conical, and about
five layers of cells thick in its deepest portion.

The longitudinal section of the root tip in L. Douglasii shows the same
structure, differing only in having a wider plerome cyHnder than Floerkea.

Stem

The stem of Floerkea is weak and flaccid. It remains green, and
rarely branches in Floerkea. In Limnanthes Douglasii the stem is stouter
and is much branched, especially at its base.

Sections of the stems of Floerkea proserpinacoides imbedded in paraffin
were cut, and stained in safranin and Delafield's haematoxyhn. In
such transverse sections, there is an external epidermal layer of small
rectangular cells, shghtly cuticularized (Fig. 5). Beneath the epi-
dermis is a layer of larger oblong cells, regular in shape, not thickened.
Within this is the cortical region, 6 to 7 cells in depth. The cortical cells
are large, round cells with thin walls. Throughout the cortical region are
definite circular inter-cellular spaces surrounded by a sheath of cortical
cells. The innermost cortical layer forms an indistinct endodermis.
Within the endodermis is the vascular strand. This consists of a num-
ber of collateral bundles, separated from each other by pith or paren-
chymatous tissue. There is no ring of wood formed. The bundles consist
of two or three spiral tracheae, and a varying number of spiral, or annu-
lar tracheids. The phloem is poorly developed, consisting of a few sieve
tubes and fewer phloem cells, forming a slightly darker stained patch
external to the xylem in the bundles. The two outermost layers of

FLOERKEA PROSERPINACOIDES AND ALLIES 409

phloem cells form a loose, open-celled outer bast fiber region. L.
Douglasii shows this more strongly developed. Between the phloem
and the xylem, there is an easily recognizable large celled intrafascicular
cambium. There is no interfascicular cambium shown however. The
intrafascicular cambium shows division to form secondary wood. In
most of the stems sectioned, 8 to 10 bundles were distinguishable. Sev-
eral join, usually two to four at the ends of the eUipse of the bundle
strand The bundles thus placed are those which pass off at the node
into the petioles of the leaves. The other bundles remain distinct.

In the transverse sections of L. Douglasii stems which were cut free
hand from herbarium material, and stained in safranin and methyl
green, the structure was found to resemble F. proserpinacoides closely.
The epidermis is as in F. proserpitiMoides, excepting for the development
of a thick cuticle. The same regular row of cells is present immediately
beneath the epidermis. The cortex is of from 7 to 8 layers of large cells;
intercellular spaces are rare, no air lacunae are present. The endodermis
is indistinct. The bundles are more strongly developed, that is, the
xylem elements are more numerous, and the phloem patches are larger.
A complete ring of bundles is not formed, and little secondary wood is
laid down. The two outer layers of the phloem form a poorly developed
bast fiber region. The cambium is pecuUar, its large thin-walled cells
resembling markedly the cambium in F. proserpinacoides. The young
stem of L. Douglasii, above the cotyledons, shows a number of
air lacunae in the cortex corresponding to those in F. proserpinacoides.
These do not appear in the older stems of L. Douglasii examined.

The bundles vary in number, in the stem of F. proserpinacoides,
according to the level at which the sections were made. In the younger
parts of the stem, the bundles may be four with four wide medullary
rays between the bundles. This is also true in L. Douglasii. These
bimdles separate in the older stem into a number of bundles. At the nodes
four bundles give off traces into the petiole.

Leaf

The leaves in F. proserpinacoides are alternate, exstipulate, and
pinnately compound, with 3 to 5 leaflets. In Limnanthes, the leaves
may be pinnate wdth 7 leaflets, to bipirmate as in L. Douglasii.

The first foHage leaf is trifoliate. The second leaf may be perfectly
5 foliate, but is more usually imperfectly divided, showing one terminal
entire leaflet, and two lateral half-divided leaflets, or four entire leaflets.
The third leaf is usually perfectly 5 -foliate in Floerkea.

410 RUSSELL— A COMPARATIVE STUDY OF

Leaflets were imbedded in paraffin and sectioned. These showed a
characteristic umbrophytic structure. In transverse section there
appeared: — An upper epidermis of rounded cells, not thickened (Fig. 6),
containing stomata. Below the epidermis is a layer of loose palisade
tissue, rich in chlorophyll, with many open intercellular spaces lying
below the stomata. Below the palisade tissue is a narrow zone of meso-
phyll tissue containing the bundles. There are, in transverse section of a
leaflet, a stronger midrib, two lateral strong veins, and two intermediate
smaller veins. The vascular bundles are poorly developed, showing
merely a few xylem and phloem cells. The lower epidermis is identical
with the upper, excepting for the presence of more stomata. The veining
in a leaflet consists of a midrib, branching at the base into two lateral
branches running along the margin of the leaflet, and uniting with the
midrib at the tip. From the margin lesser veins rim toward the middle
of the leaflet. A longitudinal section (Fig. 7) through the median part
of a leaflet shows at the tip a single hydathode. From the end of the
united bundles at the tip a mass of parenchymatous tissue spreads out
in a fan-shaped mass. These cells are filled with a denser protoplasm
than the surrounding cells. This is the epithem tissue of the hydathode.
It lies immediately below a circular area of epidermal cells at the tip of
the leaflet. The cells of this area differ from the ordinary epidermal
cells in that they are oblong in shape and have smooth, straight walls.
Stomata are more numerous than in the epidermis outside this region
and are of smaller size (Fig. 8 and 9). The guard cells of these stomata
are always open. The stomata themselves are more nearly spherical
in outline than the normal stomata, which are elliptical in shape, and
usually have their guard cells closed.

The water stomatic area at the tip in the older leaflets undergoes a
breaking down in the epidermal cells, forming in this way a water pore
(Fig. 7).

In both Floerkea and Limnanthes the leaves are glabrous. In Limnan-
thes, however, there are present tannin containing cells on the under
surface of the leaf. They are long slender cells, running parallel to the
axis of the leaflet. Sometimes they are present on the outer face of the
sepals (Fig. 13). These are mentioned by Solereder (25). They give a
positive reaction for tannin upon the application of ferric chloride.
There are no tannin cells present in Floerkea proserpinacoides.

Flower

The flowers of F. proserpinacoides are solitary axillary. All the leaves,
except the first foliage leaf, bear a flower in the axil.

FLOERKEA PROSERPINACOIDES AND ALLIES 411

In F. proserpinacoides the flowers are small and inconspicuous. They
are trimerous in symmetry. There are three ovate acuminate green
persistent sepals, united at their bases. Stomata are present in great
numbers on both the outer and inner faces of the sepals. Water stomata
are present upon the tip.

The petals are oblong and white, three in number, and only one-third
as long as the sepals. Upon the base of the petals are a few blunt short
unicellular hairs (Fig. 10).

The stamens are six, in two rows. The outermost row, opposite the
sepals, have upon their outer side a gland at the base of the filament.
The inner row, opposite the petals, are without glands. The stamens
equal each other in height, and are shorter than the petals. The anthers
are globose, and introrse, versatile, dehiscing longitudinally. The pollen
is smooth and oval. The stamens with the petals are set on the inner
edge of a flat disk of cells upon the receptacle.

The pistil consists of three separate carpels with a g>Tiobasic style
set between the carpels. There are usually only two carpels well de-
veloped, the third being a mere rudiment which is early crowded out by
the two functioning carpels. The style rarely divides into three, mostly
into two short style arms, each bearing a capitate stigmatic surface.

Sections of each young carpel show one anatropous ovule. The
funiculus arises from the base of the receptacular tissue, runs up freely
within the carpel and bears the inverted ovule (Fig. 11). When a third
carpel is present, the ovule present is rudimentary, and imperfect. One
carpel alone usually matures; maturation of two carpels is, however, not
uncommon.

In sectioning the flowers at different stages, there are shown in the
carpels and fruit certain pecuHarities of structure. Over the carpellary
wall there are papillae \-isible even to the naked eye. In transverse
sections prepared from paraffin material these papillae are seen to be
formed by pushing out of the large celled epidermis of the carpellary wall.
These in the fruit cause the appearance described as "papillose," "rug-
ose." These epidermal cells are peculiarly thickened on their outer
walls by wart-Uke protuberances. Upon application of chlorophyll
extract, these color green, but give no color reaction when phloroglucin
is applied, showing that they are thickenings of cutin, not Hgnin. Within
this epidermal layer in the carpel are four layers of large cells with thin
walls, which form the remainder of the carpellary wall. The innermost
layer of the carpellary wall is flattened, forming a distinct Une of demarca-

412 RUSSELL— A COMPARATIVE STUDY OF

tion between the carpellary walls and a broad zone of large thin-walled
cells. This is shown, by its reaction to iodine, to be a starch storing
layer. In its outer part, lying nearer to the carpellary wall, are numerous
small vascular bundles, cut transversely. This zone is, from all evidence
at hand, a wide development of the tissue of the integuments surrounding
the single ovule. The wide development and apparent fusion of the in-
teguments, and the apparent single layer of the nucellar tissue, breaking
down early in the development of the ovule, are characteristics resembling
the development of the ovules of Impatiens fulva, as described by Mr.
Franklin Carroll in the present publication ("The Development of the
Chasmogamous and the Cleistogamous Flowers of Impatiens fulva").
Floerkea represents perhaps a more reduced state than that found in
Impatiens fulva. Further careful study of the development of the ovule
of Floerkea would be necessary, however, before a definite opinion on the
matter could be expressed.

Within the fused integuments in the more nearly matured seed is a
broken down granular substance, immediately surrounding the remnants
of the embryo sac membrane. This represents the broken down nuceUar
tissue. Within the embryo sac is the embryo showmg the cotyledons
cut transversely, and the radicle between. In the spaces about the
maturing embryo are shreds of the broken down endosperm tissue almost
entirely consumed by the embryo at this stage of growth.

In L. Douglasii the flowers are solitary axillary. The flowers are
5-merous in symmetry, having 5 sepals, valvate in aestivation, green and
persistent, united at the base as in the F. proserpinacoides. The petals
are spatulate and about twice the length of the sepals. They are whitish
or yellowish in color. On the base of the petal are unicellular hairs of
the same type as in Floerkea, but longer. Fig. 10 gives the relative size
of the hairs in Floerkea and Limnanthes. On the middle regions of
the petals in Limnanthes, however, are hairs not present on the
petals of F. proserpinacoides. These are longer than the petals of
F. proserpinacoides and are shown in relative size in Fig. 10. These
hairs are peculiar in their shape and walls. They are unicellular, and
tapering. Their walls are covered with wart-hke thickenings resembhng
the thickenings of the walls of the epidermis in the carpels of Floerkea
and Limnanthes.

The stamens are 10, in two rows. The outer row, opposite the sepals,
and not the row opposite to the petals, as is stated in some descriptions,
are the ones which bear the gland, as in Floerkea. The stamens are

FLOERKEA PROSERPINACOIDES AND .\LLIES 413

shorter than the petals or sepals, and are all of the same length as in
Floerkea.

The carpels are five in number, all well developed, and having between
them the gynobasic style bearing five short style arms with capitate
stigmas as in Floerkea.

The carpellary wall in Limnanthes has papillae as in Floerkea. Upon
examination these show upon their outer epidermal cells minute thicken-
ings of cutin as in Floerkea (Fig. 11). The much developed integument is
present in Limnanthes also as in Floerkea, storing starch until late in the
growth of the seed.

In going over a great number of plants of Floerkea proserpinacoides
many were found to have on them flowers departing from trimerous
symmetry. Most of these had four sepals, three petals, six stamens,
and two carpels. The sepals upon examination showed by their venation
that the four were derived by a splitting of one of the three usually pre-
sent. Several were found with four sepals, four petals, six stamens, two
carpels. One only was found with five sepals. It, however, had only
three petals.

Conclusions

In the separation of the two genera forming this family the dis-
tinguishing features have been: —

Symmetry of flowers
Length of petals
Aestivation of the petals

In flower symmetry, however, there is now a gradation series. This
begins with those having pentamerous symmetry, such as L. Douglasii
and L. alba. The tetramerous flower of L. Macounii forms the transition
type to the trimerous form described as Floerkea. In Floerkea constant
variations from the tetramerous to the trimerous type have been found.

These range from imperfect tetramery in the sepals, to forms with
the four sepals and four petals. One flower found approached pentamer-
ous symmetry in possession of five sepals. The variations in these
cases did not extend to the stamens and carpels. The stamens remained
six in number, the carpels, as usual two to three.

In the length and aestivation of the petals there is also an intergrading
series represented in the two genera: —

Limnanthes rosea seems to be the cHmax type, possessing as it does
rose colored petals, much exceeding the sepals in length (sepals 7-8 mm.
long; petals 12-18 mm. long). Petals convolute in aestivation.

414 RUSSELL— A COMPARATIVE STUDY OE

Limnanthes Donglasii: — the petals are a more primitive color, white
or whitish yellow, still much exceeding the sepals, yet not so large as in
Limnanthes rosea (sepals 5-6 mm., petals 10-15 mm.). Limnanthes
versicolor is about as L. Douglasil and is probably a hairy form of L.
Douglasii. L. alba also has petals exceeding the sepals (sepals 6-8 mm.,
petals 10-15 mm.).

Limnanthes pumila forms a transition type from those whose petals
exceed the sepals, to those with sepals exceeding the petals. It is given
as having petals "scarcely exceeding the sepals. " The petals too in this
more nearly approach the oblong type found in forms with smaller petals.

Limnanthes floccosa comes next with "oblong cuneate petals, not
exceeding the sepals."

In the tetramerous Limnanthes Macounii the petals equal the sepals.
This forms a valuable transition species, being reduced both in the num-
ber and size of the parts (petals 3-4 mm., sepals 3-4 mm.).

In Floerkea proserpinacoides the petals are oblong and about the
length of the sepals (sepals 3 mm.; petals 1-5 mm.), flowers trimerous,
valvate in aestivation.

In Floerkea occidentaUs the petals and sepals are reduced in size
(petals 1 mm., sepals 2-3 mm. long). The petals are valvate in aestiva-
tion. This distinction is not important since it would be impossible from
the size of the petals in Floerkea to have a convolute aestivation. The
petals are so small and reduced that they do not overlap.

The two species of Floerkea are distinguished by Rydberg (1) on
account of the different lengths of peduncles, thus: —

Peduncles longer than the petioles — F. occidentaUs.

Peduncles rarely equalling the petioles — F. proserpinacoides.

This is not a difference to be depended on, for F. proserpinacoides
does show the peduncles longer than the petioles. The plants exhibiting
this were collected in different localities around Philadelphia. Rydberg
and others who accept F. occidentaUs do not report it from the region
about Philadelphia.

From the anatomy of these two forms just described, F. proserpina-
coides and L. Douglasii, it will be seen that they show a remarkable
similarity in structure. Their correspondence in such minute details
as the thickenings upon the epidermal cells of the carpellary walls, in
the possession by both of the fused integuments about the ovule, which
store food in the form of starch, of the subsequent absorption of the
starch in the growth of the embryo, and the flattening of the empty cells
by the enlarged embryo in the mature seed; by the presence upon the

FLOERKEA PROSERPIXACOIDES AND ALLIES 415

tip of the cotyledon in both of a water stomatic area, is striking. Simi-
larity of structure in all of these parts indicates surely an extremely close
relation between these two plants placed in separate genera. In com-
paring the anatomical features, it seems reasonable to look upon Floerkea
as a type of plant evolved from Limnanthes by reduction. Thus: the
stem shows a similar but reduced structure to that of Limnanthes; the
glands upon the stamens in Floerkea are not so large as are those of
Limnanthes, nor are the cells so rich in protoplasm. The hairs upon the
base of the petals in Floerkea are shorter than those in Limnanthes, where
they probably act as protecting hairs about the nectaries or glands at
the base of the stamens between the petals. In Floerkea these probably
do not function as in Limnanthes, since the petals are much reduced in
size, and the hairs are too short to reach out and protect the gland at the
base of the stamens. In the number and size of flower parts Floerkea
shows reduction from Limnanthes as has been already traced. The
geographical distribution already noted also suggests a continuous rela-
tion and close affinity between the pentamerous, tetramerous and tri-
merous forms in the two genera.

Until a complete series is made out, however, between the pentamer-
ous forms with epigeal cotyledons and the trimerous forms with h>^ogeal
cotyledons the two genera wdll have to remain distinct as Robert Brown
\dewed them.

The writer takes this opportunity to acknowledge her indebtedness
to Dr. John M. Macfarlane for his advice and assistance in preparing
this paper.

Summary

Two plants were studied microscopically and macroscopically as
representing each one of the two genera of the family Limnanthaceae —
Floerkea proserpinacoides and Limnanthes Doiiglasii.

In the anatomical features of the root, stem, leaf, flower and fruit,
these two forms show a striking similarity of structure, Floerkea proser-
pinacoides indicating by its reduced members that it is a form derived
from Limnanthes by reduction. In its present distribution, the family
Limnanthes is represented in the west by species of Limnanthes, and in
the east and west by Floerkea. Transition types are represented by
various species of Limnanthes ranging from large attractive pentamerous
types through a lesser tetramerous species to the trimerous flower types
characteristic of the genus Floerkea. Since the genus Floerkea overlaps
the distribution areas of the pentamerous and tetramerous forms it is
likely to be a t>^e evolved from them which worked its way through

416

RUSSELL— A COMPARATIVE STUDY OF

southern Canada to northern New England and the Middle Atlantic
States. Although the two genera are similar enough to form a single
genus, until forms showing transition from the pentamerous type of
Limnanthes with epigeal cotyledons, to those of the genus Floerkea with
trimerous symmetry and hypogeal cotyledons, the two genera should
remain perhaps as in Robert Brown's Classification.

1.

Rydberg, Axel

2.

Howell, Thos.

3.

Greene, Nath.

4.

Macoun, J.

5.

Gray, Asa

6.

Pursh, F. T.

7.

Nuttall, Thos.

8.

Lindley, J.

9.

J. Torrey & A. Gray

10.

Torrey, J.

11.

Darlington, Wm.

12.

Piper, C. V.

13.

Britton, N. L.

14.

Parsons, M. E. &

Buck, M. W.

15.

Clements, F. E. &

Clements, E. S.

16.

Rydberg, Axel

17.

Stone, W.

18.

Watson, S.

19.

Jepson, W. L.

20.

Baillon, H. E.

21.

Engler, A. &

Prantl, K.

22.

Engler, A. &

Prantl, K.

23.

Small, J. K.

24.

Britton, N. L. &

Brown, A.

25.

Solereder, H.

26.

De Bary, A.

27.

Goebel, K. E.

BIBLIOGRAPHY

North American Flora, vol. 25, part 2.
Flora of N. W. America 1 :108. (1897)
Erythrea3:62. (1895)
Cat. Canadian Plants 3:502. (1886)
Synoptical Flora of North America 1:363. (1897)
Flora of N. America. (1814)
Genera Plantarum 1 :229. (1818)
Bot. Register, vol. 20, 9. pi. 1673. (1835)
Flora 1:210. (1838)
Flora of New York 1 :126-7. (1843)
Flora Cestrica p. 213. (1853)

Flora of the State of Washington, (Contrib. Nat. Herb.)
vol. XI:383. (1906)

Manual of Flora of Northern States & Canada, p. 599.
(1901)

Wild Flowers of Cahfornia, p. 126. (1897)

Rocky Mountain Flowers, p. 37. (1914)
Flora of Montana, p. 269. (1900)
Report of N. J. State Museum. (1910)
Geological Survey of California, p. 95. (1876)
Flora of Western Middle California, p. 248. (1901)
Natural History of Plants, vol. 3, p. 20-1.

Die Natiirhchen Pflanzenfamilien, 3.5a, p. 136. (1896)

Die Naturlichen PflanzenfamiUen, 3.4. (1896)
Flora of S. E. United States, p. 725. (1913)

Illustrated Flora of the United States & Canada, vol.

II. (1897)

Systematic Anatomy of the Dicotyledons, vol. I, p. 169

(1908)

Comp. Anatomy of the Phanerogams & Ferns. (Eng.

Ed. 1884)

OutUnes of Classification & Special Morphology. (1887)

FLOERKEA PROSERPINACOIDES AND ALLIES 417

28. Coulter, J. ^L &

Chamberlain, C. Morphology of the Angiosperms. (1910)

29. Willdenow, K. L. Neue Schriften der Gesellschaft der naturforschenden

Freunde. 3:148. (1801)
30 Barton, W. P. C. Compendium Florae Philadelphicae, p. 171. (1818)
31. Brown, Robert London, Edin., Dublin Philos. Mag. of General Science

vol. 2:70. (1833)

DESCRIPTION OF PLATES
Plate XCI

Fig. 1. Longitudinal section tip of cotyledon Limnanthes Douglasii (X220)
H = Epithem tissue below water stomatic area.
T = Bundle suppljing epithem tissue.
S = Storage tissue of cotyledon.

Fig. 2. Longitudinal section tip of cotyledon of Floerkea proserpinacoidcs (X150).

H = Epithem tissue.

T= Vascular strand.

S = Storage tissue of cotyledon.

K = Remnants of seed coats.
Fig. 3. Transverse section root of Floerkea proserpinacoides (X220).

E = Epidermis. X=Xylem.

C = Cortex. P = Phloem.

D = Endodermis.

Fig. 4. Transverse section root of Limnanthes Douglasii (X220) labeHng as above.

Fig. 5. Transverse section stem of Floerkea proserpinacoides (X80).
E = Epidermis. Pi = Pith.

C = Corte.x. X=Xylem.

L = Lacunae. P= Phloem.

D = Endodermis.

Plate XCII

Fig. 6. Transverse section leaflet of F/oeryfeea /To^er/iiwacojJe^ (X153).
T = Palisade tissue.
I = Intercellular space.
M = Midrib.

Fig. 7. Longitudinal section leaflet of Floerkea proserpinacoides (X153).
T = Palisade tissue.
H = Epithem tissue.
P = Mesophyll.

Fig. 8. Stomata from water stomatic area Floerkea proserpinacoides (X 200).

Fig. 9. StoTOSita liovalovftT t'^idtirms Floerkea proserpinacoides (X 200).

418 RUSSELL— A COMPARATIVE STUDY OF

Fig. 10. A. Petal of flower of Floerkea proserpinacoides, showing hairs at its base
(X 15).
B. Hairs from base of petal of Limiianthes Douglasii (X 15).

Fig. 11. Longitudinal section thru ovule of Floerkea proserpinacoides (X 150).

C = Carpellary wall.

N = Nucellus.

S = Membrane of embryo sac.

E = Egg.

F = Fused indusia.

Fig. 12. Carpels and gynobasic style of Floerkea proserpinacoides (X 25).

Fig. 13. Sepal of flower of Limnanthes Douglasii, showing cells with tannin contents
(X 25).

Bot. Conlrib. Univ. Perm.

Vol. IV, Plate XCI

Fig. 1

Fig. 2

Fig. 3

Fig. 4

Fig. 5
RUSSELL ON FLOERKEA

Bot. Coutrib. Univ. Finn.

Vol. IV, Plate XC 1 1

Fig. 6

Fig. 8

Fig. 12

Fig. 1.3

Fig. 11 Fig. 10

RUSSELL OX FLOERKEA

BIOCHEMICAL STUDIES OF INSECTIVOROQS PLANTS'

CONTENTS

Page
I. The Work of Pre\aous Investigators on Nepenthes, by Joseph Samuel Hep-
burn, A.M., M.S., Ph.D 419

II. A Study of the Protease of the Pitcher Liquor of Nepenthes, by Joseph

Samuel Hepburn, A.M., M.S., Ph.D 442

III. A Bacteriological Study of the Pitcher Liquor of Nepenthes, by Joseph S.

Hepburn, Ph.D., and E. Quintard St. John, M.D 451

W. Occurrence of Antiproteases in the Larvae of the Sarcophaga Associates of

Sarracenia Jlava, by Joseph Samuel Hepburn and Frank Morton Jones 460

I. The Work of Previous Investigators on Nepenthes

BY

Joseph Samuel Hepburn, A.M., M.S., Ph.D.

Voelcker (1) studied the composition of the pitcher Hquor of Nepen-
thes. He describes the Hquor from unopened pitchers as a clear, colorless
liquid with a refreshing taste, an agreeable but not very pronounced
odor, and an acid reaction to litnfus. The acid was not volatiUzed by
evaporation of the liquor to dryness.

The total soUds of the liquor were determined by drying at 212° F.
(100° C), the ash by ignition of the soUds at a red heat. Total solids
are expressed as percent of the pitcher liquor, ash and volatile matter
as percent of the total soUds. The solids were yellow or cream colored,
very hygroscopic, and readily soluble in water. Samples of Uquor from
different unopened pitchers gave the following values: —

Total Volatile matter Ash

solids (loss on ignition)

1. 0.92

2. 0.91 25.86 74.14

3. 0.62 36.06 63.94

4. 0.27 32.92 67.08

* Through the kindness of Dr. John M. Macfarlane, these studies were conducted
in the Botanical Laboratory and Nepenthes House of the University of Pennsylvania
during 1914-1916. Papers that have appeared since that time have also been cited
in the bibliographies.

420 HEPBURN— BIOCHEMICAL STUDIES

The liquor from unopened pitchers contained potassium, sodium,
magnesium, calcium, chlorine (hydrochloric acid), and organic acids;
other bases, and sulphuric, phosphoric, oxahc, tartaric, and racemic
acids were shown to be absent.

The Hquor from opened pitchers was yellowish and not always clear;
it was acid to litmus, and contained the same acids and bases as were
present in the liquor of unopened pitchers. A determination of total
solids gave 0.87 percent; the solids were yellow and readily soluble in
water. Free volatile acids (including acetic and formic) were absent,
for distillation of one-half ounce of the liquor to dryness yielded only
distilled water as a distillate.

A further study of the organic acids and the ash was made on the
residue obtained by evaporation to dryness of the mixed liquor collected
from both unopened and opened pitchers. The organic acids were found
to consist chiefly of mahc acid, plus a httle citric acid.

The ash had the following composition: —

Potassium chloride

76.31%

Sodiimi carbonate

16.44%

Calcium oxide

3.94%

Magnesium oxide

3.94%

Total 100.63%

Using the average value for the volatile matter of the total solids of
the liquor of unopened pitchers (see above), and allowing for the carbonic
acid content of the ash, the composition of the total solids was found
to be: —

Organic matter (chiefly malic acid and a little citric acid) 38.61%

Potassium chloride 50.42%

Sodium oxide 6.36%

Calcium oxide 2.59%

Magnesium oxide 2.59%

Total 100.57%

The yellow color (e.g., of the total solids) is ascribed to a small quan-
tity of "another organic matter."

OF INSECTIVOROUS PLANTS 421

Hooker (2) found that the pitcher Hquor was ahvays acid. The
secretion of the hquor was not increased b}' the introduction of inorganic
substances, but was increased by the introduction of animal matter into
the pitcher. As substrates, Hooker used fibrin, raw meat, cartilage and
cubes of egg white. He reported: —

"After twenty-four hours' immersion, the edges of the cubes of white
of egg are eaten away, and the surfaces gelatinised. Fragments of meat
are rapidly reduced; and pieces of fibrin weighing several grains dissolve
and totally disappear in two or three days. With cartilage the action
is most remarkable of all; lumps of this weighing 8 or 10 grains are half
gelatinised in twenty-four hours, and in three days the whole mass is
greatly diminished, and reduced to a clear transparent jelly. After
drying some cartilage in the open air for a week, and placing it in an
unopened but fully formed pitcher of A\ Rafflesiana, it was acted upon
similarly and very little slower. "

All of the above experiments save the last were apparently conducted
in opened pitchers. Experiments, made in vitro with pitcher liquor and
fibrin, meat, and cartilage, gave a digestion of the substrates "wholly
different" from that occurring in the pitchers. Hence the digestive
action in the pitchers was not produced by the fluid first secreted by the
pitchers.

From these and similar experiments Hooker concluded: — "It would
appear probable that a substance acting as pepsine is given off from the
inner wall of the pitcher, but chiefly after placing animal matter in the
acid fluid; but whether this active agent flows from the glands or from
the cellular tissue in which they are imbedded, I have no evidence to
show. " However, he attributed all three phenomena — secretion of the
acid liquor, digestion (secretion of the digestive enzyme), and assimila-
tion— to the glands, and decided that Nepenthes possess a true digestive
process.

From the following experiments, it seemed probable that the tempera-
ture influenced the digestive action. When cartilage or fibrin was kept
for six days in pitchers of N. ampullaria in a cold room, it was not di-
gested; the substrate, however, was immediately acted upon when trans-
ferred to pitchers of N. Rafflesiana in a stove house.

If large amounts of substrate were introduced into a pitcher, digestion
and absorption did not attain completion, and putrefaction was noted
after the lapse of many days.

Tait (3) isolated an enzyme from the liquor of unopened Nepenthes
pitchers, and also from the liquor of opened pitchers. His procedure

422 HEPBURN— BIOCHEMICAL STUDIES

was as follows. The liquor was acidulated with dilute phosphoric acid,
and a thin suspension of calcium carbonate in water was added until
effervescence ceased; the precipitate of calcium phosphate, which ad-
sorbed the enzyme, was permitted to stand for 24 hours, then was sepa-
rated from the supernatent liquid by decantation, and was dissolved in
very dilute hydrochloric acid. A saturated solution of cholesterol in a
mixture of absolute alcohol and absolute ether was added to this solution.
The precipitated cholesterol, which adsorbed the enzyme, was dissolved
in absolute ether. The aqueous layer now contained the enzyme as a
gray, amorphous, fiocculent, suspended soUd, which was partly soluble
in distilled water, msoluble in boiling water, very soluble in glycerol, and
produced a characteristic viscid change in a small quantity of fresh milk.
Unfortunately the power of the enzyme to digest proteins was not tested;
and its action on milk recalls that of rennin, rather than that of a true

protease. -

Tait permitted the Hquor from four virgin pitchers of N. phyllamphora
to act for 28 hours on cubes of albumen (volume of each cube 1 cubic
millemeter). The substrate remained unchanged. But one sample of
the liquor was acid; the other three samples were absolutely neutral; the
enzyme described above was found in-but one sample.

After the pitchers had opened, the liquor became changed in its
reaction, and m its proteolytic power. "Fluid taken from pitchers into
which flies have previously found their way is always very acid, has a
large quantity of the ferment, and acts in a few hours on cubes of albu-
men, making them first yellow, then transparent, and finally completely
dissolvmg them. " The liquor in a pitcher, in which insects were under-
going digestion, was "very viscid and very acid. "

Tait sums up: — "In the unopened pitcher the secretion is only faintly
acid and not at all viscid. The secretion is increased therefore, . . .
, in quahty after food has been taken in."

The work of von Gorup and Will (4) was largely carried out on the
secretion of N. phyllamphora, Willd. and N. gracilis, Korth. Separate
studies were made, in vitro, on (a) liquor from pitchers which were free
from insects, and (b) liquor from pitchers which had been entered by
insects and contained their remains. These may be spoken of as non-
stimulated and stimulated pitchers respectively.

The liquor was almost colorless, famtly opalescent or entirely clear,
odorless, without any distinct taste, and of varying consistency— some
samples being thick, others being thm like water. The liquor from non-
stimulated pitchers was neutral or, at the most, very famtly acid; that

OF INSECTIVOROUS PLANTS 423

from stimulated pitchers imparted a decidedly red color to litmus paper,
and the color did not completely vanish on exposure to the air.

Liquor from Stimulated Pitchers

The insect residues were removed by filtration, and the action of the
filtrate on certain substrates was studied. Ox-blood fibrin was swollen
to a gelatinous mass in 0.2 percent hydrochloric acid, then was freed as
completely as possible from the adherent acid by means of pressure. A
flock of this fibrin was almost completely dissolved by the pitcher liquor
in ^ to 1 hour at a temperature of incubation of 40° C, and in 2 hours at
a temperature of incubation of 20° C. The resulting solution was
faintly opalescent. The time required for solution of the fibrin was
reduced to 14: hour by addition of several drops of 0.2 percent hydro-
chloric acid to the liquor. After digestion for two hours, the filtered solu-
tion remained clear on boiling, and was not precipitated by mineral acids
or by acetic acid plus potassium ferrocyanide, but was precipitated by
mercuric chloride, by tannin, and by phosphotungstic acid, and gave an
excellent rose-red biuret reaction.

Little slices of coagulated egg white were digested with liquor to which
1 or 2 drops of 0.2 percent hydrochloric acid had previously been added.
After 24 hours at 20° C, the sUces were attacked at the edges, and were
transparent; the filtrate from them gave a distinct rose-red biuret reaction.

On digestion at 20° C. with pitcher liquor and 1 or 2 drops of 0.2 per-
cent hydrochloric acid, raw meat soon became transparent at the edges
and somewhat swollen, and was partly dissolved without putrefaction.
A further change was not noted after 48 hours. The filtered solution
was not precipitated by boiling, nor by the addition of acetic acid and
potassium ferrocyanide, was precipitated by mercuric chloride and by
tannin, yielded a cloudy precipitate with phosphotungstic acid soluble in
excess of the reagent, and gave a positive biuret reaction.

Legumin became transparent on the edges and somewhat swollen
after digestion for 24 hovu-s at 20° C. with pitcher liquor plus 1 or 2 drops
of 0.2 percent hydrochloric acid; the filtered solution gave a very decided
biuret reaction.

Gelatin, upon which pitcher liquor plus several drops of 0.2 percent
hydrochloric acid had been poured, dissolved almost completely after
24 hours at ordinary temperature. The solution was filtered and con-
centrated to a small volume; it did not gelatinize, but remained a thick
syrup; digestion had occurred with the production of gelatin-peptone,
and the power to gelatinize had been destroyed.

424 HEPBURX— BIOCHEAIICAL STUDIES

The pitcher liquor did not contain a diastase. The Hquor was mixed
with a thin starch paste, and then incubated for 24 hours at 20° to 30° C.
The starch was not hydrolyzed, the filtered solution was optically inac-
tive, and did not reduce FehUng solution, hence did not contain reducing
sugar.

Sufficient liquor from stimulated pitchers was not available to deter-
mine the nature of the free acid present in it. It is stated that hydro-
chloric acid may be excluded.

Liquor from N on- stimulated Pitchers

Gelatinous, swollen fibrin was washed until the acid reaction (due to
hydrochloric acid) had almost completely vanished. Flocks of the
fibrin, which were placed in the liquor, suffered no noticeable change with-
in several hours at 20° to 30° C, and had not dissolved to the shghtest
extent at the end of 24 hours; the fibrin had contracted somewhat, and
the filtrate from it gave a scarcely noticeable tinge of rose-red in the
biuret test. However, the fibrin was almost completely dissolved in
13/^ hours by pitcher liquor to which 2 or 3 drops of 0.2 percent hydro-
chloric acid had previously been added; the resulting solution behaved
in every way as did the solution obtained by the action of the acid liquor
from stimulated pitchers.

Swollen fibrin, carefully freed from adherent hydrochloric acid, was
dissolved almost instantaneously at the ordinary temperature by liquor
to which 3 or 4 drops of dilute formic acid had previously been added.
The residue, which was scarcely noticeable, was removed by filtration,
and the filtrate was carefully neutralized. The very shght precipitate
which formed was collected on a filter; the filtrate gave none of the react-
ions of the true proteins, but did give an intense biuret reaction. When
the formic acid was replaced by acetic acid, or by propionic acid, the
fibrin was digested less rapidly, the rate of digestion now being about the
same as in the liquor of the stimulated pitchers. At a temperature of 20°
to 30° C, the fibrin was completely dissolved in 2 to 3 hours; the resulting
solution contained mainly metaproteins and gave a very faint biuret
reaction.

When the pitcher liquor was previously acidified with malic acid, the
fibrin was almost completely dissolved in 10 minutes at ordinary tempera-
ture; the resulting solution gave a faint but distinct biuret reaction.
After the digestion had been in progress for 2 hours, the precipitate on
neutralization was very slight, and the biuret reaction decidedly more
marked. When citric acid was substituted for malic acid, the fibrin was

OF INSECTIVOROUS PLANTS 425

dissolved in a considerably shorter time; after digestion for 2 hours, the
resulting solution gave a very sUght precipitate on neutralization, and
gave a biuret reaction as intense as that obtained in the case of
formic acid.

The solutions, obtained by digestion, contained much more protein
(metaprotein) immediately after the fibrin had dissolved, than in later
stages of the digestion. This protein gradually vanished, and was dis-
placed by peptone as the period of digestion lengthened. From these
phenomena, which were repeatedly obser\^ed, it followed that, apparent-
ly, peptone represented the second and not the first stage of the action
of the enzyme.

Von Gorup and Will, who made the necessary control experiments
on their reagents, do not hesitate distinctly to designate the acid liquor
of the stimulated Nepenthes pitchers as a solution of plant pepsin. The
neutral secretion of the non-stimulated pitcher, like pepsin in the absence
of free acid, exerts no digestive action.

Vines (5) dehydrated pitchers of Nepenthes species (iV. hybrida and
A'', gracilis) by means of absolute alcohol, converted their tissue into a
pulp, and extracted this with glycerol in order to obtain a solution of the
protease. As a substrate, he used fibrin which had been soaked in 0.2
percent hydrochloric acid until gelatinous. When the fibrin was in-
cubated at 40° C. with the enz>Tne solution to which a few drops of 0.2
percent hydrochloric acid had previously been added, at the end of 8
hours the substrate showed signs of digestion and, after filtration, the
solution gave a distinct peptone (biuret) reaction. Control experiments
on fibrin plus enz\TTie solution, and on fibrin plus 0.2 percent hydrochloric
acid, did not show digestion and did not give a biuret reaction. Hence
the glands of the pitchers contained a proteolytic enzyme, which was
soluble in glycerol and was active only in the presence of acid.

Pitchers were gathered from the same plants (of the species mentioned
above) at the same time. Some pitchers were immediately treated as
outlined above for the preparation of a glycerol extract. Other pitchers
were first treated with dilute (1 percent) acetic acid for 24 hours, then
the glycerol extract was prepared, using the procedure already described.
In every set of experiments, the extract from the pitchers, which received
the preliminary treatment with acetic acid, was higher in proteolytic
power than the extract from pitchers not so treated. Thus the same
volume of each extract was permitted to act on a pellet of swollen fibrin
in the presence of 2 cc. of 0.2 percent hydrochloric acid at a temperature
of 40° C. The fibrin pellets were similar. At the end of 6 hours, the

426 HEPBURN— BIOCHEMIC.\L STUDIES

extract from pitchers, which had received the acid treatment, had com-
pletely dissolved the fibrin, while the extract from pitchers not so treated
had but slightly attacked the fibrin. The solutions, obtained by filtra-
tion of the contents of each tube, always gave a positive response to the
biuret test.

Vines states: — "These experiments seem to indicate that in the gland-
cells of the Nepenthes pitchers, . . . ., the digestive ferment exists at first
in combination with some other body, as zymogen — and that in plants,
as in animals, this zymogen can be split up by the action of dilute acid,
the free ferment making its appearance as a result of this decomposition. "
In another report on this research, Vines (25) also described tests
for the presence of diastase in the pitchers. The glycerol extracts of the
pitchers were without action on starch, therefore, did not contain dias-
tase. He also noted that the glycerol extract did not contain sugar.

During a further study of the proteolytic enzyme of Nepenthes,
Vines (6) found that pitcher liquor plus 0.25 percent hydrochloric acid
completely digested fibrin in the presence of bactericides such as potas-
sium cyanide, chloroform, thymol. The fibrin was dissolved when mer-
curic chloride (approximately 0.5 percent) was used as a bactericide;
however, the proteolysis was somewhat retarded. Fibrin was also
digested by pitcher liquor to which had been added 1 drop of concen-
trated hydrochloric acid and sufficient hydrocyanic acid to render the
concentration of the latter acid one percent. The liquor plus 0.25 percent
hydrochloric acid partially digested coagulated egg albumen in the
presence of potassium cyanide or thymol, the products of proteolysis
being detected by the biuret test.

The enzyme and its proteolytic power were destroyed by the action
of 1 percent sodium hydroxide for 1 hour, or of 5 percent sodium car-
bonate for three hours, at a temperature of 35 to 40° C; the solutions
were then neutralized, and tested for protease in the usual way, but
digestion of the substrate never occurred.

The influence of the following concentrations of hydrochloric acid on
the proteolysis was studied:— 1%, 0.5%, 0.25%, 0.125%. The optimum
acidity for digestion was found to be 0.25%.

The enzyme was isolated from 100 cc. of pitcher liquor by the follow-
ing procedure. An equal volume of absolute alcohol was added, then
phosphoric acid and lime water; the solution was neutralized with ammon-
ium carbonate; the precipitate was collected on a filter, and permitted
to drain over night. A portion of the precipitate was shaken with 10 cc.
of 0.25 percent hydrochloric acid, and the solution was filtered; the

OF INSECTIVOROUS PLANTS 427

filtrate digested fibrin, while a control experiment on fibrin plus 0.25
percent hydrochloric acid gave no digestion. After the precipitate had
been kept in a bottle with chloroform vapor for a month, the enzyme
contained in it was still active.

Vines was unable to find a zv-mogen of the protease in the pitcher
liquor, although the activation of a z\Tnogen was suggested by the
requirement that acid be added in order to produce proteolysis.

Glycerol extracts, prepared from the washed and dried tissue of
relatively young, vigorous pitchers, contained a protease which acted
in the presence of 0.25 percent hydrochloric acid, and retained its activity
for as long as 2 months, but not indefinitely.

Vines concluded that the digestive action of the pitcher liquor is due
to an enzyme. He detected albumose, but not peptone, as a product of
the digestion, and suggests that peptone was formed and immediately
split into other compounds. Wheat gluten was digested in the same
manner as fibrin.

Most of these experiments were made on N. mastersiana. The liquor
in the unopened pitchers of this species usually was distinctly acid.

In his next paper on the proteolytic enzyme of Nepenthes, Vines (7)
studied the action of heat on the protease. The pitcher liquor was
heated, then cooled; acid and fibrin were added, and digestion was
carried out as in the previous study. Control experiments were made on
the unheated pitcher liquor. Heating the liquor at 80° C. for 15 to 20
minutes did not destroy the enzyme, but the proteolytic power was
decreased to a marked degree. Thus in one experiment the heated liquor
completely digested the fibrin in approximately 4 days, while the un-
heated liquor in the control experiment required but 3 hours to digest
the fibrin completely. WTien the liquor was held at 78° to 83° C. for
30 minutes, the enzyme was completely inactivated; no digestion occurred
in 4 days, while the control experiment on unheated liquor gave complete
digestion of the fibrin in Ij/^ hours. When the liquor was kept at 80° C.
for 30 minutes, the enzyme was completely inactivated, and exerted no
proteolytic action on fibrin after digestion for 1 week; the unheated
control completely digested the fibrin in 5 hours. Boiling the liquor
"for some seconds" partly inactivated but did not completely destroy
the enz>Tne. It was necessary to subject the liquor to a temperature of
100° C. "for an appreciable time, say 3-5 minutes," in order to destroy
the enzyme completely.

The destructive action of sodium carbonate on the enzyme was also
studied. The degree of inactivation of the enzyme depended on the con-

428

HEPBURN— BIOCHEMICAL STUDIES

centration of the alkali, the period of time during which it acted, and the
temperature at which it acted. Sufficient soUd sodium carbonate was
added to the pitcher hquor to bring the salt to the desired concentration.
After the resulting solution had been incubated at the desired temperature
for the desired period of time, it was neutralized, then acidified with
hydrochloric acid, and a digestion experiment was made in the usual way
to determine the enzyme activity. The concentration of the sodium
carbonate varied between 0.5 and 5 percent; it was permitted to act on
the enzyme at a temperature of either 35° to 38° C. or 50° C. for a period
of time varying between 30 minutes and 17 hours. Each control experi-
ment on untreated liquor was incubated at the same temperature and for
the same period of time as its determination proper; then its enzymic
activity was determined.

The typical series of experiments in the following table may be
quoted. The substrate in each experiment was 0.01 gram of fibrin.

Series

% of sodium
carbonate

Time of incubation

with sodium carbonate

at 50° C.

Proteolytic activity
subsequently

a

b
c

5

1

Control

5

1

Control

1
Control

\]/2 hours
\}/2 hours

\}/2 hours
\]4, hours

1 hour

no digestion in 6 days
digestion complete in 4 days
digestion complete in 3J/^ hours

no digestion in 5 days
no digestion in 5 days
digestion complete in a few hours

no digestion in 4 days
digestion complete in a few hours

"On comparing the results of a, b, and c, it would appear that treat-
ment with 1% Na2C03 for one hour at a temperature of 50° C. is an
approximate index to the stabiUty of the enzyme. "

The pitcher liquor lost its acid reaction and its coloration, when
passed through a Berkefeldt filter. "It still retains some digestive
power, but is far less active than unfiltered liquid, the period of digestion
being more than doubled. " Filtration through a Berkefeldt filter caused
a marked loss of the enzymic power of solutions of pepsin (from the
stomach of the pig) and of salivary ptyalin. Hence the partial loss of
proteolytic power by the pitcher Hquor was due to the retention by the

OF INSECTIVOROUS PLANTS 429

filter of a true proteolytic enz>Tne, and not to the retention of proteolytic
bacteria, i.e., the liquor owed its proteolytic power to a true enzyme.

Experiments (usually, though not invariably, made on washed, un-
opened pitchers) showed that "under certain circumstances, previous
treament with acid causes the glands of the pitcher to yield a more active
glycerin-extract, or to yield an active extract when otherwise the extract
would be inactive; and it can only be concluded that this must be due to
the presence of a zymogen in the glands from which the enzyme is lib-
erated on treatment with acid. " The zymogen was best converted into
active enzyme by treatment of the glandular tissue with acid for a short
time at a relatively high temperature, say ^4 hour at 50° C. Action of
the acid for a longer time at lower temperature, not only activated the
enzyme but also extracted it from the glands. In the activation experi-
ments, 0.25 percent hydrochloric acid was used; 0.5 percent acetic acid
apparently was less satisfactory. One function of the high acidity of
the liquor in unopened pitchers probably is to activate the zymogen.

Peptone and leucine were recognized among the products of digestion.

Vines concludes that the enz\Tne is derived from a zymogen which is
present in the gland cells. The enzyme is a tryptic ferment, requires an
acid medium for its action, and resembles the proteases of germinating
seeds with respect to the reaction of the medium, which it requires, and
the products, which it forms.

In his final paper on the proteolytic enzyme of Nepenthes, Vines (8)
appHes to it the name "Nepenthin." Fibrin and Witte peptone were
subjected to a somewhat prolonged digestion at 38.5° C. with pitcher
Uquor, to which hydrochloric or citric acid had been added. The follow-
ing are typical experiments: —

I. 10 grams moist fibrin

50 cc. 0.3% hydrochloric acid
50 cc. pitcher liquor (from N. Mastersiana)
Period of incubation I8I/2 hours
II. 1 gram Witte peptone
0.2 gram citric acid
10 cc. distilled water
40 cc. pitcher liquor
Period of incubation — from noon until morning of the next day.

After digestion, the resulting solutions gave the tryptophane reaction
(a violet or pink color with chlorine water), which is stated to be charac-
teristic of tryptic digestion only.

430 HEPBURN— BIOCHEMIC\L STUDIES

The solutions, obtained by digestion, were evaporated to dryness,
the residue was extracted with absolute alcohol, and the alcohol was
removed by evaporation. This residue gave a biuret reaction. Hence
tests could not be made ior free tyrosin among the products of the pro-
teolysis, for the tyrosin reactions would have been given by the combined
tyrosin groups of the albumose and peptone, which were present in the
residue from the alcoholic solution.

Dubois (9) studied the pitcher liquor of the following species of
Nepenthes: — coccinea, distillatoria, Hookeriana, hybrida, maculate, phyl-
lamphora, Rafflesiana. Before the opening of the operculum, the pitcher
liquor of all these species was limpid, sUghtly viscid and slightly acid.
In opened pitchers, the Hquor generally was thick, contained whole
insects, and, at times, emitted a strong odor of putrefaction.

When liquor was removed from closed pitchers, which were about to
open, and was immediately placed in contact with cubes of coagulated
albumen, then incubated at the temperature of the atmosphere, or at a
temperature of 35° to 40° C, the albumen was not attacked; the hquor
remained limpid at the end of several hours. It was then filtered; the
filtrate contained no peptone. The experiment was repeated by trans-
ferring the hquor from closed pitchers to Pasteur culture tubes which
contained albumen cubes. The results were the same as before; the
angles of the cubes remained absolutely intact, and neither micro-
organisms nor putrefaction were present at the end of several days.

Liquor from pitchers, which had been open for but a short time, was
still clear. However, it attacked cubes of egg-white, quite rapidly at
ordinary temperatures, and very rapidly at the temperature of the incuba-
tor. The cubes became swollen, transparent and gelatinous, and lost
their angles; the Hquor became viscid, and a distinct odor of putrefaction
developed in some of the tubes. The hquor contained numerous micro-
organisms of different species and, after filtration, gave some of the
reactions of peptone. Many open pitchers contained insects, which were
in process of putrefaction and not in process of digestion.

The manner in which coagulated egg albumen behaved in the presence
of the pitcher Hquor — contaminated or not contaminated by micro-
organisms— led Dubois to the following conclusions: —

(1) That the Hquor does not contain any digestive substance (enzyme)
comparable to pepsin, and that Nepenthes are not carnivorous plants.

(2) That the phenomena of disintegration or pseudo-digestion, ob-
served by Hooker, were due without any doubt to the activity of micro-

OF INSECTIVOROUS PLANTS 431

organisms which came from without the pitcher, and were not due to a
secretion of the plant.

Couvreur (26) supported the conclusions of Dubois, and maintained
that the digestive phenomena observed by Vines were due to the action
of the reagents on each other, and not to the presence of a protease
in the pitcher liquor of Nepenthes.

Tischutkin (27), in a paper on Pinguicula, commented on the work
of Von Gorup and Will. He considered that the protease, which these
investigators found in the pitcher liquor of Nepenthes, was entirely of
bacterial origin.

Among the insectivorous plants studied by Tischutkin (10) was
Nepenthes Mastersi. The pitchers were stimulated by means of small,
sterile cubes of albumen. Even 24 hours later, the Uquor of the pitchers
contained myriads of bacteria, as was regularly shown by the direct
microscopic examination. The bacteria were isolated by cultures on
weakly acid nutrient gelatin and were tested for their peptonizing power;
several species were always found which dissolved small, sterile albumen
cubes fairly rapidly in an acidified menstruum.

The following experiments were also carried out. Incisions were
made in the side wall of pitchers, which had not yet opened and therefore
contained no bacteria; the liquor was removed with a pipette and con-
veyed to test glasses which contained water and small cubes of albumen
(1 cube in each glass); in some glasses the water was neutral, in other
glasses it had been acidlied. The experiments were carried out with
antiseptic precautions. The results showed that the pitcher hquor did
not contain a peptonizing enz\Tne, for the substrate remained unchanged
after incubation for 48 hours at 37.5° C. In order to overcome the
possible objection that the pitchers had been too young, the experiments
were repeated in a modified form. Openings were made in the wall of
pitchers, which had not yet opened; small, sterile cubes of albumen were
introduced into the cavities of the pitchers; the openings were closed;
and the plants were permitted to remain undisturbed. When the pitchers
opened, ^ days later, the albumen cubes were found unaltered; their
angles had not been rounded off; the Hquor had a strongly acid reaction
and contained no peptone; and bacteria were present "in very shght
number." When the hquor of these pitchers was placed in test glasses
and treated anew with small cubes of albumen, it dissolved the cubes
only after 4 to 5 days had passed, i.e., at a time when the bacteria had
already multiplied to a considerable degree.

432 HEPBURN— BIOCHEMICAL STUDIES

Tischutkin maintained that the digestion of protein in the liquor of
insectivorous plants is produced exclusively by the vital activity of
micro-organisms, which are always present in the secretion of the fully
developed plant, entering from the air and also with the bodies of the
insects, etc. The role of the insectivorous plants is therefore limited;
they secrete a medium favorable for the activity of the peptonizing micro-
organisms, and they make use of the products of this activity.

Goebel (11) made an elaborate research on the pitcher liquor. Pitch-
ers of Nepenthes paradisiaca (a hybrid) contained a clear, colorless, taste-
less fluid free from insects. Fibrin flocks were placed in the pitchers,
and also in water as a control. In both cases, after 6 days, the flocks
had been disintegrated into little shreds, and innumerable bacteria were
present. The liquor, in all the experiments, was either neutral or very
faintly alkahne, and contained no peptone. The liquor from the pitchers
gave no color with Nessler's reagent; the solution from the flasks, which
contained water and fibrin, gave a strong yellow color with this reagent.
Hence, in the pitchers, the nitrogenous compounds produced by the
proteolysis of the fibrin were absorbed, either as ammonia or as some
other compound.

Liquor was removed from the pitchers, in which the fibrin had been
digested, and was sown on nutrient gelatin. The gelatin became lique-
fied to a marked degree in 2 days and acquired a green fluorescence
which is characteristic of Bacillus fluorescens liquefaciens. Neither
bacteria nor moulds were found when liquor from unopened pitchers was
inoculated into the gelatin.

Liquor was collected from unopened pitchers of N. paradisiaca; it
exerted very little digestive action after 0.2 percent hydrochloric acid
had been added, forming very little peptone; therefore very little enzyme
was present.

The absorption of ammonia by the pitcher was demonstrated by the
following procedure. A definite volume of aqueous solution of ammonia
(containing 1 part of ammonia in 20,000 parts of solution) was placed
in a pitcher, an equal volume of the solution was placed in a glass vessel
in the Nepenthes house as a control. Twenty-four hours later, the volume
of Hquid within the pitcher was practically unchanged. However,
Nessler's reagent produced no precipitate with the contents of the pitcher,
and formed a thick precipitate with the control experiment; therefore
ammonia had been absorbed by the pitcher. Further experiments
demonstrated that one-half of the ammonia was absorbed during the
first three hours, and that all the ammonia had been absorbed at the end

OF INSECTIVOROUS PLANTS 433

of 6 hours; in these experiments Nessler's reagent was used for the
colorimetric determination of ammonia.

Pieces of meat, the size of grains of rye, were introduced into pitchers;
after 2 days, the contents of 3 pitchers were acid in reaction, those of
5 pitchers neutral in reaction; after another week had passed, the con-
tents of but 1 pitcher had an acid reaction which, however, was not very
marked.

A vigorous plant of Nepenthes paradisiaca was cultivated in a glass
chamber at a temperature of 20° to 25° C. in an atmosphere saturated
with moisture. The plant bore 3 pitchers. The oldest of the pitchers
was brownish and no longer vigorous, and contained a small amount of
neutral liquor. A wasp was introduced into the liquor and soon died;
3 days later the liquor had an alkaline reaction; bacteria and infusoria
were present in abundance. The second pitcher contained an acid
liquor, in which a small fly was present. This liquor dissolved fibrin
(which had been stored in glycerol, washed, and strongly swollen) in 1
hour; after 3 hours, soluble protein was absent, but peptone was present;
the temperature of incubation was 25° C. Another fibrin flock and 0.2
percent hydrochloric acid were added, and the temperature of incubation
was changed to 16° to 18° C; the flock dissolved in 40 minutes; this solu-
tion was sown on nutrient gelatin, and bacteria were not found.

The youngest pitcher was still closed. Inoculation of its contents
into nutrient gelatin gave no growth. The liquor was neutral in reaction
and mucilaginous. A flock of swollen fibrin and 0.1 percent formic acid
were added to the liquor; the flock was completely digested in 12 hours;
even after 8 days, bacteria were absent as was shown by a nutrient gelatin
culture, which was made in duplicate. It was also found that 0.1 percent
formic acid prevented the development of putrefactive bacteria in an
approximately 0.5 percent peptone solution, which was exposed in the
open air for 8 days, and that only a few mold-threads developed. When
the acid had not been added to the peptone solution, clouding and an
unpleasant odor soon appeared, due to the development of innumerable
bacteria.

Goebel interprets these observations: —

" These facts in themselves suffice to refute the acceptance of a bacterial
digestion. They show that, in the Nepenthes hybrid studied, even in
the unopened pitcher, a peptonizing enzyme is present which exerts an
energetic digestive action on the addition of acid. Normal pitchers,
into which an insect falls, very soon secrete formic acid. With excessive
feeding, of course, the enzyme action is insufficient and putrefaction

434 HEPBURN— BIOCHEMICAL STUDIES

may occur even in otherwise normal pitchers; it is probable, but not
certain, that an increased secretion of enzyme occurs in the opened
pitchers in the presence of stimulating substances." However, an in-
creased secretion could not be detected in a pitcher stimulated by a
fibrin flock, and compared with an unstimulated pitcher as a control;
but one such experiment was made.

The tests for a protease in the glands of the pitcher-cover, which
secrete nectar, were entirely negative. Flocks of fibrin were fastened
over the glands by means of filter paper, which was kept moist. Diges-
tion of the fibrin did not occur.

The glycerol extract of the pitchers contained too little enzyme and
showed no proteolytic activity. Apparently the procedure of Vines (5)
for the detection of enzymic activity was followed.

The secretion of unopened pitchers of Nepenthes paradisiaca was
neutral in reaction. The hquor of unopened pitchers of Nepenthes
M aster siana was strongly acid; when fibrin was introduced into these
pitchers, it was dissolved in 3 days, and cubes of albumen were strongly
attacked; the protein was peptonized in the pitcher, and no bacteria
were present.

A pitcher of Nepenthes Sedeni contained a strongly acid liquor and
dissolved fibrin in 25 hours. Similar observations were made with a
pitcher of Nepenthes robusta. Cultures demonstrated the absence of
bacteria. The filtered solutions from the pitchers plus 0.1 percent
formic acid digested meat fibres in 5 hours at a temperature of 35° C.

Goebel states that free formic acid could be shown to be present in
the secretion (pitcher hquor) of the species of Nepenthes studied by him-
self. The total acidity of diff'erent pitchers, calculated as formic acid,
was:— 0.036%, 0.025%, 0.021%. All bacteria are not killed by this
degree of acidity.

The strongly acid pitcher-Hquor of Nepenthes Mastersiana was sown
on non-acid nutrient gelatin, and on nutrient gelatin rendered acid by
the addition of 0.2 percent tartaric acid. After 8 or 9 days, no growth
was noted, or else a few bacteria and molds, which were doubtless due
to air-contamination.

Goebel considered that a true enz\Tnic digestion occurred in normal
pitchers of Nepenthes, in which the Hquor had not been diluted by water;
this dilution often occurs in greenhouses. The enzjnne was classified as
a peptonizing enz\Tne, not identical with pepsin, and different from the
pancreatic protease. Bacterial digestion could occur only in the liquor

OF INSECTIVOROUS PLANTS 435

of enfeebled pitchers, which possessed a neutral or alkaline reaction; it
was recorded as much slower than the normal digestion, although its
products could be absorbed, in part at least, by the plant, provided the
putrefaction had not injured the pitcher. Goebel stated that the pitch-
ers of enfeebled plants frequently contain only water, and that such plants
may give results like those obtained by Dubois and by Tischutkin.

The absorption of peptone by the pitchers was demonstrated. Six
pitchers of N. M aster siana were washed out; and a known volume of a
0.1 percent solution of peptone, containing 0.1 percent formic acid, was
placed in them. At the end of 66 hours, the pitchers were emptied.
Resorption and evaporation of the liquid had been but slight, it was clear,
colorless, free from bacteria, and contained only a trace of the mycelium
of mold. The formic acid content had not decreased; however, but a
faint trace of peptone remained, and an enzyme could not be detected.
Hence the peptone had been absorbed. These pitchers again secreted
liquor, and digested small pieces of meat.

Clautriau (12) conducted experiments on plants of Nepenthes ynel-
amphora in their native habitat in Java. The liquor in non-stimulated
pitchers was neutral to litmus. When an unopened pitcher was merely
shaken, its liquor had a more or less acid reaction on the following day.
The reaction became acid after fine glass tubes, 1 to 2 cm. in length, were
dropped through the lid into the pitcher, or after 2 or 3 drops of tincture
of litmus were added to the liquor within the pitcher. Hence sUght
stimulation sufficed to cause the appearance of acid. The strongest
acid reaction in the stimulated pitchers was about equal to that possessed
by a solution, obtained by diluting 2 cc. of slightly fuming hydrochloric
acid with sufficient water to render the final volume 1 liter. The liquor
contained in solution a substance which was apparently thermolabile;
this substance caused insects to become wet and to sink. The liquor
did not contain a poison which kills insects. The insects were finally
digested, leaving only a residue of chitin; putrefaction did not occur.
The pitchers were sensitive to the presence of even exceedingly small
quantities of antiseptics, such as formaldehyde, chloroform, spirit of
camphor, essence of peppermint and of lemon, in the pitcher liquor;
these reagents caused a cessation of the secretion of acid and of digestion,
and the pitchers died in a few days.

When coagulated egg-white was introduced into the pitchers, diges-
tion and absorption occurred. If but a small quantity of egg-white was
used, absorption equalled digestion; proteolytic products did not remain
in the pitcher, and a substrate for bacteria was lacking. If a large

436 HEPBURN— BIOCHEMICAL STUDIES

quantity of egg-white was used, the unabsorbed products of digestion
accumulated, and the pitcher was invaded by bacteria.

When a sterile solution of egg-white, prepared as follows, was intro-
duced into the pitchers, digestion occurred; and bacterial invasion and
putrefaction were rarely noted. Ten cc. of white of egg and 90 cc. of
water were shaken together, to break up the membranes in the white
and to dissolve the albumin. The solution was filtered and 0.1 milli-
gram of ferrous sulphate was added, i.e., 2 drops of a freshly prepared,
0.1 percent solution of ferrous sulphate. If the egg had not been fresh,
more iron was added, but never an amount in excess of 1 milligram of
ferrous sulphate. The solution was then boiled; it usually remained
clear and limpid, but at times showed a faint opalescence. After con-
ducting a digestion experiment with this solution as the substrate, the
lindigested albumin was removed by the following procedure. An alka-
line salt was added to the solution, which was then acidified very quickly,
and the albumen was coagulated with heat.

In most vigorous pitchers, all the protein was digested at the end of
two days. Thus 5 cc. of the solution of egg-white, described above, was
introduced into a pitcher. At the end of two days the pitcher liquor
gave no precipitate on neutralization, and on boiling in the presence of
salts or of acids; it yielded merely a trace of a precipitate with the follow-
ing reagents:— potassium mercuric iodide, acetic acid and potassium
ferrocyanide, phosphomolybdic acid.

The plant itself played an important role in the digestion. Experi-
ments were made in vitro with liquor from both unopened and open
pitchers, using- chloroform as a bactericide. Absolutely no digestion
of the substrate occurred. In one experiment, liquor, which possessed
digestive power while in the pitcher, digested the substrate in vitro; the
albumen disappeared, and much albumose (proteose) and possibly a
httle peptone were formed. Separation of the pitcher from the plant
during the course of digestion inhibited the digestion of the albumin.

Clautriau also conducted experiments at Brussels on Nepenthes
which had been cultivated in greenhouses.

He used the following technique in studying the products of the
proteolysis. After digestion, the solution was neutralized with dilute
sodium hydroxide solution in order to separate the syntonin (acid meta-
protein), which was collected on a filter. To the filtrate were added an
equal volume of saturated solution of sodium chloride and a trace of
acetic acid; the solution was boiled to coagulate the albumin, which was
then removed by filtration. The filtrate was saturated, while hot, with

OF INSECTIVOROUS PLANTS 437

ammonium sulphate, first while acid, then while alkaline in reaction; the
resulting solution was permitted to stand; and the albumoses (proteoses)
collected, and were removed by filtration. In the filtrate, the ammonium
sulphate was removed by means of barium carbonate, and the excess
of barium was removed with sulphuric acid. The solution was then
filtered and concentrated; it contained no albumose, for it did not form
a precipitate with potassium mercuric iodide, or with acetic acid and
potassium f errocyanide ; it contained peptone, for it responded to the
biuret test, and gave precipitates with tannin, phosphotungstic acid,
and phosphomolybdic acid.

The liquor in a pitcher of Nepenthes Mastersiatia, containing insects,
was filtered and used in digestion experiments m vitro. Three cc. of
the liquor and 20 drops of albumin (egg-white) solution, prepared as
described above, were permitted to react for 3 days, with and without
the addition of 1 drop of hydrochloric acid (1 drop contained 0.01 gram
of hydrochloric acid), in the presence of camphor as a bactericide. The
albumin was completely digested to peptone in both the presence and the
absence of the hydrochloric acid, while a blank experiment, heated for
10 minutes at 100° C. at the beginning of the experiment, contained no
peptone. The digestion in vitro by the liquor plus hydrochloric acid
was far more rapid at 37° C. than at 20° C.

Experiments were made in vitro with the liquor of non-stimulated
(unopened) pitchers of Nepenthes coccinea and of a Nepenthes similar to
iV. phyllamphora. Albumin was the substrate, and hydrochloric acid
was added. After digestion in the incubator for 5 or 6 days, syntonin
and a little albumose were present, but no peptone, and enzyme action
probably had not taken place. However, Clautriau hesitated to advance
the proposition that the secretion of the enzyme, like that of the acid,
is the result of stimulation, although the two experiments were con-
cordant.

The products of proteolysis were rapidly absorbed by the pitcher.
Successive and abundant additions of albumin were supported perfectly,
without inconvenience. In four days, the pitcher of Nepenthes Master-
siana, mentioned above, digested 2.5 cc. of albumin solution, and com-
pletely absorbed the products. Then 10 cc. of albumin solution (nitrogen
content determined by the Kjeldahl method as ammonia equalled
14 cc. of 0.1 N sulphuric acid) were introduced into the pitcher; 7 days
later the liquor contained nitrogen equal to but 2.8 cc. of 0.1 iV sulphuric
acid. Another 10 cc. portion of albumin solution was added; at the end
of 7 days the nitrogen content of the liquor equalled but 2.7 cc. of 0.1 N

438 HEPBURN— BIOCHEMICAL STUDIES

sulphuric acid. For the third time, 10 cc. of albumin solution were
placed in the pitcher, and its protein was also digested. It should be
noted that a portion of the nitrogen, found in the liquor after the albumin
had been digesting for a week, was due to the enzyme and to the chitin
of insects.

Digestion also took place in the pitcher of Nepenthes coccinea. In but
two instances was peptone found in the pitchers after digestion of
albumin, once with a plant of Nepenthes coccinea, and once with a plant
of Nepenthes from Borneo similar to N . phyUamphora. The peptone,
it is stated, is diffusible, and probably is absorbed.

Clautriau classified the protease of the pitcher liquor as a pepsin,
since it acted only in an acid medium, and formed true peptone as the
ultimate product of its action. Leucine, tyrosin, and other crystalline
compounds were not found among the product? of proteolysis.

An amber color, becoming red with alkali, was. very frequently noted
in the liquor after digestion. It w^as ascribed to the presence of tannin
derived from the glands, and not to the presence of tryptophan.

The pitcher liquor did not contain an amylase, for it had no action
on starch paste, when the mixture of liquor and paste was digested for
5 days.

Stimulation was found necessary to excite an abundant secretion of
both the acid and the protease. The glands, which secrete both the acid
and the enzyme, are said to absorb the digested protein. The micro -
chemical reactions of the proteins were more intense in the region of the
glands, hence it was concluded that the peptone was absorbed by the
glands and stored as protein.

Since the liquor, obtained from healthy pitchers of Nepenthes mel-
amphora in its natural habitat, failed to produce proteolysis in vitro, it
was suggested by Clautriau that the pitchers of this variety possibly
absorb either albumin or albumose. He also pointed out that, if too
many insects accumulate in a pitcher, they may be decomposed by
bacteria without injury to the plant, which probably utilizes the ammonia
and amino acids formed by the bacteria.

Clautriau looked on the digestion in the pitcher as a source of nitrogen,
and possibly of mineral food, for the nutrition of the plant.

Fenner (13) used Nepenthes Rafflesiana Jack in his study of the pro-
teolysis in the pitchers of Nepenthes. He states that the innumerable
glands of the pitcher lie in niches which open downwards, and are from
one-half to two- thirds covered by a projecting, roof-like, epidermal
structure. Wet insects, climbing up the wall to escape from the pitcher,

OF INSECTIVOROUS PLANTS 439

come under the "roof" of a gland, and at the same time come in contact
with the gland itself, and adhere to it. The gland now secretes a slight
quantity of a sticky mucilage, which is more viscous than the other
secretion of the glands — the pitcher liquor, that is found even in unopened
pitchers. The mucilage dissolves that part of the insect-body which can
be utilized by the plant. When the surface of the gland becomes dry
by absorption, the undissolved insect -residue falls from the gland, and
is usually washed down by the pitcher liquor, either as its level rises or
through the swinging motion of the pitcher, or its leaf. The resulting
sediment consists for the most part of chitin. The glands on the border
of the level of the pitcher liquor come in contact with the insects in this
manner, and the cells of these glands show aggregation -phenomena in
consequence of the absorption of organic substances, while the cells of
glands above the level of the liquor, or beneath its level and merely
washed by it, usually have unclouded contents.

When a number of gnats were placed in a pitcher, which was just
opening, they swam on the surface of the liquor and came in contact
with the glands at its level. The pitcher was emptied, a portion of the
glandular region was dried, and some gnats were placed on the dry place;
the secretion occurred only after the course of 4 to 6 hours, was sUght
in amount and insufficient to digest the insects, but dried up about them.
When an insect was wet with the pitcher liquor and then brought on the
dry place in the glandular region, very soon secretion of the mucilage,
digestion, and absorption took place, so that only the residue of chitin
was left at the end of 5 to 8 hours. The liquor, used for wetting the
insect, must not have been diluted by water; since water may enter the
pitcher during the watering of the plant, the liquor should be taken from
a pitcher from which water has been excluded by means of a cotton plug.

Greenhouse plants are abnormal in that they usually contain only a
slight quantity of liquor, secreted by the glands; as a consequence the
greatest portion of the glandular region is not wet by the liquor. Fenner
used suitable control experiments in his research.

From his own experiments, and from the observations of Goebel(ll)
(see above), Fenner draws the following conclusions: —

(1) Normal pitchers, in which insects are found, contain a faintly
acid liquor (formic acid according to Goebel), which acts upon the glands
as a chemical stimulant when insects, wet with it, come in contact with
the glands.

(2) The liquor thus gives rise to digestion, so that insect bodies, which
are saturated with it, can rapidly be dissolved and absorbed if they come

440 HEPBURN— BIOCHEMICAL STUDIES

upon the glands of the pitcher wall beneath the roof-like epidermal
structures.

(3) Insects, which enter the pitcher liquor, can be completely dis-
solved with the exception of their chitin plates; the latter are found as a
sediment.

Fenner considered it a question for further study whether (a) the
dissolved products of the digestion were withdrawn from the liquor and
absorbed by the glands, or (b) their solution in the liquor was absorbed
as such by the glands. Apparently the same glands, that secrete the
pitcher liquor and the digestive mucilage, absorb the products of diges-
tion.

Abderhalden and Teruuchi (28) studied the action of the pitcher
liquor of Nepenthes on the dipeptide glycyW-tyrosin. This peptide is
quite soluble in water, and is not cleaved by pepsin-hydrochloric-acid
but is readily spht by tr}^sin into its components, glycine (glycocoll) and
/-tyrosin; the latter compound is very difScultly soluble in water, and
precipitates.

Liquor was obtained from pitchers with closed lids, and from open
pitchers which appeared to contain no very large quantity of condensa-
tion water. It was viscous, neutral in reaction, and exerted a very slow
but distinct proteolytic action on fibrin flocks; after digestion for 3 days,
the fluid was removed by filtration and then gave a distinct biuret
reaction.

One gram of glycyl-/-tyrosin was dissolved in 10 cc. of liquor, collected
from several pitchers,- and toluene was added as a bactericide. The
solution was kept at room temperature for 7 days; a slight cloudiness
occurred, but no precipitate formed. Then the solution was transferred
to an incubator; the opalescence did not increase; and a precipitate did
not separate, even when the solution was evaporated to half its original
volume.

The conclusion was drawn that the protease of Xepenthes is not a
tr}'psin. ''It therefore seems that the flesh-eating plants, Nepenthes,
do not act through a tr>'psin-like enzyme. We do not venture to record
our finding as a certainty, since from lack of material, it was not possible
for us to repeat the experiment under different conditions. "

Robinson (14) introduced various solutions into pitchers of Nepenthes
distiUatoria and noted their,action upon the pitchers. A dilute (M/1024)
solution of potassium nitrate had exerted no injurious action at the end
of 9 days, but the pitcher began to wither at the end of 12 days. The
nutrient solution of Sachs (calcium nitrate 6 grams, potassium nitrate

OF INSECTIVOROUS PLANTS 441

1.5 grams, dipotassium phosphate 1.5 grams, magnesium sulphate 1.5
grams, ferrous sulphate a trace, water 6000 cc.) caused a withering of
the tissues on about the eighth day. When a dilute solution of Liebig's
meat extract was introduced, and the plants then observed for a period
of two weeks, the contents of the pitchers did not become foul, and the
pitchers did not decay. A 10 percent solution of glucose, kept in the
pitchers for 4 da}s, apparently had no harmful effect on the plant; the
solution retained its power to reduce Fehling solution.

Robinson also made tests concerning the secretion of enzymes by the
pitchers. A 10 percent solution of sucrose, which had been in the
pitchers for 4 days, failed to reduce Fehling solution, whence it was
inferred that an invertase is not secreted by the pitchers. A thin starch
paste was kept in the pitchers for 4 days; it then had no reducing action
on Fehling solution, and " the iodine test showed that the starch granules
in the paste had not been broken down "; therefore the pitchers do not
secrete an amylase (diastase).

Neutral olive oil and water were mixed in the proportion of 0.4 cc.
of oil and 9.6 cc. of water. After the mixture had been thoroughly
shaken, it was introduced into pitchers, and permitted to remain in them
for a period of from 4 to 7 days, then removed and titrated with 0.01 N
potassium hydroxide solution, using phenolphthalein as an indicator.
Control experiments were carried out (a) with the emulsion of oil in
vitro, and (b) with the emulsion plus toluol in the pitchers. The results
indicated that hydrolysis of the oil did not occur, and that the pitcher
does not secrete lipase. Tap water, which had been kept in the pitchers
for 1 day, was used in an experunent with ethyl butyrate; 2 cc. of this
water and 4 drops of the ester were allowed to react at room temperature
for 24 hours; the hydrolysis of the ester was determined by titration with
0.01 N fixed alkaline hydroxide, using phenolphthalein as an indicator.
Enzymic cleavage of the ester was not detected, hence an esterase was
not present.

Jenny Hempel (15) found that "the sap of the stimulated pitcher of
Nepenthes gives values for the hydrogen ion concentration greater than
10"", but unstimulated pitchers give no definite value. "

Shibata and Nagai (16) noted the presence of flavone in the leaves
and the flowers of Nepenthes phyllamphora Willd. More flavone was
present in N. phyllamphora, growing in the open on the island of Yap,
than occurred in N. Mastersiana, growing under glass at Tokyo.

Pfeffer (17) has suggested that the insectivorous plants derive both
nitrogen and phosphorus from their prey.

442 HEPBURN— BIOCHEMICAL STUDIES

Nepenthes have found application in medicine. "The water ab-
stracted from their leaf pitchers is an article of commerce in the East
Indies. The leaves and root of N. Boschiana, Krthls. are especially em-
ployed, chiefly as an astringent. "(18)

II

A Study of the Protease of the Pitcher Liquor of Nepenthes

BY

Joseph Samuel Hepburn, A.M., M.S., Ph. D.

Broadly speaking, two hypotheses exist concerning the mechanism
of the proteolytic digestion within the pitchers of Nepenthes. One
view is that digestion results from the action of a protease, secreted by
the pitchers. The other view is that digestion is due to bacterial action.
A third factor to be considered is the autolysis of the tissues of the
dead insects.

In the present study, the volume of the liquor secreted by a single
pitcher was always so small, that liquor could not be obtained from the
same pitcher both before and after stimulation. Very rarely indeed did
two pitchers mature on the same plant at the same time, thereby permit-
ting a comparative study of the liquor from both stimulated and non-
stimulated pitchers of the same plant. Differences due to individual
plants could not be entirely eliminated, but the problem was attacked by
several methods for the study of proteolysis, and a number of experiments
were made according to each method; the results obtained by all the
methods lead to the same general conclusions.

Material for this research has been obtained from the following species
and hybrids of Nepenthes, grown in the Nepenthes House of the Univer-
sity of Pennsylvania: — ampullar ia, atrosanguinea, Chelsonli, Claytonii,
Dominii, Dyerinana, gracilis, Hamiltoniana, Henryana, Hookeriana, Mas-
tersiana, mixta, Morganiana, paradisae, Rafflesiana pallida, rufescens,
splendida, Wittei.

Pitchers were always selected prior to opening. They were closely
watched; and the mouth of each pitcher was closed with absorbent cotton
as soon as the lid opened; the entrance of insects was thereby prevented;
and possible contamination of the pitcher liquor by the tissue enzymes
of the digested prey was entirely excluded.

OF INSECTIVOROUS PLANTS 443

When the liquor from non-stimulated pitchers was studied, it was
used as soon as possible after the opening of the pitcher.

When liquor from stimulated pitchers was desired, recourse was had
to mechanical stimulation by chemically inert substances. In some
experiments, the glands of the pitcher were stroked repeatedly with a
camel's hair brush, and the cotton plug was then inserted; the liquor
was removed for study on the following day. In other experiments,
several small, round, solid glass beads, such as are used in fractionating
columns, were inserted into the newly opened pitcher; the cotton plug
was introduced; and the pitcher and its contents were shaken thoroughly
at intervals during one or more days, taking care not to wet the cotton
and thereby lose liquor; the liquor was finally removed for study.

In all the tests for the presence of a protease, a bactericide was used
in order to exclude completely the action of micro-organisms. In some
experiments, sufficient sohd sodium fluoride was added to the mixture
of pitcher liquor and substrate to render the final concentration of the
fluoride 1 percent. In other experiments, a sufficient volume of a con-
centrated (2 percent.) aqueous solution of trikresol was added to render
the final concentration of that bactericide 0.2 percent. This concen-
tration of trikresol was found satisfactory by Graves and Kober (19)
in certain of their experiments with proteases. When the mixture of
pitcher liquor and substrate was diluted to a definite volume, the con-
centrated trikresol solution was added before the dilution to the final
volume was made.

In each experiment, a blank or control determination was carried
out, using pitcher liquor which had previously been boiled, then permitted
to cool to the temperature of the room; the control was made in exactly
the same manner, in all other respects, as the determination proper.
The control or blank was always compared with the determination
proper, and due allowance was thus made for the possible action of any
thermostable catalyst present in the pitcher liquor, and also for any
action of the reagents on each other.

The following reactions for the detection of a protease were used: —

(1) The formol-titration of Sorensen.

(2) The digestion of: —

(a) carmine fibrin,

(b) edestan,

(c) protean derived from castor bean globulin,

(d) ricin (Jacoby).

444 HEPBURN— BIOCHEMICAL STUDIES

(3) The cleavage of glycyltryptophane.

The temperature of incubation was 37°C., unless otherwise stated.

Formol-titration

The following substrates were used: — ovalbumen, fibrin, edestin,
ovomucoid, Nahrstofif-Heyden, and Witte peptone. The fibrin was
prepared from ox blood; the ovomucoid was obtained by the procedure
of Eddy (20); the Nahrstoff-Heyden, according to Gotschlich (21), was
a mixture of different albumoses.

After incubation, any insoluble protein was removed by filtration,
and was washed on the filter; the combined filtrate and washings were
made neutral to phenolphthalein. If metaprotein separated, it was
filtered out and washed on the filter; the filtrate and washings were
again made neutral to phenolphthalein — if necessary — and one-half
of their volume of formol (40 percent, formaldehyde), previously ren-
dered neutral to phenolphthalein, was added. The basic amino group
in the amino acid molecule was thereby converted into its methylene
derivative by condensation with the formaldehyde, and no longer neu-
traUzed the acidic carboxyl group in the same molecule. This carboxyl
group now functioned as an acid, giving the solution a reaction acid
to phenolphthalein. This acidity, due to amino acids, was immediately
titrated with standard fixed alkaline hydroxide, using phenolphthalein
as the indicator, and served as a measure of the proteolysis. Usually 0.1 iV
sodium hydroxide was used for the titrations; however, 0.05 N sodium
hydroxide was used when it was expected that the proteolysis would
be slight on account of the small volume of pitcher liquor used.

The following experiments were made with the liquor from stimulated
pitchers.

Ovalbumen (0.05 gram) was digested with 15 cc. of pitcher liquor
for 3 days. After the addition of formol, the determination proper
required 0.15 cc, the blank experiment 0.00 cc. 0.05 X sodium hydrox-
ide for the neutralization of the amino acids.

Fibrin (0.05 gram) was digested with 15 cc. of pitcher liquor for
14 days; on titration after the addition of formol, the determination
proper required 0.45 cc, the blank 0.00 cc 0.1 N sodimn hydroxide.

Ovomucoid (0.05 gram) was dissolved in 10 cc of water, then in-
cubated with 5.5 cc. of pitcher liquor for 6 days. The formol titration
was:- determination proper 0.10 cc, blank 0.00 cc. 0.1 N sodium
hydroxide.

OF INSECTIVOROUS PLANTS 445

Nahrstoff-Heyden (15 cc. of a 1 percent, aqueous solution) was
mixed with 15 cc. of pitcher liquor, then incubated for 3 days. The
formol titration was:- determination proper 0.30 cc, blank 0.10 cc.
0.1 N sodium hydroxide.

Witte peptone (25 cc. of a 1 percent, aqueous solution) was mixed
with 25 cc. of pitcher liquor, then incubated for 4 days. The formol
titration was: — determination proper 4.60 cc, blank 1.85 cc. 0.1 .Y
sodium hydroxide.

A solution of edestan was prepared by dissolving 0.1 gram of edestin
in 15 cc. of 0.1 .Y hydrochloric acid, previously diluted to 50 cc. with
water. The pitcher liquor (8 cc.) was mixed with 25 cc. of this solu-
tion (equal to 0.05 gram of edestin), and then incubated for 25 days.
The formol titration was: — determination proper 0.90 cc, blank 0.00
cc. 0.1 A' sodium hydroxide.

When liquor from non-stimulated pitchers was used, the following
results were obtained.

In three experiments, the period of incubation was 4 days. In the
first experiment, 10 cc. of pitcher liquor and 0.05 gram of ovalbumen
were used; in the second experiment, 12.5 cc. of pitcher liquor and 25
cc. of a 1 percent, aqueous solution of Nahrstoff-Heyden; in the third
experiment, 12.5 cc. of pitcher liquor and 25 cc. of a 1 percent
aqueous solution of peptone (Witte). In all three experiments, the
formol titration of the determination proper was the same as that of the
blank, showing that enzymic cleavage of the substrates had not occurred.

A solution of edestan was prepared by dissolving 0.1 gram of edestin
in 15 cc. of 0.1 N hydrochloric acid, previously diluted with water to a
volume of 25 cc. In one experiment, 25 cc. of the edestan solution
were mixed with 11.5 cc. of pitcher hquor; sufficient water was added to
make a total volume of 50 cc, and the solution was digested for 28 days.
In another experiment, 9 cc. of the edestan solution were added to 8 cc
of pitcher liquor; sufficient water was added to make a total volimie of
25 cc, and the solution was incubated for 21 days. In both experiments,
both the determination proper and the blank remained neutral, after
neutral formol had been added in the formol titration, hence digestion
with the production of soluble proteolytic products had not occurred.

Carmine Fibrin

The directions of Grutzner (22) for the use of this reagent were some-
somewhat modified. The carmine fibrin was washed with water to re-
move the glycerol, in which it had been preserved, then was permitted to

446 HEPBURN— BIOCHEMIC.\L STUDIES

swell in 0.2 percent, hydrochloric acid, to which sufficient trikresol had
been added to produce a 0.2 percent, solution of that bactericide. The
swollen, gelatinous carmine fibrin was placed in the pitcher liquor;
sufficient hydrochloric acid (0.6 percent, solution) and trikresol (2
percent, solution) were added to make a concentration of 0.2 percent,
of each of these reagents in the resulting solution. The temperature of
incubation was that oj the room. The occurrence of digestion was made
knowTi by two phenomena— (1) the flocks of carmine fibrin decreased
in size and finally dissolved completely; and (2) the carmine was thereby
liberated, dissolved, and imparted a red color to the solution.

In the preliminary series of experiments, the carmine fibrin was
swollen in one mass, and a definite volume of the gelatinous reagent was
used in each experiment. Three experiments were made in each of which
the liquor from a single, stimulated pitcher was used. In the
first experiment, the pitcher hquor (1.5 cc.) completely dissolved 0.1
cc. of carmine fibrin in 13 hours. In each of the other experiments,
4 cc. of pitcher Hquor were permitted to act on 0.5 cc. of carmine fibrin;
in one of these experiments, the substrate was partially dissolved m 15
hours, and completely dissolved in 24 hours; in the other experiment,
the substrate was markedly digested in 48 hours, and completely dissolved
in 6 days. In still another experiment, 4.75 cc. of liquor, collected from
several stimulated pitchers, completely dissolved 0.25 cc. of carmine
fibrin in 26 hours.

Liquor (1.5 cc.) from a single non-stimulated pitcher completely dis-
solved 0.1 cc. of carmine fibrin in 13 hours. The liquor (4.75 cc.) from
several non-stimulated pitchers produced marked digestion, but not com-
plete solution, of 0.25 cc. of the same substrate in 31 hours.

In the final series of experiments, the liquor from a separate pitcher
was used in each experiment, and the carmine fibrin (0.2 gram) for each
experiment was weighed out into a separate tube.

One set of experiments was conducted on liquor from stimulated
pitchers. The carmine fibrin for each experiment was swollen in its tube
in the usual manner, then was placed in the pitcher liquor, and hydro-
chloric acid and trikresol were added as described above. The time
required to dissolve the swollen substrate was: —

Pitcher A, 3.5 cc. liquor, 48 hours.
" B, 2.5 cc. " 72 "
» C, 2cc. " 93 "

" D, 3.5CC. " 111 "
" E. Icc. " 133 "

OF IXSECTIVOROUS PLANTS 447

Another set of experiments was made with hquor from non-stimulated
pitchers; unswoUen carmine fibrin was used, and no acid was added to
the reaction-mixture. The pitcher hcjuor was placed on the substrate,
and sufficient trikresol (2 percent, solution) was added to give a con-
centration of 0.2 percent, of the bactericide. Six experiments were
made, the volumes of pitcher liquor being 1.0, 1.5, 2.5, 3.0, 3.5 and 5.5
cc. respectively. Even after 70 days, the substrate was absolutely un-
attacked. The supernatent liquid had assumed a very faint pink tinge,
no more marked than that of a blank experiment.

In a third set of experiments, unswollen carmine fibrin was incubated
with liquor from non-stimulated pitchers, in the presence of both hydro-
chloric acid and trikresol, as described above. Two experiments were
made, the volumes of pitcher liquor being 2.5 and 1.0 cc. respectively.
The substrate was markedly digested, in the first experiment in 16
hours, in the second experiment in 52 hours.

Edestan

The solution of edestan, used in these experiments, was prepared
by dissolving 0.1 gram of edestin in 15 cc. of 0.1 N hydrochloric acid,
previously diluted to 25 cc. with water. After the mixture of edestan
solution and pitcher liquor had been incubated under the conditions
stated below, it was neutralized with 0.1 .Y sodium hydroxide, using
phenolphthalein as the indicator.

Liquor from stimulated pitchers was used in the following experiments:

The pitcher liquor (20 cc.) was mixed with 25 cc. of edestan solution;
the mixture was diluted to a volume of 50 cc. with water, and was in-
cubated for 14 days. Then the solution was neutralized. The deter-
mination proper gave absolutely no precipitate, showing that proteolysis
had occurred, and that both the protean, edestan, and the meta-protein,
which is one of the first products of proteolysis, had been converted
into simpler proteolytic products. In the blank, on the other hand,
a voluminous precipitate formed.

In a second experiment, 1 cc. of pitcher liquor and 2 cc. of the edes-
tan solution were mixed ; and the mixture was diluted with water to a
volume of 5 cc, then incubated for 8 days. On neutralization, the
determination proper failed to give a precipitate, while the blank yielded
a voluminous precipitate. Hence the edestan had been digested.

Experiments were also made, using the liquor from non-stimulated
pitchers : —

448 HEPBURN— BIOCHEMIC.\L STUDIES

In one experiment, 1 cc. of liquor, 2 cc. of the edestan solution,
and sufficient water to render the total volume 5 cc., were mixed; and
the mixture was incubated for 8 days. On neutraliztion, a voluminous
precipitate formed in the blank; the precipitate, which formed in the
determination proper, was about one -half as great as that in the blank,
showing that partial digestion of the edestan had occurred.

In another experiment, a mixture of 4 cc. of pitcher liquor and 1 cc.
of the edestan solution was incubated for 13 days. On neutralization,
the blank yielded a voluminous precipitate, the determination proper,
a precipitate but one-tenth as great as that in the blank. Hence partial
digestion of the edestan had occurred.

Protean derived from castor-bean globulin

The castor-bean globulin, used in these experiments, was presented
by Dr. Isaac F. Harris, to whom I am also indebted for the outline of
its preparation. Ground castor-beans were extracted with gasoline
to remove the oil, then were extracted with a 10 percent, sodium chloride
solution. This solution was filtered, and the clear filtrate was dialyzed.
The globulin, which deposited, was dissolved in a 10 percent, solution
of sodium chloride; the solution was filtered and dialyzed. The globuUn,
which separated, was washed with water, alcohol, and ether, and was
desiccated.

A 2 percent, solution of this globuhn in a 5 percent, solution of sodium
chloride was used as a reagent for proteolytic enzymes; the solution
was filtered, if necessary. When the clear solution of the globulin was
mixed with the pitcher hquor and 0.5 cc. of 0.1 N hydrochloric acid was
added, a cloudy precipitate of the protean derived from the globulin
formed. On incubation, if a proteolytic enzyme, active in the presence
of hydrochloric acid, was present, the insoluble protean was digested
and converted into less complex, soluble compounds, which dissolved;
and the cloud gradually became less dense, and finally disappeared.

The following experiments were made on liquor from stimulated
pitchers.

The pitcher liquor (2.5 cc.) was incubated with 2 cc. of the globulin
solution and 0.5 cc. 0.1 N hydrochloric acid. Proteolysis was marked
on the third day, advanced on the seventh day, and ahnost complete
on the twelfth day.

The experiment was repeated using 0.5 cc. of pitcher liquor, 4 cc. of
the globulin solution, 1 cc. 0.1 .Y hydrochloric acid, and 4.5 cc. of water
(to secure the same concentration of substrate and of acid as in the pre-

OF INSECTIVOROUS PLANTS 449

ceding experiment). The proteolysis was marked on the fourth day, and
was almost complete on the ninth day.

Liquor from non-stimulated pitchers was used in all of the following
experiments.

The liquor (0.6 cc.) was digested with 2 cc. of the globulin solution
0.5 cc. 0.1 N hydrochloric acid, and 1.9 cc. of water. No proteolysis
had occurred at the end of 14 days.

In another experiment, 2.5 cc. of the liquor, 2 cc. of the substrate
solution, and 0.5 cc. 0.1 N hydrochloric acid were mixed and incubated;
the protean was completely digested in 14 hours.

Two experiments were made, using 1 cc. of liquor, 2 cc. of the sub-
strate solution, 0.5 cc. 0.1 .Y hydrochloric acid, and 1 cc. of water.
The protean was almost completely digested in one of these experiments
in 29 hours, and was completely digested in the other experiment in 48
hours.

It should be noted that liquor from a separate pitcher was used in
each experiment in which this protean served as the substrate.

Jacoby^s Ricin

This test was carried out with the reagents prescribed by Jacoby (23).
A solution of 1 gram of ricin (Jacoby) and 1.5 grams of sodium chloride
in 100 cc. of water was prepared, and filtered if necessary. The pitcher
liquor (1 cc.) and the ricin solution (3 cc.) were mixed; 1 cc. of 0.56
percent, hydrochloric acid was added; and the resulting mixture was
incubated. In the presence of a protease active in this concentration
of hydrochloric acid, the cloudy precipitate, which forms on the addition
of the acid, is dissolved during subsequent incubation. The reactions
involved are essentially the same as those described above for the castor-
bean globulin.

Liquor from a stimulated pitcher was used in one experiment. The
cloudy precipitate underwent a marked proteolysis in two days.

Liquor from a non ■stimulated pitcher was used in another experiment.
The cloudy precipitate was partially digested in two days, but had not
been entirely digested at the end of 1 week.

Glycyltryptophane

Liquor (10 cc.) from stimulated pitchers was incubated with 2 cc.
of an aqueous solution of the dipeptide, glycyltryptophane (so-called
Ferment diagnosticum). In this series of experiments only, toluene
was used as a bactericide. After incubation, the test for free trypto-

450 HEPBURX— BIOCHEMICAL STUDIES

phane was made in the usual way with dilute acetic acid and bromine
vapor; the production of a red color showed the presence of free trypto-
phane, and cleavage of the dipeptide.

In the first experiment, the period of digestion was nine days in the
incubator; the test for free tryptophane was negative.

In a second experiment, the period of digestion was 21 days in the
incubator, followed by 7 days in the room. A distinctly positive test
for free tryptophane was obtained.

General Summary

The following conclusions may be drawn from the experiments re-
ported.

The formal titration showed that the Hquor from stimulated pitchers
produced proteolysis of ovalbumen, fibrin, ovomucoid, Nahrstoff-Heyden,
and Witte peptone, while the liquor from non-stimulated pitchers lacked
proteolytic power. This method also showed that, in the presence of
very dilute hydrochloric acid, edestan was digested by the hquor from
stimulated pitchers, but not by that from non-stimulated pitchers.

Carmine fibrin was dissolved by the Hquor from both stimulated and
non-stimulated pitchers, in the presence of 0.2 percent, hydrochloric
acid. This substrate was not dissolved by Hquor from non-stimulated
pitchers in the absence of acid.

In the presence of very dilute hydrochloric acid, the pitcher liquor
produced proteolysis of edestan; digestion proceeded more rapidly in
Hquor from stimulated pitchers than in Hquor from non-stimulated
pitchers.

The protean derived from the globulin of the castor bean was usuaUy
dissolved by the liquor from both stimulated and non-stimulated pitchers,
in the presence of very dilute hydrochloric acid. The same statement
may be made concerning Jacoby's ricin.

The Hquor from stimulated pitchers apparently hydrolyzed glycyl-
tryptophane, provided the period of incubation was sufficiently long.

The liquor from stimidated pitchers possessed proteolytic power in
both the absence and the presence of acid.

The Hquor from non-stimulated pitchers exerted no proteolytic power
in the absence of acid, but possessed such power in the presence of acid.

Further study is required to determine the manner in which stimula-
tion imparted active proteolytic power to the pitcher Hquor. Possibly
stimulation gave rise to a change in the reaction (hydrogen ion concen-
tration) of the Hquor and thereby created a favorable environment for

OF INSECTIVOROUS PLANTS 451

the activity of an enzyme already present; or stimulation may
have produced the activation of a zymogen present in the liquor; or
it may have increased the secretion of protease by the glands of the
pitcher.

In the presence of acid, liquor from stimulated pitchers digested cer-
tain substrates more rapidly than did liquor from non-stimulated pitchers;
this was especially true of edestan.

The proteolytic enzyme of the pitcher liquor undoubtedly plays a
highly important role in the digestion of insects within the pitcher.

Ill

A Bacteriological Study of the Pitcher Liquor of Nepenthes

By
Joseph S. Hepburn, Ph. D. and E. Quintard St. John, M. D.

Since certain investigators (9-10) have attributed the digestive action
of the pitcher Hquor of Nepenthes to the activity of micro-organisms, it
has seemed desirable to study the bacterial content of the pitcher Hquor,
and the proteolytic power of its bacteria.

Description of the Media

Bacterial counts were obtained by sowing the pitcher liquor — un-
diluted, and in several successive dilutions — on plain nutrient agar,
incubating the plates at 37°C., and counting the colonies in the usual
manner.

For the study of the proteolytic activity of the bacteria of the pitcher
liquor, certain protein media were used, following, to a large extent, the
directions of Crabill and Reed (24). The basis of thesfe media was a
stock solution which contained magnesium and ferrous sulphates, dipo-
tassium phosphate, and potassium chloride. For soUd media a stock
agar was prepared by addition of 2 percent, of agar to this solution. The
protein sohd media were obtained by addition of approximately 1 per-
cent, of one of the following proteins to the stock agar: — casein, egg
albumen, carmine fibrin, edestin, ricin (Jacoby), protein (prepared from
aleuronat). After these media had been steriUzed, the proteins were
present as suspended, insoluble particles. Whenever proteolytic bac-
teria were present in the pitcher liquor plated on such media, their
colonies gradually digested and dissolved the suspended particles over
which they grew.

452 HEPBURN AND ST. JOHN— BIOCHEMICAL STUDIES

In a few experiments, plates of plain nutrient gelatin were also sown
with the pitcher liquor, in order to detect the presence of liquefying
(proteolytic) micro-organisms.

To test for the Hberation of tryptophane and the formation of indol,
a hquid medium was prepared, containing 0.4 gram protein (from aleuro-
nat), 20 cc. 0.1 iV sodium hydroxide solution, and 80 cc. of the stock
solution of inorganic salts already mentioned. The resulting suspension
of protein was placed in tubes (10 cc. to a tube), and was sterilized.
The protein gave a purple color with glyoxylic acid and sulphuric acid
(reaction of Hopkins and Cole), and therefore contained a tryptophane
group in its molecule.

The formation of basic compounds (e.g. ammonia) from simple or-
ganic compounds of nitrogen was also studied, using as substrates: —
glycocoU (an amino acid), acetamide (an acid amide), asparagin (which
is both an amino acid and an acid amide), and ammonium lactate (an
ammonium salt of an organic acid). For this study, recourse was had
to solid media, prepared by addition of one of the compounds just named
to the stock agar. One percent, of asparagin was used, and the other
compounds in molecular concentration equal to that of the asparagin.
One-half percent., by volume, of a two percent, solution of rosolic acid
in sixty percent, alcohol was added to each medium as an indicator.
These media were always sterilized by the discontinuous method. The
production of basic compounds by bacteria growing on these media
was indicated by a red color of the medium beneath and surrounding the
colony. Sterile plates of the rosolic acid media were always poured as
controls, to be used in determining the changes in color in the experiments
proper.

In inoculating all of the special media just described, 1 cc. of the
undiluted pitcher liquor was sown into each plate or tube.

Lactose bile-salt broth was used to test for the presence of members
of the colon -aerogenes group of bacteria.

The temperature of incubation was always 37°C., except for the
gelatin plates which were kept at 20° C.

Sources of the Pitcher Liquor Examined

In the majority of the experiments, the liquor was obtained either
from unopened pitchers or from active, open pitchers containing insect
remains. A few experiments were conducted on liquor from pitchers
partly opened and not yet invaded by insects. The liquor from each pitch-
er was studied as a separate experiment.

OF INSECTIVOROUS PLANTS 453

Unopened Pitchers

The prolonged midrib or tendril, which carries the pitcher, was severed
at the end of the basal part of the lamina; and that portion of the tendril,
which interfered with the subsequent manipulation, was removed.
Sterile scissors were always used in cutting the plant tissues. The top
portion of the pitcher was rapidly passed through the flame, and was
then cut off. The cut edge of the pitcher was then rapidly flammed;
and the Uquor was immediately withdrawn by means of a sterile pipette,
and plated in the usual way on plain nutrient agar. The liquor from
12 unopened pitchers was studied in this manner, and was invariably
found to be sterile, as was shown by the absence of colonies on the plates
after incubation for 4 days.

Opened Pitchers

The liquor was removed from opened pitchers with sterile pipettes,
placed in sterile test tubes, and immediately plated.

Partly opened pitchers, free from insects, were used in two experi-
ments. The liquor from each of these pitchers contained a goodly num-
ber of bacteria which grew on plain nutrient agar.

The remaining experiments were conducted on liquor from open,
active pitchers, containing insect remains. The number of bacteria
present in 1 cc. of liquor was determined in each of five pitchers with
the following result: —

Pitcher 1 450,000

2 8,000,000

3 1,200,000

4 1,900,000

5 48,000

The morphology of these bacteria was studied. Smears were made
from several colonies, which differed in physical appearance, and were
stained with Loeffler's alkaline methylene blue. All the micro-organisms
were rod-like, and therefore belonged to the family of the Bacteriaceae.
A few of the organisms contained spores; none of them produced gas
when inoculated into lactose bile-salt bouillon.

The Hquor in an old pitcher, which was becoming brown at the top,
contained 104,000 bacteria per cc.

Gelatin. In two experiments, gelatin plates were poured. The bac-
teria grew and completely liquefied the gelatin in 48 hours.

454 HEPBURN AND ST. JOHN— BIOCHEMICAL STUDIES

Action on Special Media

After the plates of the special media had been inoculated, they were
held in an incubator at 37°C., and were examined at intervals as stated
below, until drying of the media rendered further observation useless.

Casein agar. Eight experiments were made using this medium. In
seven experiments, growth, but no proteolysis, had occurred at the end
of 3 days; digestion of the casein had begun by the fifth day, had become
more marked by the ninth day, and still more marked by the twelfth
day. In the eighth experiment, bacterial colonies failed to develop.

Egg albumen agar. Two series of experiments were made with egg
albumen agar as the substrate. The first series included eight experi-
ments; colonies had appeared in three experiments by the third day
and in a fourth experiment by the ninth day. The plates in the other
experiments remained sterile. Digestion of the albumen was not noted,
even at the end of 12 days.

In a second series, which consisted of seven experiments, growth of
the bacteria occurred during the first five days of incubation, but pro-
teolysis of the albumen had failed to develop at the end of 14 days.
Possibly, in both series, sufficient ovomucoid (a non-coagulable protein)
was present in the dried egg albumen to supply the bacteria with the
necessary carbon and nitrogen.

Carmine fibrin agar. Six experiments were conducted on carmine
fibrin agar. Washed, unswoUen carmine fibrin had been used in the
preparation of the medium, and the flocks were rather large. Colonies
developed by the third day, and showed a marked tendency to grow over
the flocks. Proteolysis had not become apparent on the fifth day. On
the ninth day digestion of the fibrin was distinctly under way.

Edestin agar served as the medium in four experiments. Growth of
the bacteria, and possibly incipient proteolysis of the edestin, occurred
by the third day. No further change was noted on the fifth day. The
digestion of the edestin had advanced somewhat by the ninth day, and
was still more marked on the twelfth day.

Ricin agar was used as the medium in three experiments. Colonies
developed in one experiment by the third day, and in another experi-
ment by the fifth day. Digestion of the ricin had begun in both
experiments by the ninth day, and had become very marked by the
twelfth day. Bacterial growth failed to occur in the third experiment.

Protein agar. Agar, containing protein from aleuronat, served as
the medium in eight experiments. On the third day colonies were pres-
ent in all the experiments, and proteolysis had probably begun in six

OF INSECTIVOROUS PLANTS 455

experiments. On the fifth day distinct evidences were noted of in-
cipient digestion of the protein in all eight experiments; this digestion
was more marked on the ninth day, and still more marked on the twelfth
day.

Each pitcher did not always contain sufficient liquor to permit a
complete set of experiments on all six of the agar media which contained
suspended proteins. However, a general tendency existed that, if the
micro-organisms present in the liquor grew on one of these media, they
grew on all of the media, and usually exerted a proteolytic action on all
the proteins.

A spar agin rosolic acid agar. This medium was used in three series
of experiments. In the first series which included seven experiments,
colonies had developed by the fifth day, and a red (alkaline) color had
been imparted to the medium. By the fourteenth day, the medium had
become yellow in color (acid in reaction).

The second series consisted of eight experiments. Colonies appeared
on all the plates by the third day. The entire medium next became alka-
line in reaction; this change had occurred on from the third to the fifth
day. While the colonies themselves remained alkaHne, the medium
finally became acid in reaction; this change had taken place in over one-
half of the experiments by the ninth day, and in the remaining experi-
ments by the twelfth day.

The third series included seven experiments, in six of which good
growth of the bacteria and an alkaline reaction of the medium were
apparent by the third day. The medium had become acid in reaction
in one of these experiments by the tenth day. In the seventh experi-
ment of this series, growth had not occurred by the third day, but both
colonies and the alkaline reaction of the medium had developed by the
tenth day. The liquor from this series of seven pitchers was also sown
on glycocoll rosolic acid agar, acetamide rosolic acid agar, and anmionium
lactate rosolic acid agar; the results are given below.

Glycocoll rosolic acid agar. In six experiments, good growth of the
bacteria had occurred, and an alkaline reaction had been imparted to
the entire medium at the end of three days; in one of these experiments,
about one -half the total area of the agar had become acid in reaction
by the tenth day. In the seventh experiment, bacterial growth did
not occur.

Acetamide rosolic acid agar. In six experiments, good growth of the
bacteria and an alkaHne reaction of the medium were noted by the third

456 HEPBURN AND ST. JOHN— BIOCHEMICAL STUDIES

day. On the tenth day, the medium was still alkaline in three experi-
ments, and had become distinctly acid in two experiments, while the
change from an alkaline to an acid reaction was almost, but not entirely,
complete in the sixth experiment.

In the seventh experiment, colonies were absent on the third day,
but had developed by the tenth day, and had caused the entire medium
to assume an alkaline reaction.

Ammonium lactate rosolic acid agar. Six experiments were char-
acterized, on the third day, by good growth of the bacteria and an
alkaline reaction of the medium. On the tenth day, the reaction of the
medium had begun to change from alkaline to acid in five of these ex-
periments.

In the seventh experiment, growth of the bacteria had not occurred
by the third day, but had taken place by the tenth day, and the medium
had then become alkaline in reaction.

In the set of seven experiments, which were studied on all four
rosolic acid agars — asparagin, glycocoU, acetamide, and ammonium
lactate — the odor was also recorded. The plates were quite frequently
characterized on the third day by an odor recalling that of ammonia or
amines. This odor was rarely noted on the tenth day.

Protein liquid medium. The bacteria of the pitcher liquor were per-
mitted to act on the suspension of protein (from aleuronat) in the stock
solution of inorganic salts. Two series, of 8 experiments each, were
made. In each experiment, 1 cc. of pitcher liquor was added to a tube
of the medium.

In the first series of experiments, the test for indol was made after
incubation for three days, and again after incubation for ten days.
The test was always negative.

In the second series of experiments, neither indol nor free tryptophane
was present after incubation for twelve days.

Lactose bile-salt bouillon. In two experiments, 1 cc. of pitcher liquor
was sown in lactose bile-salt bouillon; gas developed within 72 hours,
showing the presence of organisms of the colon-aerogenes group in both
pitchers.

The liquor from several pitchers was mixed, and five tubes of the
bouillon were inoculated, using 1 cc. of the mixture, and of its 1:10,
1:100, 1:1,000, and 1:10,000 dilutions, respectively. Gas developed in
all five tubes within 72 hours, hence it may be stated that at least 10,000

OF INSECTIVOROUS PLANTS 457

micro-organisms of the colon-aerogenes group were present in 1 cc. of
the pitcher liquor examined.

General Summary

The following conclusions are based on the bacteriological experi-
ments.

The liquor taken aseptically from unopened pitchers was found to be
sterile.

The liquor in partly opened pitchers, which were free from insects,
contained a goodly number of bacteria.

Liquor from open, active pitchers, containing insect remains, had a
bacterial count of from 48,000 to 8,000,000 per cc. These organisms
were rods (Bacteriaceae). The bacteria in the liquor from such pitchers
liquefied gelatin, and grew on agar in which the sole source of nitrogen
and carbon was either a protein (casein, egg albumen, carmine fibrin,
edestin, Jacoby's ricin, protein from aleuronat), or a simple organic
compound of nitrogen (glycocoll, acetamide, asparagin, ammonium
lactate). The bacteria usually digested the protein in the medium, but
the rate of digestion was exceedingly slow. The bacteria decomposed
the simple organic compounds of nitrogen; an odor recalling that of
ammonia and amines was frequently produced; the medium became al-
kaline in reaction; later on, this reaction changed to acid, but the bac-
terial colonies themselves remained alkaline. These bacteria did not
Hberate tryptophane nor produce indol in their action upon protein (from
aleuronat). The pitcher liquor, on the average, contained at least
10,000 micro-organisms of the colon-aerogenes group per cc.

The following conclusions are supported by the results of our studies
on the protease and the bacteria of the pitcher Hquor.

The slowness, with which bacterial digestion of the protein occurred,
shows that bacteria play but a secondary role in the digestion of the
insects in the pitcher. The leading role in the digestion is played by the
protease of the pitcher liquor.

The bacteria live in symbiosis with the Nepenthes plant, drawing
their nutrition from the digested insects, and assisting somewhat in the
digestion of the insects.

Needless to remark, the tissue enzymes of the insects may produce
autolysis of their tissues, and thereby assist in the digestion.

458

HEPBURN— BIOCHEMICAL STUDIES

1. Voelcker

2. Hooker

3. Tait

4. Von Gorup and
Will

5.

Vines

6.

Vines

7.

Vines

8.

Vines

9.

Dubois

10.

Tischutkin

11.

Goebel

12.

Clautriau

13. Fenner

14.

Robinson

15.

Hempel

16.

Shibata and

Nagai

17.

Pfeffer

18.

Hare, Caspari

and Rusby

19.

Graves and

Kober

20.

Eddy

21.

Gotschlich

22.

Grutzner

23.

Jacoby

24.

Crabill and

Reed

25.

Vines

BIBLIOGRAPHY

Annals and Magazine of Natural History, 1849, (II), IV, 128 136.
Nature, 1874, X, 366-372.

Report of the Forty-fourth Meeting of the British Association
for the Advancement of Science, 1874; Notes and Abstracts of
Miscellaneous Communications to the Sections, 1875, 102-116.
Nature, 1875, XII, 251-252.

Sitzungsberichte der physikalisch-medicinischen Societat zu
Erlangen, 1875-6, VIII, 152-158.

Berichte der deutschen chemischen Gessellschaft zu Berlin,
1876, IX, 673-678.

Journal of the Linnean Society, Botany, 1877, XV, 427-431.
Annals of Botany, 1897, XI, 563-584.
Annals of Botany, 1898, XII, 545-555.
Annals of Botany, 1901, XV, 563-573.

Comptes rendus des seances de 1' Academie des Sciences, 1890,
CXI, 315-317.

Botanisches Centralblatt, 1892, L, 304-305.
Pflanzenbiologische Schilderungen, 1893, II, 186-193.
Memoirs couronnes et autres memoires, publics par 1' Academie
Royale des Sciences, des Lettres et des Beaux Arts de Belgique,
Collection in 8°, 1899-1900, LIX, third memoir, 56 pages.
Flora oder allgemeine botanische Zeitung, 1904, XCIII, 335-
434, (especially pages 358-363).
Torreya, 1908, VIII, 181-194.

Bidrag til Kundskaben om Succulenternes Fysiologi, Copen-
hagen, H. Hagerup, 1916, 147 pp., abstract in Physiological
Abstracts, 1917, II, 146.

Botanical Magazine (Tokyo), 1916, XXX, 149-178.
Landwirtschafthche Jahrbiicher, 1877, VI, 969-998.

National Standard Dispensatory, 3d edition, 1916, pp. 585-586.

Journal of the American Chemical Society, 1914, XXXVI,

751-758.

Dissertation, Faculty of Pure Science, Columbia University,

1909, page 22.

Kolle und Wassermann, Handbuch der pathogenen Mikro-

organismen, 2 Auflage, 1912, I, 102.

Archiv fiir die gesammte Physiologie des Menschen und der

Thiere, 1874, VIII, 452-459.

Biochemische Zeitschrift, 1906, I, 53.

Biochemical Bulletin, 1915, IV, 30-44.

Journal of Anatomy and Physiology, 1876-1877, XI,

124-127.

OF INSECTIVOROUS PLANTS 459

26. Couvreur Comptes rendus des seances de I'Acad^mie des Sciences,

1900, CXXX., 848-849.

27. Tischutkin Berichte der deutschen botanischen Gesellschaft, 1889,

VII, 346-355.

28. Abderhalden and

Teruuchi Zeitschrift fiir physiologische Chemie, 1906, XLIX, 21-25.

IV.

Occurrence of Antiproteases in the Larvae of the Sarcophaga

Associates of Sarracenia flava

By

Joseph Samuel Hepburn and Frank Morton Jones

The parasitic intestinal worms of man and the domestic animals con-
tain antiproteases (antipepsin and antitrypsin), which effectively pre-
vent the digestion of the parasite by the proteolytic enzymes in the
digestive fluids of the host. This is especially true of A scar is (1).

The pitcher liquor of Sarracenia flava contains a proteolytic enzjTne
(2). The larvae of certain species of Sarcophaga {sarraceniae Riley,
Rileyi Aldrich, and Jonesi Aldrich) habitually occur in the pitchers of
Sarracenia flava, where they are constantly bathed in the digestive
liouor of the pitcher. This phenomenon suggested the examination of
Sarcophaga larvae from Sarracenia pitchers for the presence of anti-
proteases. Live larvae obtained from open pitchers of Sarracenia flava
were used in the study. Two series of experiments were made.

In the first series, 16 larvae (total weight L86 grams) were crushed,
and ground with sand and 4.5 cc. of distilled water. The turbid solution
and suspended ti?sue were removed by decantation, and were mixed
with sufficient 95 percent alcohol to render the final concentration of
the alcohol 60 percent. The precipitate which formed was collected
on a filter and dried over calcium chloride in a dessicator. The filtrate
was mixed with alcohol until the concentration of the latter was 85
percent; the precipitate which formed was negligible, although the
antiproteases should have separated at this point.

The thought, that possibly the rather tough larval tissue had not
been ground sufficiently to liberate the antiproteases, led to an examina-
tion of the first precipitate for these antienz>Tnes. When thoroughly
dry, the precipitate was separated from the filter paper, ground inti-
macely with glass powder, and then triturated with 10 cc. of distilled
water. A supernatant liquid was obtained by centrifugation; 2.5 cc.
of this liquid and 2.5 cc. of a 1 percent solution of pepsin in 50 percent
glycerol were mixed; and sufficient hydrochloric acid and trikresol were
added to produce a concentration of 0.2 percent of each of these reagents.
The resulting solution was allowed to stand for 2 hours at room tempera-
ture to permit the pepsin and the antipepsin (if present) to combine. A

OF INSECTIVOROUS PLANTS 461

control experiment was made in which 2.5 cc. of physiological salt solu-
tion were s\ibstituted for the solution derived from the larvae. Carmine
fibrin (0.2 gram, weighed, then swollen in 0.2 percent hydrochloric acid)
was added to both the experiment proper and the control, and both were
then incubated at room temperature. In the control, the carmine
fibrin was completely dissolved in 1.75 hours. In the experiment proper,
the carmine fibrin was not dissolved at the end of 12 days, but had been
completely dissolved at the end of 17 days. Therefore antipepsin,
an antiprotease, was present in the larvae, since the solution derived
from the larvae markedly retarded the peptic digestion.

In the second series of experiments, 82 larvae (total weight 8.30
grams) were used. From the same gathering of larvae a number were
bred to the adult fly, and proved, by examination of the male genitalia,
to be Sarcophaga sarraceniae Riley, the first recognized Sarcophaga asso-
ciate of Sarracenia. The larvae were ground with glass powder to an
intimate mixture, which was thoroughly triturated with distilled water.
The pasty mass was subjected to a pressure of 50 kilograms per square
centimeter in a Buchner press; 48 cc. of press juice were obtained. The
press juice was so cloudy that the edestan and casein tests could not
be applied in the examination for antiproteases, and only carmine fibrin
was used as a substrate.

Antipepsin. In the experiment proper, 12 cc. of press-juice and 12
cc. of a freshly prepared 0.2 percent aqueous solution of pepsin were
mixed and allowed to stand at room temperature for .30 minutes to per-
mit the pepsin and the antipepsin (if present) to combine. Suflficient
hydrochloric acid (2 percent) and trikresol (2 percent aqueous solution)
were then added to make the concentration of each 0.2 percent in the
resulting solution; lastly, 0.2 gram of carmine fibrin was added. A con-
trol experiment vvas carried out in exactly the same manner as the ex-
periment proper, save that 12 cc. of distilled water were substituted
for the press-juice. The temperature of incubation was that of the room.
In the control experiment, the carmine fibrin was completely dissolved
in 45 minutes; in the experiment proper, it was partly dissolved in 14
hours and completely dissolved in 17 hours.

Antitrypsin. In the experiment proper, 12 cc. of press-juice and 12 cc.
of a freshly prepared 0.2 percent aqueous solution of pancreatin (owing
its proteolytic power to trypsin) were mixed, and held at room tem-
perature for 30 minutes to permit the trypsin and the antitrypsin (if
present) to combine. Sufficient 4 percent solution of sodium carbonate
and 2 percent solution of trikresol were added to make 0.4 percent of

462 HEPBURN AND JONES— BIOCHEMICAL STUDIES

the former and 0.2 percent of the latter reagent in the final solution;
then 0.2 gram of carmine fibrin was added. A control experiment was
made exactly like the experiment proper, except that 12 cc. of distilled
water were substituted for the press juice. The incubation was made
at room temperature. In the control experiment, the carmine fibrin
showed signs of incipient digestion in 45 minutes, and had completely-
dissolved in 14 hours. In the experiment proper, the carmine fibrin
was only partly dissolved at the end of 17 hours, but was completely
dissolved at the end of 22 hours.

Since the press juice markedly retarded the digestion of carmine
fibrin by both pepsin and trypsin, both antipepsin and antitrypsin were
present in the larvae of Sarcophaga sarraceniae.

T her mo-stability of the anti proteases. The experiments proper were
also carried out as described above, except that the 12 cc. portion? of
press juice were boiled and cooled to room temperature, then used
without filtration. The protein in the press juice was coagulated by
the heat on boiling. The digestion of carmine fibrin by both pepsin and
trypsin was retarded to about the same extent as when unboiled press-
juice was used. The coagulated protein of the press juice was dissolved
completely by pepsin and by trypsin (pancreatin) only after digestion
at room temperature for 7 to 8 days, the coagulum remaining long after
the carmine fibrin had disappeared. These results indicate that the
antiproteases — antipepsin and antitrypsin — of the larvae were ther-
mostabile. They also indicate that the coagulated protein of the press-
juice either adsorbed antiprotease and thereby resisted digestion, or
else was in itself not readily digestible.

The methods used in the preceding experiments were based on those
described by Fischer (1) and by Wohlgemuth (3) The trikresol served
as a bactericide.

In this study, antiproteases have been found in the larvae of the
Sarcophaga associates of the pitcher plant, Sarracenia flava. The larvae
of other species of Sarcophaga, and of several other dipterous genera,
are likewise able to live and escape digestion in an environment rich in
proteolytic enzymes; probably these larvae also contain antiproteases
which protect them from digestion. Thus Sarcophaga haemorrhoidalis
Fall, can live in the human intestinal tract; Haseman (4) has recently
pubUshed a detailed account of a series of cases of intestinal myiasis
in man, due to the presence of the larvae of this species in the intestines;
Aldrich (5) gives an additional and similar case in which the parasite

ON INSECTIVOROUS PLANTS 463

was also positively identified as S. haemorrhoidalis, and cites other
records where this species may have been the one concerned.

LITERATURE CITED

1. Fischer Physiology of Alimentation, 1st edition, p. 133, New York, 1907.

2. Hepburn Proceedings of the American Philosophical Society, 1918, LVII,

112-129.

3. Wohlgemuth Grundriss der Fermentmethoden, Berlin, 1913.

4. Haseman Entomological News, 1917, XXVHI, 343-346.

5. Aldrich Sarcophaga and Allies in North America, Thomas Say Foundation,

Volume I, 1916.

Index

Antiproteases in Larvae of Sarcophaga 460

Bacteriological Study of Nepenthes Liquor 451

Beach Plum, Observations on the 231

Biochemical Studies of Insectivorous Plants 419

Carroll on Chasmogamic Flowers 144

Carroll on Cleistogamic Flowers 144

Clark, Water Content and Transpiration of Leaves 105

Comparative Histology and Irritability of Sensitive Plants 185

Comparative Morphology of Myricaceae 339

Comparative Study of Floerke a .401
Floerkea, Russell on * 401

Groth on Sweet Potato • 1

Hepburn on Nepenthes 419

Impatiens fulva, Carroll on 144

Jones on Larvae of Sarcophaga 460

Myricaceae, Morphology, Taxonomy etc. of 339

Nepenthes, Hepburn, Jones and St. John on 451

New Cell Formations in Plants 271

Peanut, Waldron on the 301

Pennypacker on the Beach Plum 231

Protease of Nepenthes Liquor 442

Russell on Floerkea 401

Sarracenia flava, Hepburn and Jones on 460

Sensitive Plants, Steckbeck on 185

Steckbeck on Sensitive Plants 185

St. John on Nepenthes Liquor 451

Sweet Potato, Groth on 1

Taylor on New Cell Formations 271

Transpiration of Leaves, Clark on 105

Waldron on the Peanut 301

Water Content of Leaves 105

Youngken on Myricaceae 339

•

\

New York Botanical Garden Library

IIIMIIIIII III 1

3 5185 00293 3404