UNIVERSITY OF CALIFORNIA
AT LOS ANGELES

GIFT OF

CARNEGIE INSTITUTION
OF WASHINGTON

MUTATIONS, VARIATIONS, AND RELATION-
SHIPS OF THE OENOTHERAS

D. T. MACDOUGAL, A. M. VAIL, AND G. H. SHULL

WASHINGTON, D. C.:

Published by the Carnegie Institution of Washington
1907

87 43

MUTATIONS, VARIATIONS, AND RELATION-
SHIPS OF THE OENOTHERAS

D. T. MACDOUGAL, A. M. VAIL, AND G. H. SHULL

WASHINGTON, D. C. :

Published by the Carnegie Institution of Washington
1907

CARNEGIE INSTITUTION OF WASHINGTON PUBLICATION No. 81.
PAPERS OF THE STATION FOR EXPERIMENTAL EVOLUTION, No. 9.

209134

tl

A^t

^iJ

%

MUTATIONS, VARIATIONS. AND RELATIONSHIPS
OF THE OENOTHERAS.

BY D. T. MACDOUGAL, A. M. VAIL, AND G. H. SHUU,.

SCOPE OF INVESTIGATION.

The oenotheras have furnished so much evidence of importance in connec-
tion with saltatory action in heredity that it has been deemed important to
continue the cultural investigation of the group begun in 1902.

Seeds representing the species in cultivation in the principal botanical
gardens of the world have been procured, and these, in addition to a large num-
ber of forms native to eastern North America, have been grown in guarded
cultures.

Attention has been paid to the occurrence of mutants in Oenothera lamarck-
iana with a view to testing the coefficient of mutability and the influence of
environmental conditions on mutation. Extensive sowings have been made
for the purpose of finding derivatives hitherto undetected, with a coefficient of
mutability so small as to have escaped observation. Descriptions of known
mutants have been made independently for the purpose of comparison with
supposedly identical forms in Amsterdam and of facilitating observations
of all kinds upon the subject.

Many important relations between mutants and their parents may be most
advantageously considered by statistical methods, and the studies begun by
one of the authors in 1904 have been continued and extended to include addi-
tional mutants. The height and branching of the stems and the width and
length of the leaves have been again taken into account, but owing to the
great susceptibility of these organs to variation in direct response to environ-
ment, measurements have also been made upon the buds. The lesser varia-
tion of the latter in correlation with vegetative characters makes them much
more satisfactory for the study of hereditary relations, and it is clear that
their statistical study in connection with pedigree-cultures will demonstrate
in several generations the permanence or evanescence of the mutant types
and give decisive answers to such questions as the relation between fluctua-
tion and mutation and the "fixing" of variations through self-fertilizations
or their disappearance through crossings.

2 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

Guarded pedigree-cultures have been made from pure seeds of native species
of evening-primroses which had not previously been brought under close
observation, with a view to procuring additional evidence on mutability in
this group.

The preliminary examination made it apparent that the group comprises a
swarm of clearly separable species, many of which had not been recognized
by the taxonomists and which gave basis for the current opinion as to the
wide variability of the common evening-primrose (0. biennis}.

The study of genetic relationships and of the phenomena of hybridization
in general has necessitated the organization of extensive cultures, which will
need continuance for two or more seasons before definite results may be
announced, and their discussion is reserved for a future paper.

A few cases of fixed hybrids have been encountered, however, in which the
crossing of two species resulted in the production of a progeny consisting of
one fixed type constant in successive generations. Forms arising in this way
are, in reality, new species, and doubtless many such have arisen naturally
and now constitute a part of the native flora.

Striking cases of vegetative mutation, or bud-sports, having arisen, atten-
tion has baen devoted to a study of the inheritance of the saltatory groups of
characters.

A systematic attempt to localize mutatory changes in the life-cycle of the
sporophyte has been made, and has met with marked success in a preliminary
way, since it has been possible to induce new mutants by the use of solutions
of strong osmotic activity and by highly dilute preparations of mineral salts,
some of which are poisonous to plants in high concentrations and stimulative
in low concentrations. Finally, an attempt has been made to bring the facts
disclosed in the descriptive part of the paper into correlation with prevailing
theories as to phylogeny and evolutionary procedure.

The principal results were obtained from cultural investigations in the
experimental grounds and greenhouses of thr New York Botanical Garden
and of the Station for Experimental Evolution at Cold Spring Harbor, Long
Island, and were also extended in the later stages to the Desert Botanical
Laboratory of the Carnegie Institution of Washington at Tucson, Arizona,
thus securing the advantages of a wide range of climate and soils.

A note regarding the various mutants was presented at the weekly botanical
convention at the New York Botanical Garden, November i, 1905, and a
second one, dealing especially with sports or vegetative mutations, before the
Torrey Botanical Club, on November 4, 1905. A lecture dealing with his-
torical aspects of the phases of the subject under investigation was presented
to the Barnard Botanical Club, December 18, 1905, in which the first announce-
ment was made of the induction of mutants by chemically and osmotically
active stimuli. A full description of the technique of pedigree-cultures and

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 3

of the methods employed in the stimulation of ovaries was given in a lecture
at the Woods Hole Marine Biological Laboratory, July 20, 1906, in conjunction
with which derivatives thus obtained of the second generation in Raimannia
and of the first of O. biennis were exhibited. The results of Dr. Shull's
statistical inquiries were presented before the Botanical Society of America at
New Orleans, January 3, 1906, and various notes have been presented before
other societies.

PEDIGREE-CULTURES OF OENOTHERA LAMARCKIANA.

Arrangements were made by which seeds were obtained from four separate
individuals, purely fertilized with their own pollen. This was done by apply-
ing the pollen to the pistil of the same flower or of another flower of the same
plant. All sowings were made in earthenware pans, 30 by 30 cm., filled with
soil sterilized in an autoclave as in previous cultures, and when sufficiently
developed the seedlings were removed, to be transplanted or discarded, as
the conditions of the experiments demanded. The cultures are given under
the key-numbers by which they were designated in the journal of the experi-
mental garden.

A. i. 0.— A packet of seeds which was given the foregoing designation,
which had been harvested in Amsterdam in 1901, and was obtained directly
from Professor Dz Vries. The sowing was made in August, 1904, and the
small rosettes were inspected by Professor De Vries late in September, 1904,
and he kindly assisted in the identification of a few of the mutants included.
Some confusion in the record makes it impossible to give the exact census of
the culture, but it comprised between 500 and 600 seedlings, among which
26 mutant derivatives were identifiable, and, so far as possible, two represent-
atives of each type were transplanted to the experimental garden in May,
1905, coming into bloom about 60 days later.

The authors have not had the opportunity of inspecting living specimens
of all of the mutants which have appeared in Amsterdam and which have
been described by Professor De Vries, and of the 9 forms seen but 5 have
been conclusively recognized. Of these, 0. oblonga was represented by 12
individuals, constituting 46 per cent of the total number of mutants. The
average frequency of this type among the mutative progeny is i per cent, or
20 per cent of the mutants, although in one series De Vries found 176 oblongas
among a total of 334 mutants derived from a culture consisting of over 14,000
individuals. In this instance oblonga. constituted nearly 53 per cent of the
mutants or 1.25 per cent of the entire progeny. (In De Vries, 1905, p. 545,
the proportion of oblonga in cultures is given as 10 per cent, which is a mis-
print for i per cent.)

In the New York culture under discussion, lata was represented by one
individual, nanella by one, albida by two, and gigas by one. Of these, the

4 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

appearance of gigas is the most notable, since it has occurred but few times
in the last twenty years in pedigree-cultures of 0. lamarckiana.

The remaining four forms among the mutants were represented by one,
one, two, and five individuals. Of these, one type with orbicular leaves
presented a most striking departure from the parental form, but failed to
perfect flowers, so that but little information concerning it was gained. The
form in question was reprseented by one individual only, and did not appear
in any other culture

It is to be noted that the seeds used in this culture had been stored for three
years. Doubtless the parental type and the derivatives are characterized
by varying and different powers of endurance, so that the coefficient of fre-
quency of the several forms in question might be expected to be slightly
different from that found in freshly harvested seeds.

C.i. 2. — A lot of purely fertilized seeds derived from one individual, har-
vested in the New York Botanical Garden in 1903, was sown in sterilized soil
in the greenhouse in August, 1904. In September the plantlets were so far
advanced as to admit of recognition of a number of forms, and a number of
duplicates of the parental type were discarded. This process was repeated
at intervals throughout the winter, and the remaining rosettes were trans-
planted to the experimental garden in May, 1905. Seven derivatives were
seen, including 33 individuals. Among these, were 2 of scintillans, 3 of albida,
1 8 of oblonga, and 2, 2, 3, and 3 of four other unknown types. Some of these
unknown types also occurred in other cultures, and 3 of the known and
unknown were not duplicated in other progenies.

The total number of the seedlings included was probably not much above
500, and if this were conclusively confirmed it would furnish an illustration of
a progeny in which the mutants constituted about 6 per cent of the entire
number. Unfortunately some confusion was found in the record. Before
this was discovered, the senior author announced that he had succeeded in
modifying the coefficient of mutability, but, as is to be seen from the above,
the evidence is not conclusive. The large proportion of oblonga is noticeable,
and this form constitutes 54 per cent of the total number of the mutants.

D. i. 7. — A lot of purely fertilized seeds taken from one individual, grown
in the New York Botanical Garden in 1904, was sown in the garden in August,
1904, furnishing a total of 499 plantlets. Development proceeded so rapidly
that it was possible to select and discard 244 duplicates of the parental type
on December 4, 1904. The remaining ones were transferred to small pots.
Forty-three more duplicates were discarded on February 9, 1005, 34 on Feb-
ruary 1 8, 40 on March 3, 6 on March 21, 2 on March 24, 3 on March 27, 7 on
March 30, and i on April 17. On May 15, 19 mutants and 2 of the parental
type were transplanted to the experimental garden. The detail given above
illustrates the general procedure in all such cultures.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 5

As a result of the continued inspection of the culture it was seen to include
3 of nanella, 2 of lata, 6 of albida, and 3 of oblonga, while the remainder could
not be identified. One was purely pollinated to test its relationship with
gigas. The coefficient of mutability in this culture was seen to be barely 3
per cent, and the proportion of oblonga was much lower than in any other
culture.

D. i. 9. — A lot of purely fertilized seeds designated as above, harvested in
the New York Botanical Garden in August, 1904, was sown in two earthen-
ware pans filled with sterilized soil, early in November, 1904; 604 plantlets
were produced ; 390 duplicates of the parental type were discarded on Decem-
ber 4th, 1904, and 66 on December 15th; 77 were transplanted for further
inspection on December 2ist, and the remaining 71 on January i5th, 1905; of
these, 59 were discarded on February 9th, 33 on February i8th, 15 on March 3d,
3 on March 2ist, 5 on March 24th, and 4 on March 3oth; 19 mutant individuals
were transplanted to the experimental garden on May i6th, 1905, of which
i was nanella, 2 were lata, 2 scintillans, i albida, and 2 oblonga, and the
remainder were of types not identified by the authors.

The coefficient of mutability in this instance is seen to be about 3 per cent,
and that of oblonga very small in comparison with the general frequency of
this derivative.

IDENTITY AND DISTRIBUTION OF OENOTHERA LAMARCKIANA.
Since the cultures of the evening-primroses was begun, a few years since,
no opportunity has been neglected to attempt to trace O. lamarckiana to its
original habitat and to establish its relationship to other species of the genus.
Among numerous bibliographical discussions of interest in this connection,
one by Miller (1760) is of great interest. He says regarding the "Tree Prim-
rose with oval spear-shaped indented Leaves, and Flowers proceeding from
the wings of the Leaves on the Upper Part of the Stalk:"

This plant is also a Native of North America; but was the first species of the Genus which
was brought to Europe, so it is more commonly seen in the Gardens than any of the other
species. In some parts of Europe, this is spread about from the Gardens in such plenty,
that it might be supposed a native there. In a small wood near Haarlem in Holland, this
plant covered the ground insomuch that many skilful persons supposed it was a native of
that place. But it may be easily accounted for, because the gardeners who live near that
place are chiefly florists, and they annually change the earth of the beds in their gardens;
so by carrying out of their old earth from their beds, in which many of the seeds were
scattered, the plants came up there; and those being suffered to scatter their seeds, had
filled the whole wood with the plants.

This differs from the first sort (described and figured as 0. biennis) in having broader
leaves; the stalk grows taller, and the flowers are much larger. Both these sorts will thrive
in the Smoak of London better than most plants.

The appended descriptions and the plate (No. 189, dated 1757) very fit-
tingly characterize O. biennis and 0. lamarckiana, and as the descriptions were

6 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

made before O. grandiflora, the only known species which might be confused
with O. lamarckiana, was discovered, the inference may be drawn by exclu-
sion that nothing but O. lamarckiana was referred to, and this inference is
strengthened by the fact that O. lamarckiana was recognized and cultivated
in the Paris gardens a few years later.

Of the cultures that have been carried out from material obtained from
various sources, the following are selected as of unusual interest :

Two sheets of dried material of O. lamarckiana, collected by Mr. E. P. Bick-
nell at Nantucket City, were received in September, 1904. These plants had
been found in the previous month, growing near a cottage, and had spread to
an adjoining waste lot. A visit was made to the place in August, 1905, by
the senior author, but no trace of the plant could be seen, although it was
found again in 1906 by Mr. Bicknell. This journey was inspired by the behav-
ior of the seeds taken from the herbarium specimen and sown in sterilized soil
in November, 1904. No particular attention was paid to the lot, but on
January 27, 1904, 24 plantlets representing the widest diversity observable
were transplanted to small pots in accordance with the usual practice. Six of
these corresponded quite exactly to the mutant 0. albida. Four of these
died before May, 1905, at which time the remainder were transplanted to the
experimental grounds. All of the other individuals developed in accordance
with qualities of 0. lamarckiana, with a maximum amount of color in the buds,
and also a maximum number of basal branches of some length. The pale
mutants, however, did not proceed beyond the rosette stage in 1905.

The owner of the grounds in which the original plants were growing in
Nantucket could not give any information as to the origin of the culture,
except to say that it had been started from seeds a great many years before.

A lot of seeds taken from a trade packet sold by Vilmorin & Company, of
Paris, were sown in sterilized soil on January 3, 1905. The record shows that
192 duplicates of the parental type had been discarded when the 1 1 remaining
ones were transplanted to the experimental garden on May 16. Of these but
3 were mutants, including i lata, i nanella, and i albida. The mutants were
thus seen to constitute but 1.5 per cent of the total culture.

Far the most interesting material and records were obtained from England,
and it was found that O. lamarckiana has been growing in some profusion
in certain localities in that country for several years.

In 1905 Mr. H. Stuart Thompson (Thompson, 1905) called attention to
Oenothera biennis L- and 0. odorata Jacq., growing on sandhills in Lancashire
and Somersetshire, and with the view of comparing the native North American
O. biennis with the plant so called growing in England, Mr. C. Theodore Green
was applied to for seeds of the plant growing in Liverpool district. To him,
therefore, we are indebted for seeds, a photograph of the Lancashire O. biennis,
and the note on the following page (Green, 1902).

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 7

OENOTHERA "BIENNIS" LINN^US.

Form as found on both sides of the River Mersey, from the estuary of the River Dee to
Southport, 1 8 miles north of Liverpool. It is especially luxuriant about Form by Sandhills,
9 miles north of Liverpool, and about Bidston railway junction, 3 miles north of Birken-
head. It flourishes in sandy waste ground, chiefly among the hills of blown sand along the
coast, and extends a few miles inland in scattered groups of plants. It also is found fre-
quently in cultivated ground in cottage and other gardens from the Dee to Southport. I
do not know the date of its first appearance in the Liverpool district, but it was mentioned
in Dickinson's Flora of Liverpool publication, 1851 and 1855.

F. W. Webb saw it near Leasowe (1860), 5 miles southwest of Liverpool, and T. Sansom
at New Brighton (Dick, Flora).

I have known it in various parts of this peninsula of Wirral (between Mersey and Dee)
since 1892.

It has in the last five years extended greatly about Bidston Junction among the sand .
There is much more of it north of Liverpool. In C. C. Babington's Flora it is given as
"American" and its home "Lancashire Sandhills." When newly opened the flowers are
very fragrant, but only last twenty-four hours.

Seeds inclosed, more of which I can send if desired; came from Bidston Junction. This
plant is identical with that of North Liverpool.

At Bidston, the characteristic sandhill plants among which it flourishes are Erysimum
cheiranthiud.es, Brassica monensis, Anthyllis -vulneraria, Psamma arenaria, Equisetum max-
imum, Lychnis alba, Senecio jacobaea, Salix arenaria.

It extends chiefly along the railway lines, which is suggestive as to its origin.

Inclosed enlargement from one-fourth plate shows the plant in situ, with flowers and
fruit stems in end, September, 1905. This year I saw the same type of this plant about the
railway at Llangollen in the northwest.

Mr. H. Stuart Thompson (Thompson, 1906) quotes Watson's New Botan-
ist's Guide (1837) as the first record of O. biennis on the coast of Somerset and
says, "indeed it is quite a feature in the landscape." It was this sentence
that aroused interest in these special plants, as anyone conversant with the
weed-like aspect of the O. biennis of this region would scarcely credit them
with anything like a spectacular appearance.

Hall in 1845 (Hall, 1845, p. 37) speaks of the plant as being extensively
naturalized on the Liverpool sandhills, and James Edward Smith in 1806
(Smith, 1806, p. 1534) figures a plant that was gathered on the extensive and
dreary sandbanks on the coast a few miles north of Liverpool, where millions
of the same species have been observed by Dr. Bostock and Mr. John Shep-
herd, perfectly wild, covering a large tract between the first and second range
of sandhills. Some natural cause has no doubt established it here, though
possibly from the opposite side of the Atlantic. The Smith plate and the
plant figured by Baxter (Baxter, 1839, p. 257) certainly do not represent the
common North American 0. biennis, and the closing note in Baxter, referring
to the sudden expansion of the flowers, applies more readily to such large-
flowered species as 0. grandiflora, 0. lamarckiana, and 0 . argillicola than to the
smaller-flowered ones, like 0. biennis, 0. muricata, and 0. oakesiana.

8 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

A recent writer on the flora of Lancashire records that 0. biennis has been
established there for the last 70 or 80 years and that "whenever the land is
disturbed, or the sand removed to form new roads, this plant is one of the
earliest to grow upon it, and although its conspicuous flowers make it an easy
prey for constant plucking, it survives these depredations and continues to
spread more and more."

"With Mr. Green's photograph and these references in mind, it was with no
little curiosity that the cultures from his seeds were watched. The seed-
lings raised from them early in 1906 proved to be indubitable O. lamarckiana.
In the seed-pan 2 seedlings of O. lata and 4 seedlings of O. rubrinervis were also
recognized among a preponderating number of the 0. lamarckiana. These
plants were observed unto maturity. 0. rubrinervis displayed all the char-
acters ascribed to it by Professor DeVries, with the exception that they were
not quite as large as some that had previously been studied in the New York
Botanical Garden. The two plants of 0. lata, on the contrary, were more
robust than any we had had in New York from Professor De Vries's seed, and
the pollen being more abundant than is usual in the species, efforts were made
to self-pollinate the flowers in the attempt to obtain pure seed. In due time
a few capsules ripened, containing a very small amount of seed, from which
were raised in December, 1906, 8 0. lamarckiana, 10 0. lata, 2 O. oblonga, and
i 0. albida seedlings.

Before these facts were ascertained, the presumption was that the British
0. biennis was a slightly larger-flowered evening-primrose, and in our cul-
tures from foreign seed we usually named it "European biennis," of which
no real equivalent appears to be known here. Whether our native small-
flowered species is also known in Great Britain is not apparent from the man-
uals, as the species there recorded indicates a more ornamental plant than
the one with which we are familiar here, and ours is certainly not "adapted to
the shrubbery. "

The misunderstanding in regard to the identity and provenience of O.
biennis, O. lamarckiana, and O. muricata seems to be widespread. There is
now no doubt as to the identity and original habitat of O. grandiftora Aiton.
The time and manner of its introduction into England are also known. It
escaped from gardens after it was taken into England in 1778, and is now
found growing wild in many places, according to unverified reports.

But as to the advent of lamarckiana we are still at a loss for positive proof.
The plant referred to by Miller, if correctly identified with this form, would
place it in Haarlem in 1757, 21 years before the discovery of grandiftora.
The next we hear of the grandiftora of Lamarck is in the Paris garden in 1797,
which was seen by Seringe not to be the grandiftora of Aiton as known in
England and was renamed after its illustrious discoverer. Next we have the
striking form which is described above as appearing on the coast of Somerset

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 9

in 1837, and at other places since then, forming now quite a feature of the
Liverpool sandhills, which has been shown by our cultures to be the lamarck-
iana around which so much interest clusters. It would seem difficult in the
light of these facts to find any basis upon which the conclusion that this plant
had been derived from 0. grandiflora might rest.

The plant originally introduced into Europe under the name of 0. biennis
seems to have been a large-flowered form altogether different from the plant
under that name as known in America. Then O. lamarckiana grown from
imported seeds or their progeny, and 0. grandiflora from Alabama, present
a very different aspect and have but little resemblance to the 0. biennis of
the waste lands of eastern America.

THE MUTABILITY OF OENOTHERA LAMARCKIANA.

About 2500 plantlets of 0. lamarckiana derived from various sources have
been brought under observation, as described above. Of these 106 or a little
over 4 per cent were mutants. One culture (C.i.2) appeared to give a much
higher proportion of mutants, but a failure to make a record of one lot of
discarded plants makes it impossible to confirm this. The behavior of this
lot of plants and also the confusion of fasciated specimens of 0. lamarckiana,
in the rosette stage, with mutants, led the senior author to assert that the pro-
portion of mutant derivatives in pedigreed cultures of 0. lamarckiana had
been increased from 5 to 6 per cent of the progeny (MacDougal, 1905).
In no culture of later date has it been possible to exceed the maximum propor-
tion of 5 per cent, and the greater majority fall much below this figure. While
the coefficient of mutability has not been modified yet, one or two forms have
appeared which were not recognizable in a rosette stage by Professor De
Vries, and probably constitute an extension of the range of mutability. This,
however, may not be interpreted to mean an effect of local conditions. All
of the mutants now known did not occur in the cultures in Amsterdam
during the first few years of the cultures. Many, in fact, occur so rarely
that by the law of probability it is necessary to grow many thousands of speci-
mens in order to secure one individual of the rare type. In accordance with
this same principle it is to be seen that the spreading of the cultures of 0.
lamarckiana may be expected to bring to light other rarer derivatives. When
such forms are found it is not to be interpreted as a result of local conditions
of the culture in which they appear, but simply in accordance with the coeffi-
cient of mutability of the parental type with respect to the form in question.

The uniformly low frequency of mutants in material procured from seeds-
men is a matter not altogether understood. Among the probabilities to be
taken into account is that the seeds may have lain in storage for more than
one season before being planted, in consequence of which the unequal hardi-
ness of the mutants and the parent might operate to decrease the apparent

10 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

frequency of the derivatives. No proof has been produced to show that the
frequency of mutations might be increased beyond the limit of 5 per cent
found by De Vries, but it is a well-known fact that mutations of all kinds may
be decreased by inadequate nutrition, and it is quite possible that the parental
plants grown to furnish seeds did not have the care and nourishment of pedi-
greed individuals used in the cultures. The appearance of the rare rubrinervis
in the culture might be due to a hybridization with that form, since no evi-
dence has been offered to show that the fertilizations were guarded.

The five cultures described together furnished albida, nanella, oblonga, lata,
gigas, scintillans, and nine other forms not positively identifiable at this time.
Seeds have been preserved, however, and it is hoped we may be able to
present descriptions of at least some of the forms at the close of another year.

Since these cultures were carried through, De Vries has published the
results of his cultures from seeds obtained from seed merchants. 2000 plantlets
from seeds obtained from Haage & Schmidt of Erfurt contained i specimen
of rubrinervis, which arises but rarely in any culture, i of oblonga, and 3 of
nanella.

Two cultures were made from seeds furnished by Vilmorin and grown in
1898-1899. The first gave 14 nanella, 3 lata, 3 scintillans, i albida, i oblonga,
and a few other divergent forms in 3500 seedlings. A second test yielded
3 lata, i nanella, and i rubrinervis ( ?) in 600 seedlings. In both cases the pro-
portion of the mutants was not above 5 per cent, and was even less than that
described above (De Vries, 1905).

In view of the numerous observations described above upon material from
the most widely separated sources, which show mutability, suggestions that
the mutability of Lamarck's evening-primrose has been consequent upon
its segregation in the Amsterdam Botanical Garden may no longer be taken
seriously, although no doubt the well-worn phrase will be duly rehearsed
from time to time by careless critics. 0. lamarckiana may be of hybrid origin,
and what species may not be; but if so this origin is not recent, nor can
sufficient time be found to hunt down all of the unprofitable alternatives that
are offered for consideration.

Exact records have already been made of lamarckiana, rubrinervis, nanella,
and gigas in previous papers (MacDougal, Vail, Shull & Small, 1903 and
1904), and below are the principal taxonomic characters of albida, oblonga,
scintillans, brevistylis, and lata, by which systematic descriptions of 9 of the
15 mutant derivatives iteratively yielded by lamarckiana are afforded from
material grown in New York (see plate i).

NOTE. — The illustrations of O. lamarckiana and O. rubrinervis previously published by
MacDougal (1903) were reduced without proper indication of the scale.

The capsule of O. gigas is incorrectly given in another paper as 2 mm. in length when it
should be 2 cm. (MacDougal, Vail, Shull & Small, 1905.)

Oenothera scintUlans. 1, leaf from rosette; 2, leaf from stem; 3, bract; 4, bud with
bract; 5, flower with petals removed; 6. petal; 7, capsule.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. II

OENOTHERA SCINTILLANS.

O. scintillans was first seen to originate as a mutant in 1888 and has appeared
in this manner in 14 individuals at different times in De Vries's cultures. As
previously described in this paper, it has appeared in 4 specimens in the New
York Botanical Garden during the present year.

O. scintillans is to be noted as an ever-sporting form which in some strains
gives a progeny 60 or 70 per cent of which is composed of lata, oblonga, lamarck-
iana, and nanella, while in other strains but 30 per cent will be included in
these forms. The forms thus derived from scintillans are in no wise different
in their hereditary qualities from the same types derived from other and pure
lineages. Cultures were made from seeds furnished by Professor De Vries
from a strain which was supposed to yield 60 to 70 per cent of the other types
named.

Seventy-eight individuals were obtained from the culture, 15 of which resem-
bled scintillans , 46 lamarckiana, 16 oblonga, and i was present which did not
reach maturity, but suggested some of the incomplete forms mentioned in ' ' Die
Mutationstheorie," probably sublinearis. In the seedlings 3 months old, the
rosettes were 14 to 17 cm. in diameter, dense, and spreading flat on the ground ;
outer leaves, 6 to 7 cm. long, blades about 5 cm. long, 2.5 to 3 cm. wide,
oblong-obovate, widest about the middle, obtuse or acutish at the apex,
tapering gradually to the white-margined petiole, blue-green, shining, puber-
ulent, with occasional small red spots above, glabrous or nearly so, and lighter
green beneath, somewhat brittle. At 6 months of age the rosettes were
dense, the leaves lanceolate, obscurely and minutely denticulate, green and
shining, puberulent on margin and sparingly on upper surface, midvein broad,
petioles variously winged.

The adult plant was short, with few branches which were fairly rigid and
ascending. The lower part of the stem appeared to be terete or nearly so,
while the upper portion was angled and channeled. The stem and branches
were hirsute with spreading hairs.

The stem-leaves were finely pubescent on both surfaces and much crinkled
on the margins and midribs; irregularly toothed, oblong, 7 to 9 cm. long, 30 to
35 mm. wide, abruptly tapering at base, acutish at the apex. The petioles
were short and the entire organ was a deep shining green.

Conic portion of the bud 15 to 20 mm. long, finely pubescent, with short,
spreading hairs. Free tips 5 mm. long, erect. Hypanthium 30 mm. long, and
reflexed calyx-lobes 25 cm. long. Ovary 6 to 7 mm. long.

The petals are thin, 25 mm. long and 30 mm. in width, barely emarginate.
Filaments 15 to 18 mm. long, very slender and strongly upcurved. Anthers
7 to 8 mm. long. Pistils slightly longer than the stamens ; lobes of the stigma
very slender, 4 to 5 mm. long (plate 2).

12 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

OENOTHERA BREVISTYLIS.

A culture was made from seed harvested by De Vries in Amsterdam in 1 899
and said by him to contain about 25 per cent brevistylis, the seed being the
product of the first generation of a cross between 0. lamarckiana and O.
brevistylis. Twenty-one plantlets appeared and were inspected, and some
seemed to form rosettes with much more rounded and abruptly pointed apices
than O. lamarckiana. This was the only distinction that could be made
between the parental type and the derivative, during what would naturally
be the first year of development. In addition one example of 0. nanella as a
mutant from the parental type was included. With the beginning of the
formation of lengthened internodcs in the stem but little difference between
the plants was discernible until flower-buds appeared. At that time the
leaves on the terminal portions of the branches and the bracts seemed rela-
tively much broader, and the flower-buds might be distinguished at some
distance by the fact that they were more cylindrical than the parental type,
and were .abruptly short-pointed. An examination of the structure of the
flower revealed the fact that the style is much shorter than in 0. lamarckiana
and that the stigma generally appears in the throat of the flower and some-
times is not to be seen without tearing the hypanthium apart.

It has been noted by De Vries that the length of the style varies widely
(as much as i cm.), and in this is offered an example confirmatory of results
communicated by the authors in a recent paper (MacDougal, Vail, Shull &
Small, 1905), in which the mutant characters were found to offer a wider
amplitude of variability than the correspondent characters of the parental
type. In addition to the above general characters, it is to be noted that, as a
consequence of the extreme shortness of the style, pollination fails in many
cases and comparatively few capsules are matured.

The evidence offered by this plant is of the greatest interest in connection
with questions concerning the survival of mutants in the habitat of the paren-
tal form, especially in view of the intemperate indulgence which many authors
show in theoretical discussions upon questions of this character.

O. brevistylis was discovered August 20, 1886, by De Vries, as represented
by 2 individuals. Despite the greatest care and the most rigid inspection it
has never been observed to arise in any of the cultures in Amsterdam or New
York, and the supposition is certainly allowable that it is no longer included
among the mutants given off by the parental type. The characteristic qual-
ities of O. brevistylis are recessive when hybridized with 0. lamarckiana, and
consequently the first generation resembles 0. lamarckiana in outward form
but carries O. brevistylis, which reappears in the next or the second generation
of the hybrid, forming on an average one-fourth of the progeny, according to
the simple Mendelian formula. O. brevistylis is a retrogressive departure from

Oenothera bmislylis. \, leaf from mature rosette; 2, bract and capsule; 3, stem-leaf;
4, flower with petals removed; 5, bud; 6, hypanthium split open, showing the short,
imperfect style; 7, petal; & bract; 9, capsule.

2 4

Oenolhera lain. 1, leaf from mature rosette; 2, bract and capsule; 3, bract; 4, bud and bract;
5, flower with petals taken off; 6, petal.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 13

the parental type, matures comparatively few seeds, and is dominated by the
parental characteristics in crossing, and it has survived in its original habitat
and blooming specimens have been observed there for twenty years in com-
petition with the parental form.

The more important anatomical features of O. bremstylis were described
by Pohl (1895), but it seems desirable that its general taxonomic aspect, as
presented in the cultures in New York, should be put on record.

The rosettes are open, leaves ovate-spatulate, crinkled, the upper ones
approximately denticulate, bright, dark green, pubsrulent beneath, less so
above, conspicuous broad veins, petiole winged to near the base, much more
obtuse and rounded at apex than O. lamarckiana.

The adult shoot resembles O. lamarckiana very closely, even in the form
of the stem-leaves. The bracts, however, seem slightly thinner and appear
somewhat broader than those of the parental type.

The conic parts of the buds are 2 to 5 cm. long, appearing more cylindrical
than the parental form, and are covered with very short, spreading hairs.
The buds are distinctly tinged with red, and the free erect tips are unequal
and are not spreading.

The petals are 45 to 50 mm. long and 40 mm. wide, being firm and more or
less deeply emarginate. The filaments are 2 cm. long, the anthers i cm. long
and slender. The pistil does not usually emerge from the tube of the hypan-
thium, the stigma being seen in the very throat of the flower. The stigmatic
lobes are very irregularly developed.

But few capsules are matured. The capsules are 15 to 18 mm. long and
6 to 7 mm. in diameter at the base, being finely pubescent with a few spreading
hairs, bright green and slender, tapering slightly toward the apex.

The hypanthium in this species bears about the same relation to the reflexed
sepals as in the parental form (plate 3).

OENOTHERA LATA.

O. lata was first seen by De Vries in the plantlets produced by seeds har-
vested at Hilversum, Holland, in 1886, and brought into bloom in 1887, and
he has observed the origin of a total number of 493 as mutants in a progeny of
the parental type embracing 130,000 plants.

It has been observed to arise in New York from the third generation of
a strain grown from seeds furnished by Professor De Vries, and also in seeds
obtained from Vilmorin in Paris.

A sowing was made of seeds of O. lata X 0. lamarckiana, furnished by Pro-
fessor De Vries. This form is characterized by atrophied stamens which
produce only a few pollen grains, of fairly normal structure, which seem to
be incapable of producing fertilization; but when the plant is pollinated by
0. lamarckiana, lata characters appear in the first generation as forming from

14 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

4 to 45 per cent of the progeny, or an average of 22 per cent. In the cultures
made in New York, lata, if present, was mistaken for lamarckiana in the seed-
ling stage and was discarded. Among the plants of this progeny which were
finally brought to bloom were albida, oblonga, nanella, and another type not
recognizable by the authors.

Lata is to be regarded as a retrogressive departure from the parental type,
and is most readily recognized by the rounded leaves of the rosettes, and these
organs become very much crinkled in later stages of development. The
leaves are of a deep green, and the flower-buds are much thicker than any
other evening-primrose known, except gigas, perhaps.

The principal taxonomic characters are as follows: The general habit of
this species resembles that of 0. gigas more nearly than any other of the
mutants, being sparingly branched, and having short, stoutish branches on
the upper portion of the stem. All of the branches are ascending, and in
many of the plants the terminal portion of the stem is bent over.

The stem is channeled and angled, brittle, and both stems and branches
are hirsute. The stem-leaves are finely pubescent, 10 to 15 cm. long, 4 to 5
cm. wide, numerous, and with the laminae remotely and shallowly toothed.
The leaves of the rosettes, even in the very early stages, are noticeably rounded
and obtuse at the apex, and those of the lower part of the stem are spatulate-
oblong and vary to ovate-oblong on the upper portions of the stem. Those
on the upper part of the stem are either obtuse or acutish at the apex and
taper to a margined petiole at the base. The leaves, even in young rosettes,
are closely bunched into a head and are so heavily crinkled and thrown into
convexities and concavities between the veins as to be easily distinguishable
from all other forms in all stages of development.

The bracts are large, ovate to ovate-oblong, acute or obtuse at the apex,
cordate or subcordate, and clasping at the base.

The conic portions of the heavy buds are 2 cm. long and i cm. in diameter in
the basal portion, tapering so slightly as to appear nearly cylindrical, and are
finely pubescent with spreading hairs. The free tips of the buds are spreading,
stout, 4 to 5 mm. long. The hypanthium is about 35 mm. long, stout, finely
pubescent and much longer than the reflexed sepals (plate 4 and plate 5, A).

The petals are thick, crinkled, and do not expand fully, being about 35 to 40
mm. long, and 45 to 50 mm. wide. The filaments are 18 to 20 mm. long,
while the anthers are imperfect, being very slender and about 8 mm. in length.
The microscopical examination of the anthers showed a few pollen grains
apparently perfect, yet no fertilization has ever been accomplished with them
in Amsterdam.

In a recent paper R. R. Gates (1907) describes the results of an investigation
of the pollen development of 0. lata in which he ascribes the failure of the

Kosette
Rosette

of Ocn
of Oe

lata, four
albida, fo

onths old, and separate le
months old, and separate

of the same atfe.
es of the same at

HELIOTYPE CO., BOSTON

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 15

pollen to other causes than the ingrowth of the tapetum to fill the loculus.
He found that development may proceed to the formation of tetrads, but
frequently degeneration begins in the resting stage of the first mitosis. The
successful pDllinations recorded below, however, demonstrate the functional
maturity of the pollen in many cases. Heterochromosomes were found to
arise in the prophase after synapsis, both in lata and in the lamarckiana hybrid
with lata. Inferential support was secured for the supposition that the
changes constituting mutation occur during the reduction divisions.

The pistil is barely longer than the stamens, with the stigmatic lobes, heavy
and club-shaped, 6 to 7 mm. long. The capsules are 13 mm. long, about 6 to 7
mm. in diameter in the thickest portion, finely pubescent, many angled, and
tapering slightly toward the apex. The terminal rosettes on the stem and
branches were small and regular.

While this form has hitherto been found incapable of producing mature
pollen capable of effecting fertilization in repeated trials made in Amsterdam
and New York, a notable exception is to be recorded. A package of seeds of
O. lamarckiana, which is growing wild near Birkenhead, England, was received
from Mr. C. Theodore Green in 1905. A sowing made from this lot yielded the
parental form, rubrinervis, and a few individuals of lata, which agreed with
this form as derived directly from pure cultures so far as all external charac-
ters were concerned. Many individuals previously examined in various
laboratories showed pollen which appeared normal under the microscope, yet
no fertilizations could be secured. In the present instance, however, pollen
appeared so plentiful that another attempt was made, the treated pistils
being carefully guarded from pollination from the parental or other forms.
A few pods were matured and some seeds were secured as a result. These
were sown as soon as ripe, in September, 1906. The progeny showed 10 lata,
So lamarckiana, i albida, and 3 oblonga. No material difference between this
and progenies resulting from pollination with lamarckiana was thus shown.
In the latter case the proportion of lata is often as low as 4 per cent of the
entire progeny and has not been recorded to have gone beyond 45 per cent.

.OENOTHERA OBLONGA.

O. oblonga was first observed by De Vries at Amsterdam in 1895, although
it doubtless appeared in his cultures previous to that time, as he suggests. In
all it has arrived in 700 mutants and in the pedigree-cultures described in this
paper in 35 individuals, only a part of which came into bloom in August, 1905,
agreeing with the observations in Amsterdam. Oblonga is one of the most
easily recognizable of the derivatives of 0. lamarckiana, although the distin-
guishing characters do not readily lend themselves to taxonomic description.

The rosettes are not very dense ; the leaves in the young rosettes are nar-
rowly ovate-lanceolate, rather thick and fleshy, with broad midveins which

1 6 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

are sometimes distinctly reddish. The laminae are shining green above, more
or less puberulent on both surfaces, and obscurely denticulate.

The adult plant is less than i meter in height and is very sparingly branched
below, with a few very short branches above. The basal branches are ascend-
ing and do not reach above half the height of the stem. The stems are
strongly channeled and angled.

The stem-leaves are crowded, hang down, and are pubescent on both sur-
faces, 7 to 10 cm, long, oblong-elliptical, remotely and shallowly toothed,
acutish or obtuse at the apex, and irregularly narrowed into a short margined
petiole. The laminae are strongly crinkled, dark green, and become tinged
with crimson with age.

The bracts are oblong or ovate-oblong, acute at the apex, cordate or sub-
cordate at base. The buds are reddish, the conic portion being about 2 cm.
long and finely pubescent with short spreading hairs and a few longer ones.
The free erect tips are about 3 to 5 mm. in length. The petals are thin,
generally crinkled, somewhat emarginate and crenate, and are 3 cm. long and
4 wide. In the unfolding of the flower the petals open only so far as to form
a cup-shaped corolla. The hypanthium is about 3 cm. long, slightly longer
than the calyx-lobes. The slender ovary is about 8 mm. long. The filaments
are 15 mm. long and the anthers 8 mm. The pistils are slightly longer than
the stamens, the stigmatic lobes being about 5 mm. in length.

The capsules are 22 to 26 mm. long and 6 to 7 mm. in diameter at the thick-
est part, finely pubescent, slightly angled and shining green, tapering slightly
to the apex (plate 6).

OENOTHERA ALBIDA.

O. albida was first recognized by De Vries as a mutant in 1888 and was
brought into bloom in 1896. It is distinguishable even in the earlier stages
by the paler color of the leaves, which in the rosettes have upturned margins
and are variously twisted. Seventeen examples appeared in the cultures
previously described during 1905, of which 6 were mutants from O.lamarckiana
as found introduced on Nantucket. The following taxonomic characters
were observed :

Seedling about 3 months old: Rosettes 15 to 17 cm. in diameter, thin,
somewhat raised above the slender rootstock; leaves 7 to 10 cm. long, 2.5 to
3 cm. wide, erect-spreading, the apex recurved and nearly touching the
ground ; blades 5 to 7 cm. long, oblong, broadest at about the middle, acutish
or mostly obtuse at the variously rounded apex, gradually tapering into the
almost translucent, margined petiole, pale light green, very thick and brittle,
margins undulate-denticulate, some of the younger leaves strongly wrinkled
on the margins. About 2 months later the rosettes were irregular, leaves
twisted and various-erected, convexed upwardly, pale yellowish-green, approx-

Oenothera oblonga. 1, leaves from mature rosette; 2, stem-leaf; 3, bracts; 4, petal; 5, capsule;
6, bud; 7, flower with petals removed.

1, leaf of mature rosette; 2, stem-leaf; 3, bract; 4, bud; 5, flower
petals removed; 6, petal; 7, capsule.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 1 7

iraately denticulate, oblong, soft-hairy on both surfaces, petioles winged half-
way to base.

The adult plant was seen to be irregularly branched, the branches often
exceeding the main stem in length, which never reached a height of a meter.
Stem zigzag, rather stout, channeled and angled, branching mostly at base, both
branches and stems very brittle. Branches and stems hirsute, with spreading
hairs and closely pubescent. Internodes short, with the leaves crowded.

Stem-leaves oblong to oblong-lanceolate, more or less regularly denticulate,
mostly obtuse or acutish, nearly sessile, tapering at base, finely pubescent
on both surfaces, very brittle and somewhat crinkled.

The bracts are oblong, acutish, or obtuse, slightly clasping, subcordate,
erected when young, reflexed when mature, margins upturned.

The buds are distinctly reddish, 3 cm. long, finely pubescent, with spreading
hairs and the free tips erect and 6 mm. long.

Hypanthium 4 cm. long, slightly longer than the sepals. Petals thin,
35 to 40 mm. long, and 40 to 50 mm. wide, reflexed from the middle, deeply
emarginate at the apex. Filaments 2 to 3 cm. long, anthers 12 mm. long.
Pistils longer than the stamens, with the stigmatic lobes 5 mm. in length.

Capsules 35 mm. long and 5 to 6 mm. in diameter at thickest portion, finely
pubescent and reddish, tapering slightly to apex. It is noticeable that com-
paratively few capsules are matured by this species. The terminal rosettes
are symmetrical and the buds erect and prominent (plate 5, B, and plate 7).

HYBRIDIZATION OF A MUTANT AND ITS PARENTAL TYPE.

In order to test the results which have been presented by Professor De Vries
with respect to the behavior of mutants in hybridization, a number of flowers
of 0. lamarckiana were castrated early in order to avoid entirely any proba-
bility of self-pollination, and these flowers were pollinated from unopened
buds of 0. rubrinerms. The resulting capsules were harvested early in Sep-
tember, 1904, and sown in sterilized soil in the greenhouses in November, 1904.
Early in January the seedlings were large enough to make the separate types
discernible, and the task of discarding the duplicates was begun. The entire
progeny was found to include 90 of lamarckiana (and one of its mutants, of
which i was present) and 221 rubrinerms. The rubrinerms, which is to be
regarded as a progressive mutant of lamarckiana, has been seen to surpass it in
vegetative vigor and rapidity of growth, besides producing seeds quite as
plentifully, and is sometimes dominant when crossed with the parent. Thus
De Vries found that when the above cross was made he obtained 19, 24, 68,
and 74 per cent of rubrinerms in four different lots, giving an average of
totals of 46 per cent rubrinerms produced in such crosses. The culture
described above yielded 7 1 per cent rubrinervis.

THE FLUCTUATIONS OF OENOTHERA LAMARCKIANA AND
ITS MUTANTS.

BY GEORGE HARRISON SHULL.

The results of statistical studies made in 1904 on Oenothcra lamarckiana Ser.
and two of its mutants, O. nanella and O. rubrinervis, were so suggestive that
it was thought desirable to make a further attempt to trace the relations of
the mutants to their parental form, as indicated by the variations in some
of their differentiating characters. In order to reduce the probable error, it
was necessary to secure data from a larger number of individuals than the
cultures at the New York Botanical Garden could supply. Consequently cul-
tures of several of the oenotheras were begun at the Station for Experimental
Evolution.

The cultures which form the basis of this comparative study were the
following :

(a) About 90 specimens of O. lamarckiana were grown from seed supplied
under the designation D.r.8 by the senior author, being seed from plants
guarded and purely self-fertilized for three generations in the New York
Botanical Garden, and thirteen generations preceding in the Botanical Gar-
den at Amsterdam. In this, as in all the following cultures, no selection of
specimens was made, the desired number of plants having been taken consecu-
tively, beginning at one side of the seed-pan ; and in all cases the arrangement
in the garden was the same as that adopted at the New York Botanical Garden,
i. e., the plants were placed at a uniform distance of i meter from each other.
One of these 90 individuals was an O. lata and was omitted from the statistical
studies on this plot. The remaining 89 specimens were typical 0. lamarckiana.

(6) Ten rosettes of Oenothera lamarckiana were received by mail from Prof.
De Vries April 7, 1905, having been collected by him in the open field near
Hilversum, Holland, where over twenty years ago he secured the original
stock for his pedigree-cultures; 9 of these came to bloom and were uni-
form in appearance, agreeing well with the other cultures of O. lamarckiana
in all characters but those whose differences could be appropriately attributed
to the fact that these plants had grown as biennials, while the others had been
forced to annual bloom by giving an early start in the propagating house.
The chief differences between these plants and those of the culture described
above were seen in the greater height and more numerous and longer branches
of the plants from Hilversum. The rosette leaves were more numerous and
larger in both dimensions, but had the same form and the same degree of
crinkling.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 19

(c) Eighty-eight specimens were grown from seed of an 0. rubrinervis
which had appeared as a newly arisen mutant among the first generation
hybrid progeny of O. lamarckiana X biennis discussed in our earlier account
(Mutants and Hybrids, p. 18). The occurrence of this mutant along with
several individuals of O. lamarckiana was there attributed to an accidental
self-fertilization of the pistil parent, but there is no sufficient reason perhaps
for ruling out the occurrence of monolepsis producing Millardet hybrids in
which the maternal characters alone appear. Whether this specimen of O.
rubrinervis was the product of self-fertilization or of monolepsis has no especial
significance for this study, since, whatever may have been its origin, it was a
true 0. rubrinerms, though having no rubrinervis individuals in its ancestry,
at least for many generations.

The pollination of this specimen was left entirely uncontrolled, and as it
grew in close contiguity with other species of Oenothera, where insects were
busily engaged carrying from flower to flower the pollen of a dozen or more
different species, the progeny might have been expected to show a confusion
of hybrid forms of unknown affinities. It was a matter of some surprise,
therefore, that only i specimen in the 88 appeared to be a hybrid. The char-
acters of this obvious hybrid indicated that it was probably produced by pollen
from one of the small-flowered species, and it was considered by the senior
author to be nearly if not quite identical with his O. lamarckiana X biennis,
No. 2.32, though no definite analysis of its characters was made. The other
87 individuals consisted of 80 O. rubrinerms and 7 which were looked upon as
typical 0. lamarckiana, although, as will be seen later, the results of the statis-
tical studies present some interesting exceptional features.

(d) Twenty- seven individuals of O. rubrinerms were grown from guarded
and self-fertilized seed of one which arose as a mutant at Amsterdam in 1895
in a pure 0. lamarckiana pedigree guarded for six generations.

(<?) Ninety-eight specimens of Oenothera gigas were grown from pure-bred
seed received from Prof. DeVries, the pedigree being known and fully guarded
for five generations from the time it appeared as a mutant in 1895. Unfortu-
nately only 4 of these came to bloom the first year, so that the studies here
given on Oenothera gigas must be considered wholly inadequate.

(/) Another lot of plants, 44 in number, were the offspring of Oenothera
lata which had grown surrounded by O. gigas in an isolated portion of the
Botanical Garden of Amsterdam. On the side of the pistil-parent the pedigree
of this lata had been continued for five generations from the time it arose as
a mutant, and in most of these generations 0. lamarckiana was the pollen par-
ent. As O. lata is nearly or quite incapable of self-fertilization, owing to its
lack of viable pollen, these plants must have been hybrids between 0. lata
and O. gigas, and as they were wholly unselected they formed a motley group.
Among them Oenothera lata was the most frequent, 1 7 in 44 belonging to this

20 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

species; 15 were 0. lamarckiana, and 2 were O. nanella. The remaining 10
were not positively identified, though several which were at first thought to
be O. lata but which were afterwards seen to differ from that species in certain
bud characters, especially in the possession of pollen-bearing anthers, may
have been true hybrids between 0. lata and O. gigas.

The advantages of bringing together to one place for use in this line of
investigation material having such different past histories will be obvious. If
the various types of structure which have been separated and described under
specific names are not perfectly natural centers of stability, or rather of equi-
libration, but are the product of selection, it is inconceivable that the forms
resulting from selection by different persons for different periods of time and in
different environments should produce identical forms, and especially that
there should be identity in characters that were not taken into account and
could not well be taken into account by those making the selections.

Some differences are naturally to be expected in the characters of these
several lots of plants of the same species coming from different sources,
because of the known behavior of purely fluctuating characters, and the degree
of these differences will be contemplated with some interest. On the other
hand, the various sources of the material will make the comparisons between
the different species less satisfactory than might be wished.

A part of the advantage gained in the way of larger quantities of material
by bringing these cultures to the Station for Experimental Evolution is offset
by the fact that differences in soil, climate, and methods of culture from those
to which the cultures were subjected at the New York Botanical Garden in
1904 will make impossible the drawing of any conclusions this year regarding
the problem which originally incited these investigations, namely, the question
as to the presence or absence of regression in the mutants. As a partial
measure of the difference in conditions at the New York Botanical Garden in
1904 and at the Station for Experimental Evolution in 1905, a comparison of
the heights of the 0. lamarckiana used for the studies in the former year may
be made with those of the first lot (a) described on page 1 8. Both of these lots
were raised as annuals and should therefore give a reasonably fair indication
of the relative value of the resultant of all the factors affecting the fluctuating
characters of the two lots of plants.

The height-curves reduced to equal areas are shown superposed in fig. i,
and the principal constants are as follows :

Mean.

Standard
deviation.

Coefficient
of variation.

New York

Station for

Botanical Garden, 1904
Experimental Evolution, 1905.

Centimeters.
88.68 ±0.55
86.98 ±0.69

Cen'imeier 'i '.
4-76±o.39
8-37±o.i9

Per cent.
5-37 ±0.44
0.62 ±0.57

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 21

It will be noted that there has been a slight fall in mean heights, but the
most noteworthy thing in this comparison is the great increase in the varia-
bility. A study of fig. i shows the difference in variability quite clearly in
the greater amplitude and less height of the 1905 curve. The interesting
fact becomes apparent that although the mean value of stem-heights in 1905
was less than that for 1904, some of the 1905 plants were taller than the tallest
in 1904. The explanation of these facts is very simple. The garden at the
Station for Experimental Evolution has been under cultivation so short a

66-70 71-75 76-80 8I~85 86-9O 91-95 96-100 101-105 106-110

FIG. 1 . — Variation in height of stem of Oenothera lamarckiana. Broken curve represents
the condition at the New York Botanical Garden in 1904, continuous curve that at the
Station for Experimental Evolution in 1905.

time that it still lacks much of that homogeneity that is desirable for work of
this nature. One corner of this plot of Oenothera lamarckiana dipped into a
slight depression which received the drainage from the roof of the laboratory
and was therefore kept, during a portion of the growing season, too wet for
the best development of the plants. A few specimens on this corner did not
make flower-stems at all, and it was to be expected that those nearest this
corner of the plot would be somewhat influenced by the same unfavorable
conditions and made to produce flower-stems of a slightly less altitude.

The stature of Oenothera rubrinerms. — In 1904 a comparison was made
between the statures of Oenothera nanella and O. lamarckiana, and it was a

22 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

part of the plan for the present study that this comparison should be contin-
ued, but unfortunately the seed of 0. nanella gave only 3 germinations and
the continuation of this comparison must be postponed until another season,
but meanwhile it will be particularly interesting to make another comparison
of heights which will present something of a contrast with that between O.
nanella and its parent. It will be recalled that in that comparison the co-
existence of two striking features was taken to be more than a mere coinci-
dence, namely, the great departure of the mean value in O. nanella from the

30 1

66-70 71-75 76-80 81-85 86-90 91-95 96-100 101-105 106-110 111-115 116-120

FIG. 2. — Comparison of the variability of stem-heights of Oenothera lamarckiana (broken curve)
and O. rubrinervis, as grown at the Station for Experimental Evolution in 1905.

mean of its parent, and a very great increase in the coefficient of variability
in the heights of the mutant. In comparing the heights of O. rubrinerms with
those of O. lamarckiana there is seen a very much less marked departure from
the parental condition in the mean values, and a correspondingly slight
increase in the coefficient of variability. This comparison is shown graphic-
ally in fig. 2 and is expressed numerically as follows:

Variation in heights, 1905.

Mean.

Standard
deviation.

Coefficient
of variation.

Oenothera lamarckiana
Oenothera rubrinervis

Centimeters.
86.98 ±0.69
92. 18 ±0.99

Centimeters.
8.37 ±0.49
9.86 ±0.70

Per cent.
9.62 ±0.57
10.70 ±0.78

This may still be merely a coincidence, but two such coincidences are very
much less likely to occur than one, and the attempt will be made in the study
of other characters to trace the suggested parallel between the departure of
the mean from the parental condition and the increase in the value of the
coefficient of variability.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 23

The branching habit of Oenothera rubrinervis. — In discussing the branching
of Oenothera rubrinervis in 1904 the statement was made that "such a char-
acter as this is too largely influenced by the individual physiological vigor to
be of any value as a diagnostic character, " and this statement receives striking
confirmation in the conditions observed in 1905. The specimens of 0. rubri-
nervis grown at the Station for Experimental Evolution in the latter year were
totally different in their branching from those studied the year before at the

16-18 19-21 82-24 35-27 28-30 31-33 34-36 37-39 40-42 43-45 46-48

FIG. 3. — Comparison of the number of branches of Oenothera lamarckiana (shaded with lines rising
toward the right) and O. rubrinervis, as grown at the Station for Experimental Evolution in 1905.

New York Botanical Garden. There was in 1905 almost a complete sup-
pression of the large basal branches which were such a marked feature of the
specimens earlier described. Instead of finding Oenothera lamarckiana and
0. rubrinervis entirely distinct and separated by a wide gap in respect to
number of lateral branches, the curves in figure 3 show how greatly they
overlap. Their numerical relations may be studied in the following table :

Range.

Mean.

Standard
deviation.

Coefficient of
variation.

Number of the lateral

branches :

Per cent.

Oenothera lamarckiana . .

1 8 to 39

26.20 ± 0.33

3-99 ± 0.33

15.23 ± o 91

Oenothera rubrinervis . . .

19 to 47

30 18 ± 0.72

6.12 ± 0.51

20.29 ±1.76

Total branch length in

decimeters :

Oenothera lamarckiana . .

40 to 150

91.40 ± i 49

17.97 ± 1.05

19.66 i i .20

Oenothera rubrinervis . . .

41 to 174

87.22 ± 2.36

23.46 ± 1.67

26.89 ± 2-°5

24 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OEXOTHERAS.

With respect to the number of branches, the figures in this table do not
fully represent the tendency in 0. rubrinervis to develop a large proportion of
its axillary buds into branches, as many of these buds had started into growth,
but for some reason had been arrested before they had attained a length of
i cm., under which length none were counted. Comparing these results with
those of 1904 it is found that with respect to variability Oenothera lamarck-
iana presents essentially the same condition in the two years, though the
values of both number and length of branches has increased in the latter year.
In Oenothera rubrinervis, on the other hand, although there has been in 1905
a great reduction both in number and length of branches and the coefficients
of variability are considerably changed, they are still much higher than in the
parent species. This will be best appreciated by a comparison of the results
for the two years as shown in the following table :

Variation in —

Years. Range.

Mean.

Standard
deviation.

Coefficient
of variation.

1

Per cent.

Number of branches of

1904 10 to 27

20.27 ± 0.49

3.24 ± 0.35

15-70 ± I./O

O. lamarckiana.

1905 i8to 39

26.20 ± 0.33

3 -99 ±0.33

15-23 ± 0.91

Number of branches of 1904 34 to 63

42.35 ± J -04

6.34 ± 0.73

15.00 ± i .70

O. rubrinervis. 1905 19 to 47

30. 18 ± 0.72

6. 12 ± 0.51

20.29 ± i .76

Branch-length of O.

1904 36 to 85

66.8 ± 2.0

13-5 ± i-4

20.2 ±2.2

lamarckiana. 1\ 1905 4010150

9i-4 ± i-5

18.0 ±1.1

20 . 0 ±1.2

Branch-length of O. , | 1904 78 to 300

181.9 ± 13-0

79-5 ±92

43-7 ±5-i

rubrinervis. | 1905 4110174

87.2 ± 2.4

23-5 ±i-7

26.9 ± 2.0

The greater variability of the mutants was taken in 1904 to indicate that
these forms might not yet have acquired, perhaps, that fixity of habit exhib-
ited by an older species, and the fact that Oenothera rubrinerms underwent
greater change in its mode of branching on being cultivated in a new environ-
ment than was exhibited by O. lamarckiana when subjected to the same
change would accord well with this suggestion.

The increase in the number and length of the branches in 0. lamarckiana
and the considerable decrease in those of O. rubrinervis must not be inter-
preted, however, as in any sense an approach to actual identity in these two
species, in regard to branching habit, for in the latter year they were even
more distinct in general type of branching than before. Oenothera rubri-
nervis had simply lost its long basal branches, while 0. lamarckiana had retained
them.

THE SIZE AND SHAPE OF LEAVES OF OENOTHERA RUBRINERVIS.

Some of the difficulties attending the statistical study of leaves were pointed
out in our earlier paper, and the complete suppression of the basal branches of
O. rubrinervis in 1905 has made the continuation of the comparison between

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 25

the leaves of that species and those of its
parent of little value except as a demonstra-
tion of the difficulties which were to be
overcome. It seemed fair in the former year
to consider the basal branches of these two
species homologous, and leaves for the study
in 1904, being taken from about the middle
of those branches, were consequently as
nearly comparable as possible. In 1905 the
leaves of 0. lamarckiana were again taken
from the middle of the basal branches and
are thus comparable with those of 1904,
but in O. rubriner-vis the leaves were taken
from the middle of the lowest branches pres-
ent, and as these were not quite basal, the
leaves were not homologous with those of
0. lamarckiana. The first striking result
of this lack of homology is seen in fig. 4,
where the variation in leaf-lengths is com-
pared. The corresponding curves for 1904
showed almost exact identity in the length
of the leaves, while in this figure it is seen
that the length of most of the 0. rubri-
nervis leaves corresponded in 1905 with the
lower part of the range of O. lamarckiana.
When figure 5, representing the variation in
leaf-width, is compared with the corres-
ponding figure for 1904 it is seen that with
respect to this character the lack of homol-
ogy has had but little modifying effect,
and there is close resemblance between the
sets of curves representing the variation in
leaf-width for the two years, though both
species show a slight increase in the coef-
ficient of variability. The considerable
difference in respect to leaf-length, taken
in conjunction with essential identity in
leaf- width in 1905, has the effect to change
the ratios between length and width in
about the same degree as the change in
length, but in the opposite direction. Thus
the change in length was away from identity,

FIG. 4. — Comparison of leaf-length of Oeno-
thera lamarckiana (broken curve) and O.
rubrinerms.

26 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

while that in the leaf
ratios was toward iden-
tity, as will be readily
seen on referring to fig.
6, which shows a much
more complete over-
lapping than do the
corresponding curves
for 1904. In other
words, since this ratio
was taken as a measure
of leaf-form, the form
of the leaves of O.
rubrinerms in the latter
year was but slightly
different from that of
the parent species, ex-
cept in those minor
details which were
pointed out in the
earlier study as pre-
venting the ratio be-
tween length and
width from being a
correct measure of
form. The lack of
homologyinthe leaves
measured in the two
species takes away all
20 25 30 35 40 45 50 significance of these

FIG. 5.— Comparison of leaf -width of Oenothera lamarckiana (broken curve) facts Ul their bearing
and O. rubrinervis. on the question of the

regression of 0. rubrinerms. The constants of the several curves representing
the variation in leaf measurements may be compared in the following table :

Mean.

Standard
deviation.

Coefficient
of variation.

Variation in length of leaf:
Oenothera lamarckiana . ....

Millimeters.
88 91 ±0.35

Millimeters.
18.41 ±0.25

Per cent.
20.71 ±0.29

Oenothera rubrinervis

70 96 ±o. 24

1 1 .42 ±o. 17

16. io±o.24

Variation in width of leaf:
Oenothera lamarckiana . . .

36 56 ±0.091

4. 796 ±0.064

13. I2±O. l8

Oenothera rubrinervis

27 30 ±0.079

3 • 687 ±0.056

1 3 . 50 ± o . 2 1

Variation in the ratio between width and
length :
Oenothera lamarckiana ....

Per cent.
42 27 ±o. 10

Per cent.
5.411 ±o 072

i2.8o±o. 17

Oenothera rubrinervis . .

39 34 ±o 13

5.888 ±o 089

14. 97 ±o. 23

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 27

The original measurements upon which the above figures and conclusions
are based are presented in the form of correlation tables in figs. 7 and 8. A
comparison of the tables with the corresponding figures for 1904 shows that
those representing 0. lamarckiana are very much alike in the two years, the
degree of correlation being even higher in the latter year than before, due in
part possibly to the slightly greater heterogeneity already pointed out in the
cultures at the Station for Experimental Evolution. On the other hand, in
Oenothera rubrinerms the corresponding tables for the two years are entirely

10

A

65

20 £5 30 35 AO 45 50 55 60

FIG. 6. — Comparison of the ratio between leaf-length and leaf-width in Oenothera
lamarckiana (broken curve) and O. rubrinenis.

different, as would be expected from what has already been said. In this
species the coefficient of correlation has been considerably lowered.

The coefficients of correlation for the two species for the two years may
be compared in the following table:

1904.

1905.

Oenothera lamarckiana

0.7916 ± 0.0090

0.83-^0 ± o.oo<)8

Oenothera rubrinervis

0.6604 ± 0.0119

0.4452 ± 0.0171

28 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

Variation in bud characters .—It is a well-known fact that floral characters
fluctuate less in response to change of environment than vegetative characters,
though I know of no exact studies showing to what extent this is true. The
comparatively larger flowers of alpine regions find an explanation on this
basis since the alpine plants have generally migrated from the lowlands.
I have seen a specimen of Lobelia syphiliiica less than 8 cm. tall and dwarfed

X,

-11

-s

-7

-u

--'

-1

-;;

-2

-1

0

1

2

3

1

•-'

C,

7

s

0

10

11

12

18

15

50

63

lid

w

70

75

SO

S3

90

1C,

Km

11 '5

11(1

d

].;„

l'J5

130

lid

115

15: i

!55

Xi

-1-13

21

1

1

••12J24

1

1

1

3

-•11

25

3

HF

- 10

21 i

2

2

3

1

8

-9

27

1

2

1

3

7

-8

28

1

1C

10

7 |3

31

-7

2'j

t

7

7

9

1

1

31

-li

3(1

2

1

S

10

9

9

1

13

-5

31

1

7

18

u

5

3

3

3

11

-4

32

5

12

11

27

0

S

(i

8

SI

-3

3'!

2

10

12

21

18

ID

12

(i

1

2

98

-2

31

o

12

20

22

I!)

10

i;

1

5

1

103

-1

39

1

1

3

u;

10I27J18

18

;o

1

2

108

0

86

(i

3

5

8

14

10

IS

si

5

1

79

1

37

(i

11

7

14

21

20

!G

S

6

2

103

2

38

2

8

6

9

12

2(i

11

21

4

4

6

103

3

311

2

8

1

3

(i

S

1C,

18

10

3

2

75

4

40

1

2

2

1

5

17

•5

:r>

•4

5

3

85

5

!1

2

2

1

1

7

s

12

S

3

53

6

42

1

1

1

3

lo

H

a

7

11

2

1

CO

7

13

1

2

1

i

7

4

11

2

1

1

i::

8

;i

i

3

9

9

1

27

9

15

2

1

10

5

1

2

29

10

iij

1

l

3

2

1

Hi

11

17

1

2

3

2

1

K

12

IS

1

1

1

1

13

III

3

1

1

5

14

5!J

2

1

2

1

7

15

51

1

2

1

•:

!

•.a

',:,

in

111

HI

121

110

US

Id:;

,S1

(i:5

5:5

::5

i

270

"|

FIG. 7. — Correlation between width (subject) and length of leaf inOenotheratamarckiana.
Measurements in millimeters. Coefficient of correlation, 0 . 8330 ± 0 . 0058.

specimens of numerous other species bearing flowers almost as large as is
normal for the species.

Having sufficiently demonstrated the unsatisfactory character of height,
branching, and leaf -form for statistical comparisons because of their great sen-
sitiveness to even slight changes in the environment, I have decided that in the
buds is to be found probably the most satisfactory material for a continuation

Hft

Buds of Oenothera lamarckiana and four of its mutants, showing- characteristic differences and
variability, ist row at top, Oenothera lamarckiana; 2nd row, Oenothera rubrinervis; yA
row, Oenothera lata; <|th row, left, Oenothera gigas, right, Oenothera nanella.

HEUOTYPE CO., BOSTON

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 2Q

of these studies. The oenotheraceous bud offers distinctive characters fcr
the separation of the several species (plate 8). It has, as is well known, three
distinct regions, the ovary, the hypanthium, and the cone — which are in most
cases marked off from each other in such a way as to make their accurate
measurement easy. As proof of this fact it is only necessary to say that fre-
quent repetitions
of the measure-
ments showed only
a variation of one-
tenth of a milli-
meter in the thick-
ness of ovary and
hypanthium, and
from o. 2 too. 3 mm.
in thickness of the
cone and lengths of
the several regions,
the maximum error
in the smallest
measurements
taken being less
than 4 per cent of
the magnitude
measured, and in
the larger measure-
ments, such as
length of hypanthi-
um and cone, less
than i per cent.

In deciding to
make statistical
studies on buds a
careful preliminary
survey was made
to determine what
precautions would
need to be taken

in order to free the work from error. The several points that seemed to need
consideration were: (a) A probable periodicity in the size of the buds, so that
those in the early part of the season are doubtless a little larger than those
producednear the end of the season, but as the flowering season is very long this
periodicity is certainly slight and would probably have no appreciable influence

X.,

- 1

-3

-2

-1

0

1

2

3

1

.-,

(i

7

s

!)

10

11)12

13

11

15

a

:,o

55

(iO

C,5

70

75

80

89

'.Hi

95

UK) 105

110

115

120J125130

I:L->

110

30

.-,.-,

60; 05

70

75

80 85

M

95

100

105

110

115

r:o

1251

,.".,

i:!5

,10

H5

Xi

-•11

1C,

1

1

-10

17

0

- ii

IS

1

1

1

1

1

-8

111

1

1

1

1

1

1

6

-7

20

1

1

1

1

1

2

1G

-G

21

3)7

U

2

2

23

-5

22

2

3|l3

12

7

3

1

43

-4

2.;

r, i!

Hi

U

5

4

1

55

-3

-i

1 | 5

12

12

13

3

1

2 2

59

-2

2o

•:

23

•:o

33

17

c,

I

I

1

118

-1

„'!,

i

10

2)

111

15

s

3

95

0

27

1 4

1)

21

31

IS

7

li

1

2

2

119

1

28

8

IS

211

32

10

(i

1

.-,

1

120

2

21)

1

10

Hi

22

5

3

•2

3

G7

3

30

1

i

2

U

is

13

7

U

4

1

2

I

2

19

4 J31

1

I

7

1!

ll

G

3 1 1

1

1

1

59

32

3

I

(i

12

3

1

2

3

39

3",

1

1

1

7

7

13

7

3

1

1

2

i

1

1C

31

2

1

3

4

4

1

IS

33

2

!

3

2

1

1

3

13

3:i

2

1

2

1

6

10

37

1

1

1

1

1

11

38

1

1

2

12

3!)

2

2

13

10

0

14

-11

0

15

12

0

16

13

0

17

41

0

18

!.-,

0

Hi

Hi

0

20

17

I

1

7

:ii

10.

\:->

;i

l.r;

S3

17

1.M

17

I

995

FIG. 8. — Correlation between width (subject)and length of leaf in Oenothera
rubrinervis. Measurements in millimeters. Coefficient of correlation,
0.4452±0.0171.

3O MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

during the greater part of the season if the earliest and latest flowers are
omitted. (6) A variation in the size of the buds in response to different con-
ditions of the weather. While this variation is not often great enough to be
noticeable, there were times at which the buds seemed distinctly larger after
several days of warm, humid weather. Both of these difficulties would be
rendered comparatively harmless by a collection of a small number of buds
each day from each of the lots of plants studied, so that all were being subjected
to the same conditions. There are still some elements of error, however,
which this method does not fully meet, such, for instance, as the existence of a
different period in the different species, or a different kind or degree of reaction
to changes in the weather. Without an extensive study of the periodicity and

response in the several
species to changed cli-
matic conditions it can
not be known just how
serious errors these
factors introduce, but
it is believed that they
are slight, (c) There
is always present a
complete series of
buds, beginning with
those ready to bloom
and ending with those
just beginning to de-
velop, and the develop-

1 1

X,

-nUo

-9

-g

-7

-a

-5

-1

-.",

-2

-1

0

1

2

3

4

•>

<i

7

•22

23

21

2o

2<i

27

2-

2y

30

31

32

:;:;

31

K

31;

37

38

yj

10

x,

c:

1

1

-5 65

70

0

-4 70

70

1

1

1

1

1

5

-3 75

.SO

1

2

1

1

5

-•2 St.

;;j

1

1

B

2

1

1

2

1

i

1

2

1

i

25

-1 85

<M

1

1

1

3 | 1 | 4 | 1

2

7] |_3j

i

23 i

0 90

•J.J

i

1

•1

1 ' 4 ! 2 ! 1

in 3 : 4 i 2] 2

2

30;

1 95

101

!

1

3

2

2

5

1

C

4

4

55

1 ! 1

4£

2 100

no

1

1

1

4

3

2

3

2

17

3 KS

11C

2

1

1

1

2

10

4 110

11.-.

1

1

5lll5

1-.I

1

1

•1

1

i

•2

i

i

'J

r,

a

1 1

12

5

Is

20

If

13

17

8

3

60

FIG. 9.— Correlation in the buds of OenotheralamarckianaiTomHilvei- 1S S°

sum. Length of ovary subject, length of hypanthium relative. buds Collected at One
Coefficient of correlation, 0.3872 ± 0.0453. hour Qf ^ day ^

not the equivalent of those collected at any other hour. It was found on
examination that the buds which would bloom the following day could be
readily distinguished by their size and color. Rarely there would be a doubt
about some particular bud, and such were omitted from the study. To
avoid the collection of buds at different stages of development they were
collected from all the lots at 5 p. m. and measured as soon thereafter as pos-
sible. This makes all the buds used in these studies as nearly comparable as
could well be secured, though it is realized that a source of slight error is to be
found in the fact that on some days the development of the buds is more
rapid than on others, so that they are not always quite equivalent at the time
of collection, when this is determined by the clock without regard to the
weather.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 31

This work was not begun until late in the season, the first collection of buds
for this study having been made on September 15 and the last on October 30,
and this will need to be taken into account in making future comparisons,
though it seems likely that the periodicity is too slight to make this a modify-
ing factor of any considerable moment.

The number of buds from each lot of specimens, except one, was from 140
to 161. In Oenothera gigas only 91 buds were secured before frost cut short
the investigation. The lateness of the time at which the work was begun
prevented the measurement of a larger number of buds of each lot, though it is
realized that a much larger number would be desirable. On this account it
is not thought necessary to present variation curves of each lot of measure-
ments. Instead, the constants have been calculated and are presented in the
form of a table :

Length of ovary.

Length of hypanthium.

Mean.

Standard
deviation.

Coefficient
of variation.

Mean.

Standard
deviation.

Coefficient
of variation.

Deci-millimeters.

Deci-millimeters Per cent.

Millimeters.

Millimeters.

Per cent.

a

95 -67 ±o 50

9-33 ±0-35 ; 9-75±o.39

33-o6 ±0.23

4.20 ±0.16

i2.7i±o.49

6

91.76 ±0.47

^•73 ±0-33 9-52±o.36 33.07 ±0.22

4.18 ±0.16 i2.63±o.48

c'*

92.29 ±0.34

6 40 ±0.24 6 94±o.26 29.10 ±0.22

4.14 ±0.16 i4-23±o.55

e

107.56 ±0.53

9.89 ±0.37 g.20±o.35' 30.37 ±0.21

3.88 ±0.15 i2.77±o.49

d

118.81 ±0.47

8.40 ±0.33 7.o7±o.28, 30.33 ±0.18

3.24 ±0.13 !io.69±o.43

t

IO2.I2 ±0.64

9.01 ±0.45 8.83±o.44 25.18 ±0.37

5.29 ±0.26 2i.oo±i.O9

/

83.42 ±0.44

7-77 ±0.31 9.3i±o.38| 26.96 ±0.32

5.61 ±0.23 2o.8o±o.87

1

Length of cone.

Thickness of ovary

Millimeters.

Millimeters.

Deci-millimeters. \Deci-millimeters

a

39-47 ±0.19

3-47 ±0.13

8.79±o.34

32.37 ±0.10 i.932±o.073

5- 97 ±0.23

b

39-47 ±0.18

3.45 ±0.13

8.74±o.33 33.51 ±0.11

2.O78±o.o78

6.20±0.24

c'

37.80 ±0.14

2.65 ±0.10

7.oi±o.26 3i.26i±o.O9i

1.711 ±0.064

5-47±o.2i

c

35-52 ±o 17

3-29 ±0.15

9-26±o.35 33-76 ±0.14

2. 540 ±0.096

7-52±o.28

d

36.35 ±0.13

2.277±o.o9o

6.26±o.25 36.48 ±O.TI

2.04i±o.o8i

5-59±o 22

e

39.51 ±0.36

5.29 ±0.26

i2.75±o.6s 44.46 ±0.31

3.69 ±0.22

8.3o±o.49

t

30.27 ±0.24

4-27 ±0.17

i4-09±o.58

38.24 ±0.11

i.938±o.Q78

5.07±o 20

Thickness of hypanthium.

Thickness of cone.

Deci-millimelers. Deci-millimeters

Deci-millimeters. \Deci-millimelers

a

25.29i±o.o68 i.26i±o.048 4-99±o 19 84.59 ±o 39 7.21 ±0.27

8-53±o.33

b

26.o69±o.o7oji .32o±o.o5o 5.o6±o.i9 81.81 ±0.32 ,6.00 ±0.23

7-33±o.28

c'

25-453±o. 0741. 389±o. 052! 5.46±o.2i 87.60 ±0.35 6.63 ±0.25

7-56±o.29

c

28.46 ±0.10 ji.940±o.073l 6.82±o.26 93.26 ±o 43 8.01 ±0.30

8-59±o.33

d

30.356±o.o7o i.259±o 050 4.i5±o!i6 93.47 ±0.29 5.17 ±0.20

5-53±o.2i

e

36.45 ±0.24 3.46 ±0.17

9.48±o.48io9.76 ±0.69 9.79 ±0.49

8.92±o.45

t

29- I57±o.o8i i.422±o.057

4.88±o 20

96.41 ±0.47 8.18 ±0.33

8.48±o.34

*d = Oenothera lamarckiana from lot

32 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

One of the things of first interest that will be looked for in this table is the
relation of the fluctuation found in different lots of plants of the same species.
Particularly is it interesting to compare the conditions of entry a with entry
b, the first representing Oenothera lamarckiana self-fertilized for sixteen genera-
tions, the second the same species which has never been self-fertilized, except
as that process may have taken place by natural causes.

One represents 16 generations of artificial selection, the other many genera-
tions of natural selection, but in neither case is it probable that the propor-
tionality in the parts of the bud entered into the selection in any specific way.
In the mean values of the several parts, these two lots of plants are found to be

XJ-13

-itr-ii

-It

-9

-8

-7

-<;

-.-,

— !

-3

-2

0

1

2

3

1

:,

G

7

B

•.)

10

20

21

22

21

21

2o

20

27

28

21)

:;u

:;i

32

:;:;

M

;;.j

:;r,

:\-

:>

:;j

10

11

42

4,

x,

-11 35

M

1

1

-10 40

15

0

-a 15

H

0

-S 50

0

-7 55

0

-660

0

-5 65

0

-4 70

0

-3 75

1

1

2

4

-2 80

1

2

3

-1 85

2

1

1

2

2

1

1

S

2

i

23

0 90

'.

1

1

1

2

2

1

3

i

0

M

2

5

1

1

42

1 95 11A

1

2

1

1

2 3

!

2

K

7

2

1

1

i

38:

2 1UOI03

1

23

1

?

->

2

'

2

19 1

3 105jllO

2

1

2

1

i

1

3

1

1

17 (

4 HOilU

i

2

I

2

3

5 UJT'V

1

1

3

G 121125

1

1

|

!

o

1

0

2

i

1

1

1

!:;

7

11

17

2ii

S

12

u

IS

7

7

1

i

0

1

158

FIG. 10. — Correlation in the buds of Oenothera lamarckiana. No. 04129. Length of
ovary subject, length of hypanthium relative. Coefficient of correlation, 0.4215 ±

identical only in regard to length of hypanthium and length of cone. Consid-
erable differences are to be seen in length of ovary and thickness of cone, and
slight differences in thickness of ovary and of hypanthium. In the standard
deviation and coefficient of variability, on the other hand, there is essential
identity in every character except in the thickness of the cone, which was
significantly more variable in the self-fertilized plants. There are too many
possible sources of such difference to make any speculation regarding its
significance profitable at this time.

When the data in the third entry (c'}, is compared with those in the first
and fourth entries, the very interesting fact develops that in certain charac-

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

y _

-n

— s

-7

-'„

-&

-3

— I

0

I

J

•;

i

.',

j

7

s

!i

lu

i.

-0

a

22

23

2!

25

-G

•11

•>

29

!';

j!

Ai

,!..

•••

So | 30] 37

-

.:

'->.:

*1

•1

-4 70

7.".

I

-3 75

SO

-2 80

*.">

I

j

•2

1

2

2

i

1

2

!'•

-1 JO

w

1

1

1

7

a

:.

:;

::

i

!

:

1

.;

17

0 90

<.!'„

1

1

1

j

1

7

-

1

'"

"'

2

3

1

2

1

4o

2 100

<(i.J

,;,

,

,

]

1

,

,

i

,

" Co

'.'.<,

i

i

:

_

,

!

y

1

ters, as thickness of hypanthium and length and thickness of cone, the plants
here represented stand intermediate between those of a and c, and that the
mean length of hypanthium was even less than in either, though it should
have been expected
to be as great as
that of a ; in other
words, the speci-
mens of O. lamarck-
iana which sprang
from a rubrinervis
parent had certain
characters interme-
diate between pure-
bred 0. lamarckiana
and 0. rubrinervis,
and in length of
hypanthium they
were even less than

those of the latter species. These plants were considered typical Oenothera
lamarckiana, not only by myself, but by others who are familiar with that
species. Here again, as in the. rest of the data presented in this table, the
general validity of any conclusion reached may be questioned because of
the small amount
of data.

Comparison of
the entries c and
d, representing the
variation in the
buds of two dif-
ferent pedigrees of
0. rubrinervis,
shows that only in
length of hypan-
thium and thick-
ness of cone were

FIG. 11. — Correlation in the buds of Oenothera lamarckiana, No. 04113.
Length of ovary subject, length of hypanthium relative. Coefficient of
correlation, 0.4719 ± 0.0413.

the two
essentially

lots
alike

1

X,

rll

-in

0

1

•J

:!

1

;,

6

7

8

e

10

.!;

J'..

:iO

1!

3J

;.;

51

';.->

.iii

37

3,s

.V.I

40

U

12

13

ii

!o

10

17

l.s

19

x,

!

-i> i.u

U

1

1

-5|_65j

n

o

-i ',<•

n

1

:

!

1

1

5

-3 75

M

j

J

1

0

-2 80

;,:,

1

.

!

?

8

1

:•-

l

2

-

•til

0 'JO

M

1

1

1

1

i

»

S

1

Sj

i 05

Ai.

1

2

2

j

i

.

7

:.

?

S

.

-12

2 IOC

1C,

i

_

2

'•:

:

:

•

L

!

!

17

3 Hi

IK

i

:

•

'.

1

1

to

4 llo

li:.

:

1

5115

HI

1

^_

1

1

0

0

1

i

'

7

;,;

i!

Ii!

1»

J

i.>

i :

1'

i!

:

-

2 i 1

-

|

in mean value,
the means of the
other parts differ-
ing from each other by more than the sums of the probable errors. In respect
to variability, entry c shows considerably greater coefficients in every set of

FIG. 12. — Correlation in the buds of Oenothera lamarckiana from Hilversum.
Length of ovary subject, length of bud-cone relative. Coefficient of correla-
tion, 0 . 6340 ± 0 . 03 1 9.

34 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

measurements than entry d. The two lots of plants here compared may be
contrasted with respect to their origin and treatment as follows, and seme
one, several, or all of these contrasted factors will probably satisfactorily
account for the difference in variability :

c.

d.

Seed from the New York Botanical Gar-
den.
Of hybrid (?) origin, O. iamarckiana X
biennis.
Pollination not guarded.
Seeds sown Feb. 16, 1905.
Buds collected Sept. 15 to 28, being the
last half of the season.

Seed from the Botanical Garden of
sterdam.
From pure-bred O. Iamarckiana.

Pollination guarded, self-fertilized.
Seeds sown Mar. n, 1905.
Buds collected Sept. 30 to Oct. 10,
early in the flowering season.

Am-
being

It remains to compare the three mutants here represented — 0. rubrinerms,
0. gigas, and O. lata — with the parent species in regard to mean values of
their several characters and the coefficients of their variability. This can

_ not be done with entire
satisfaction because of the
inconsistencies in the meas-
urements of plants belong-
ing to different pedigrees
in the same species. Such
comparison would be just
only when the lots of mate-
rial of all the species had
received identical treat-
ment, and this, as we have
seen, has not been true of
these. Wherever several
lots of plants of a single
species have been used, the
average condition of the
several lots has been taken
for comparison, and as a
given amount of deviation
in a constant of one value
is unfairly compared with
a given deviation from a
constant of a different value, all the deviations have been reduced to per cents
of the equivalent measurement in Oenothera Iamarckiana. These values are
arranged in the following table in such a way as to allow a comparison

x,

-7

-0

-5|-4

->

-2

-1

0

1

2

:;

4

.-,

(i

1

8

B

:;•:

:;;;

Si So

m

r,

!5

:J'J

-10

4142

43

44

45

M

47

4S

Xi

-llj 3j

10

1

1

-10! io

i;

0

-9:i5

ou

0

-8J5U

'"•'

0

-7 05, 60

0

-GI eo; 65

0

-oj CSJ 70

r<r

—1 70J 75

0

-3| 75 80

i

1

i

i

4

-2\ 80 85

1

1

1

3

-1 85J90

•;

2

23

0 9t.|95

i

i

i

10

6 | 9

i

2

2

1

42

1 951100

1

1

1

i

.-,

5

5

7

1

•2

1

3tf

•2 100:105

i

i

2

2

1

1

1

1

1

19

3 105110

2

1

1

a

a

4

1

1

17

4 UO'115

1

1

i

1

2

7

5 115J18,

i

1

:;

6 120'125

1

1

i

:;

2

8

9

•1-2

li,|

•JO

16

11

17

(i

0

li

1

:•:

3

15S

FIG. 13. — Correlation in the buds of Oenothera Iamarckiana, No.
04129. Length of ovary subject, length of bud-cone relative.
Coefficient of correlation, 0 . 563 1 ± 0 . 0366.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 35

between the deviation in the mean, dM, and that of the coefficient of vari-
ability, dV, not only of the characters studied in 1905, but also those of
the preceding year derived in the same way.

Deviation of means and "variabilities, in per cents of the corresponding values for Oenothera

lamarckiana.

dM

dV

dM

dV

O. rubrinervis, 1904.
Length of leaf
Width of leaf

1.02

20.95
24-75
104-59
172.30

74.28

20. 19
25-33
6-93
5-98
I5-I9
4-57
21.38
4-38
7.66

- 7.21
6.91
8.08
- 4.46
116.34

492.90

— 22.26
2.89
16.95
11.23

33-22

-366799
:'s:3

O. rubrinervis, 1905.
Thickness of ovary
Thickness of hypanthium
Thickness of cone .......

O. gigas, 1905.
Length of ovary
Length of hypanthium. . .
Length of cone
Thickness of ovary
Thickness of hypanthium
Thickness of cone

O. lata, 1905.

^8.46
14.84
10.26

9 52
20.67
1.54
37-03
42-38
29.63

!0 . 53
15.06

22.21

i8.n

13-87
13-87

ii .40
6.00
— 9.60

1.03

57-70
55-87
41 . 16
83-37
14.21

6.53

57-70
72 25
-13-77
- 5.61

8.58

Leaf ratio

Number of branches . .
Total branch length

O. nanella, 1904.
Height

O. rubrinervis, 1 90 5.
Length of leaf

Width of leaf . .

Leaf ratio

Height

Number of branches
Total branch length ....
Length of ovary

Length of hypanthium. . .
Length of cone
Thickness of ovary
Thickness of hypanthium
Thickness of cone .

Length of hypanthium . . .
Length of cone

In this table all the deviations which are less than the sum of the probable
errors are printed in bold-face figures, and may be left out of the discussion
as having no significance. If the signs of the remaining numbers are taken
into account it is found that 16 of the 30 characters studied showed a greater
variability in the mutant than in the parent, while only 4 were significantly
less variable in the mutant. Thus it is found again, as in 1904, that there is a
well-marked tendency for the mutant to show a greater variability than its
parent, though this relation is far from being absolute. The hint then recog-
nized, that the degree of departure from the mean of the parent might be
related to the degree of increase in the variability, is also faintly repeated, but
is too slight to be now considered of any significance.

The single study in the correlation of characters in 1904 showed a consid-
erably less correlation in the mutant than in the parent species, and we have
already seen in the study of leaf characters above that, though that study
was unsatisfactory, such evidence as there is seems to accord with the view
then expressed that this represents a general fact. In order to gain further
evidence on this point the coefficients of correlation of the parts of the buds
were calculated, and all the original measurements are given in the correlation
tables appended.

36 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

The coefficients of correlation are as follows :

0

I)

c'

c

Length of ovary
Length of hypanthium . . .

o.42i±o.o44
o.s63±o.037
o.477±o 052
o.s64±o.037
b. I57±o.o52
o.873±o.oi3
o.55S±o.037
o.65i±o.03i

> 0.518^0.039

'

o.387±o.045
o.634±o.o32
o.6oo±o.o34
o.58o±o.o35
o.044±o.053
o.785±o.o2o
o.573±o.o36
o.569±o.036
o.532±o.o38

o.472±o.04i
o.572±o.036
o.272±o.o49
o.535±o.038
o. I33±o.o52
o.829±o.oi7
o.556±o.037
o.59i±o.o35
o.500±o.040

o 298±o.049
o-598±o.o34
o.596±o.o34
o.3/5±o.o46
o.oi9±o.o5i
o.835±o.oi6
o.703±o.o27
o.63i±o.o32
o.7i7±o.o26

Thickness of ovary ...
Length of hypanthium . . .

Length of hypanthium. .
Thickness of hypanthium
Length of cone
Thickness of cone
Thickness of ovary
Thickness of hypanthium
Thickness of ovary
Thickness of cone
Thickness of hypanthium
Thickness of cone

d

e

/

Length of ovary
Length of hypanthium
Length of ovary
Length of cone
Length of ovarv

1 0.22I±0.053

o.246±o.o52
- o. 336^0.050
o.507±o.o4i
o.i53±o.o55
o.840±o.oi6
o. 5i4±o.o42
0.50 1 ±0.042
0.540 ±0.035

o.o94±o.o56
o.534±o.o5i
0.636 ±0.042
o.527±o.o5i
o.438±o.o57
o. 456±o.o56
o.838±o.o2i
o.475±o.o75
0.482 + 0.054

0.624 ±0.043
0.638 ±0.034
o.357±o.oso
o.o25±o 057
o. in ±0.050
o.837±o.oi7
0.640 ±0.034
o.s84±o.o38
0.555*0.039

Thickness of ovary

Length of hypanthium. . . .
Length of cone
Length of hypanthium ....
Thickness of hypanthium .
Length of cone
Thickness of cone
Thickness of ovary
Thickness of hypanthium .
Thickness of ovary

Thickness of cone

Thickness of hypanthium
Thickness of cone

A comparison of the coefficients of correlation in the characters of the sev-
eral mutants with those of the corresponding characters of Oenothera lamarck-
iana shows that in 14 cases out of 27 there is a decrease of correlation in the
mutant ; in 1 2 instances there is an increase, and in i the correlation remained
the same. These results seem to negative the view that the characters of the
Oenothera mutants are generally less closely correlated than those of the parent.
However, the differences between the coefficients of the mutants and those
of Oenothera lamarckiana are not generally greater than those between the
several lots of O. lamarckiana, or between the two lots of 0. rubrinerms, and
it is evident that the question is still to be considered undecided, but with
the burden of evidence in favor of the view that little or no change takes place
in the correlation of parts when mutation changes their proportions.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 37

X-.

-6

-i

0

1

2

3

1

8

(i

7

30

31

32

3:;

31

3.-.

3!i

37

3S

;!'.i

10

41

•12

13

11

X|

-4 70

?:,

1

1

"3 75

SO

0

-2 80

X.-j

i

E

1

1

1

:;

15

-1 85

JO

1

i

a

!)

11

r,

0

6

i

i

1

47

0 90

<l.l

i

:;

i;

2

7

s

8

5

1

1

4o

1 95

.00

::

:;

r.

s

.s

5

1

1

37

2 100

HI;,

i

2

1

:;

1

1

12

3 105

110

i

2

2

2

2

9

1

0

0

:;

10

23

17

22

2!

11

17

(i

:>

!l

::

81

FIG. 14. — Correlation in the buds of Oenolhera lamarckiana ,
No. 04113. Length of ovary subject, length of bud-
cone relative. Coefficient of correlation, 0.5720 ± 0.0358.

1

I

X

-7

-ii

-B

- 1

-3

-*|-l

0

1

2

3

1

5

ii

7

:.i;

•-1

2s

2U

30

31

32

n

31

35

3;

37

to

3'J

!C

x,

-C Ot

U

i

1

-5 to

7(

0

-4 70

7.':

1

1

1

1

1

5

-3 75

M!

2

2

1

5

-2 SO

;;..

I

1

25

-1 80

'.:•,

1

11

i

2

1

8

1

23

0 %

;:.,.

2

B

.->

B

2

2

1

30

1 95

icy

7

y

to

1)

1

3

42

2100

Ui

1

I

8

8

1

3

17

3ia

:iu

2

3

2

1

1

1

10

4110

n.-)

1

1

5115

Ito

1

1

i

i

1

1

7

7

3(5

21

32

23

I.')

10

1

o

1

160

FIG. 15. — Correlation in the buds of Oenolhera lamarckiana
from Hilversum. Length of ovary subject, thickness of
ovary relative. Coefficientof correlation, 0.5997 ±0.0341.

209134

38 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

\

X.

-.->

- I

-:;

-1

o

1

2

2

4

5

i

27

•2-

2'J

:>u

:fl

:;2

;:;

:;i

:;:,

:;i;

37

:;•-

X,

-1135

10

1

1

•1040

to

0

-9 45

:,ii

0

-8 50

:>:.

0

-7 50

0,.

0

-G 60

i.-.

|_0_

-5 00

:•;

0

-4 70

7.".

0

-3 75

su

1

1

1

1

4

—2 80

H

1

1

£

-1 80

y'o

1

1

:,

9

.".

-i

1

a

0 90

'..;,

1

I

ii

1U

<;

I

2

42

1 95

uu

1

2

!l

-

11

i

2

38

2 100J105

1

1

1

2

7

i

1

2

U

3 105110

1

•2

5

i

I

2

1

17

4 110110

1

2

1

2

1

7

5 115

1*

1

!

3

6 120

X,

1

1

1

1

•2

17

2;

41

5

a

10

:,

1

2

58

FIG. 16. — Correlation in the buds of Oenothera la-
marckiana, No. 04129. Length of ovary subject,
thickness of ovary relative. Coefficient of cor-
relation, 0.4766 ±0.0522.

I

XJ-3

-2

-i

0

1

2

1

1

1

1

28

28

30

31

32

33

SI

GO

3d

37

x,

-4 70

70

J

1

-3 75

60

0

-2 80

SO

1

I

B

2

1

1

15

-1 85

%

!

5

13

9

9

4

1

47

0 90

90

1

7

S

S

S

1

2

2

i

40

1 95

mi

1

2 3 9

6 : 6 | 4 [ 1

37

2 100

I,,:,

1

2

3

1

1

1

1

12

3 105|llO

1

1

1

2

I

1

9

7

i<;

:',ii

M

30

22

10

4

1

1

161

FIG. 17. — Correlation in the buds of'Oeno-
thera lamarckiana, No. 04113. Length
of ovary subject, thickness of ovary rel-
ative. Coefficient of correlation, 0.2724
±0.0492. i

1

X,

-ii

-10f-9

-8

-7

-r,

— ."»

—4

-S

-2

-1

0

1

2

8

-1

.-,

C

7

8

!)

10

2S

29

M

II

!2

.;:;

?)

35

30

37

38

IB

U

11

12

13

41

43

HI

47

IS

ill

Xi

• -11

22

1

1

2

••10

2:;

1

1

-9

2)

1

1

-S

2.1

1

1

2

-7i

a ;

1

1

1

1

4

^|27

1

3

2

1

1

g

-5J28

2

2

2

1

1

1

9

-4|29

2

1

1

1

5

-3|SO

1

2

1

1

2

1

8

-21 SI

1

1

2

3 1

3

14

-132

2

1

3

2

2

1

1

12

0

:;.:

2

1

1

1

5

1

:;i

i

1

2

I

2

4

2

2

1

18

2

K

i

1

2

I

1

32

]

1

20

3

M

1

1

3

1

2

1

1

10

4

37

1

I

1

4

|

1

1

13

5

3-

1

2

1

2

2

1

3

1

1

1

17

6

•

2

3

1

8

7

n

2

1

3

1

1

0

0

1

1

7

7

10

U

1'J

10

2(i

10

13

12

11

7

2

2

1

2

60;

1

FIG. 18. — Correlation in buds of Oenothera lamarckiana from Hilversum.
Length of hypanthium subject, length of bud-cone relative. Coefficient of
correlation, 0.5797 ±0.0354.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 39

H-B.

™ _, B

!!

a E8>

sit

III

I'H

40 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

H8

v r,

?i S

rr

?i

-

-

-

M

H

a

rt

-

-

M

H

'-.-,
s

rt

—

—

M

~

-

i~

-"

-

w

-

_

f-

0

^

rt

d

^

«

^

^

-i

«

:7

_

_

rl

—

s

*?

n

-

-

H

-H

N

-

M

M

:~-

-

2

X

8

^

a

i

?;

a

?.

t-i

f,

^

g

S

|

9

-

il

S

i-:

x

8

P

=

^

3

X

-

—

y

is

-

—

i H)

*

8

-

M

"S P

CO

f;

«4

M

<n S

ffl

?i
IT

•*

-•

^

2

-'

n

CO

**

°*

-1

•*

^

0

r,

9

~

—

rt

—

,,

„

^

„

„

^

Si

5^

V

1

tl

M

-

n

7

a

H

*

_

H

-H

H

•J

•3 §

?j

-H

rt

Cl

y

x^

?!

!!

n

^

"i

r,

i

,',

a

8

S

9

IS

3

^5

S

i

ss

ft

?

« 8

x

-4—

-i •

'

i

-*-

•A

-1-

—

-i—

2

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 41

jx2L2

-1

0

1

8

m

5

65

70

70

so

•

9095

100

70

V.i

80

S5

'JO

96

m

105

Xi

•-H'2!<

1

1

1-10,29

1

1

hyiso

0

hS 31

0

f-7>32

1

1

j-C 33

1

1

|-534

2

3

1

i

7

-4135

1

4

2

7

-8

:ii.

1

2

11

2

1C

-2 b7

i

3

7

11

-i i

2

5

10

i

1

ia

o 3D

2

5

:;

10

i

40

2

7

6

s

20

2

\:

i

.-,

1

1

16

3

\.

'l

8

i

13

1

!:<

1

7

1

12

5

!!

l

1

2

i

11

C

•ir.

:;

2

2

7

7

!•:

2

2

3

47

1

1

^2_

9

!>

1

1

10

i..

1

]

P2~

5

10

38

55

:b

io

2

1

100

FIG. 24. — Correlation in buds of O. lamarcki-
ana from Hilversum. Length of bud-
cone subject, thickness of bud-cone
relative. Coefficient of correlation,
0.7852 ±0.0204.

X2|-4

-3|-2|-1

0

l|2

3|

{•

70|75|80

81

'M

1)5

100!

70 75 80 85 90 05

11"

10-

Xil

!-7i 30

1

l

-6 31

0

-5; 32

0

-4|33

1

1

1

3

-3

31

1

1

5

10

-2

88

G

12

I

1

23

-1

3!i

8

8

G

2

17

0

37

i

4

u

1

22

1

88

2

13

8|1

24

2

39

1

7

10

a

1

21

3

40

5

0

3

17

4

41

2

1

6

5

42

8

5

6

43

1

7

1

9

7

41

1

2

3

1

2

14

32

48

38

22

4

61

X,

-4

-3

-2

-l

0

1

0

3

.1

.-,

0

CO

H

7(1

75

M

K

HO

M

;*

UM

no

i

--

32

1

a

1

4

CG

:;;

2

1

3

-0

81

2

2

-1

:;.,

2

i

2

8

-3

3,

1

i!

2

,9 |

-2

8

10

li

:;

£

-1

38

g

ft

4

1

19

_£_,

3!)

3

it

7

1

20

1

40

li

8

2

16

2

41

3

5

I

11

3

42

1

6

B

1

17

4

4:i

l

r,

6

5

41

4

1

5

6

45

8

3

C

7

4(1

2

2

4

8

47

1

2

3

9

4S

1

1

1

3

1

0

s

30

34

30

y

3

0

i

58

1

FIG. 25. — Correlation in the buds of Oenothera
lamarckiana,No. 04129. Length of bud-cone
subject, thickness of bud-cone relative. Coeffi-
cient of correlation, 0.8732 ±0.0127.

X,

-4

-3

-2

-1

0

1

2

8

1

22

2:i

24

25

2(i

27

28

23

:iO

Xj

-7

20

1

1

-6

27

1

1

2.S

1

1

-4

211

1

1

-3

30

1

2

I

1

7

-2

31

1

1

1

3

1

7

-1

32

2

4

11

1)

5

8

36

0

S3

1

;;

12

a

2

24

1

31

2

1

11

11

9

32

2

35

1

i

7

13

1

23

3

30

1

2

8

8

1

15

4

37

i

1

j

s

10

5

88

1

1

6

3!»

0

7

40

1

1

1

G

11

27

in

17

1C

1

1

160

FIG. 26. — Correlation in the bud of Oeno-
thera lamarckiana, No. 04 11 3 . Length
of bud-cone subject, thickness of bud-
cone relative. Coefficient of correla-
tion, 0.8289 ±0.0166.

FIG. 27. — Correlation in the buds of Oenothera
lamarckiana, from Hilversum. Thickness
of ovary subject, thickness of hypanthium.
relative. Coefficient of correlation, 0.5727
±0.0358.

42 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

|X2-2

-1

0

1

t

:;

4

•2",

-I

2.->

21,

27

2*

2!>

Xi

-5 27

1

1

1-4 > 28

1

4

1-3 '20

1

2

|-2.«0

I

I

2

17

i-1 31

7

i

•1 2

23

|0 ;32

•2

22

8 3

-11

i 1 :33

2 ;io

21 4

37

i - ,34

2 4

3 ; 3

12

! 3 35

5 | 1

•2

2

10

4 36

1

3

1

5

5 37

1

1

2

4

6

33

1

1

2

Hi

24

48 47

17

3

3

581

I

FIG. 28. — Correlation in the buds of O.
lamarckiana, 0.4129. Thickness of
ovary subject, thickness of hypan-
thium relative. Coefficient of cor-
relation, 0.5579 ± 0.0370.

Xo-3-2-1 0| 1 ;2 '3 i4 !5 I

22 23 24

25 20 27 28 29 30

X,

-3:28

•2

1

1

1

7

-2:29

1

3

3

3

6

16

|-1|30

1 1 |4 |» J13

6|2[l

36

1 0 |31

2 i 6 |16

5

5|

34

1 i32

2

1

12

10

1

H

! 2i33

1

'J

1

1

1

1

22

3l34

•2

1

1

t!

10

4 35

2

1

1

4

5

3U

1

1

6

*?

1

1

•2

12

23

19

-10

22

11

1

1

g

FIG. 29. — Correlation in the buds of Oeno-
thera lamarckiana, Xo. 04 1 1 3. Thickness
of ovary subject, thickness of hypan-
thium relative. Coefficient of correlation,
0.5565± 0.0367.

XJ-2

-1

0

i

•2

i

4J5

GO

70

n

so

SO

;'n

jtjJM

7U

75

8U

so

'»

00

100105

X

-72C

1

1

-6J27

1

1

-5' 28

1

1

-4 2'J

1

1

-330

1

•2

3

1

7

-231

1

1

•2

3

7

-l|32

1

8

17

(1

1

36

~24

0 |33

II

7

a

2

1 34

2 |ll

10

ii

32

2 135

4

g

a

4

23

3 ;36

2

:.

4

3

1

15

4 j 37

3

a

1

1

10

5 38

i

1

G

J'J

0

7

10

i

1

.')

1U

M

55

38

Hi

2

1

60

FIG. 30. — Correlation in buds of Oenothera
lamarckiana horn Hilversum. Thick-
ness of ovary subject, thickness of bud-
cone relative. Coefficient of correla-
tion, 0.5691 ±0.0360.

Xoj-4

-3-2-1! Ojl 23

4|5 J6|

1 60 Co 70 75 '80 85 90 1 9o ioo'lOs'llo!

C5 70 75 80 85 90 90 HXH05110 llo|

Xj

-5^27

1

1

-4:28

4

4

-3 29

1

1

2

-2 30

1 7

8

1

17

i-1 31

3 H, 3

5 1

23

0 32

1 2 7 i!4

14 4 1

41

1 33

126

S 17 2

1

37

2 34

3 2

2 ! 3 1 2

12

3 35

3

4 2 1

10

4 '36

1

1

1

5

pffB

1

2

1

4

6 38

1

1

2

1

0

s

3ii

3fi

34

so y

3

0

1

158

|

FIG. 31. — Correlation in the buds of Oenothera
lamarckiana, No. 04 1 29. Thickness of ovary
subject, thickness of bud-cone relative.
Coefficient of correlation, 0 . 6506 ± 0 . 0309.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 43

X-_

— I

-:

-_

-1

0

1

2|3|

ta

70

70

M

H

'M

95 U

;.)

»

so

SO

no

K

100

100

Xj

-3i2b

•2

1

1

7

-2|29

2l 3

i;

5

16

-1

30

7

0

12

10

1

3G

0

31

1

14

10

7

2

34

2

3S

g

r

5

i

3

31

:;

(i

1

10

4

3:>

1

!

1

4

5

30

1

1

G

37

1

1

i

2

14

.1-

Is

:js

22

1

61

A-.

- 1

-3

--

-1

0

I

2

>

!

«| G

G:.

70

18

H

•Si

•JO

'.10

101

lu:,

11011.-

Xj

-2

23

°i

1

2

1

2

-

16

0

25

3

to

US

'.1

10

1

4S

1

21 1

I

I

11

15

1

1

1

47

2

27

1

2

3

o

1

:

17

3

2s

1

1

1

3

4

->!l

1

1

i

3

1

0

S

31 i

3li

34

30

!)

3

0

1

58

FIG. 32. — Correlation in buds of O.
lamarckiana, No. 041 13. Thickness
of ovary subject, thickness of bud-
cone relative. Coefficient of cor-
relation. 0.5910 ±.0.0346.

FIG. 34. — Correlation in the buds of Oenothera
lamarckiana, No. 04129. Thickness of hy-
panthium subject, thickness of bud-cone rela-
tive. Coefficient of correlation, 0.5180
±0.0393.

XJ-2 -I

0

i

j

a

4|5

M

TO

75

M

•

90

95 100

70

70

SO

so

DO

90 100

:

Xi

-4 22

1

1

2

-3,23

1

•1

1

G

-2 24

3

4

2

1

1

11

-1(25

3)9

12

8

27

Oi26

16 10 | 3

•2.

49

1 27

I10

15 17 1 5

1

47

2 i28

rma

q

16

fJjS

i

I

4;30

i

1

1

u

yj

55

3s

10

•1

i

IGO;

XoLj

-3

-i

0

i

2J3

fi">

70

79

so

SO

N

90 100

7o

7i

so

SO

to

'JO

iJlOo

X!

-3

22

1

1

2

-2

23

5

(i

1

12

-1

-'1

1

1

7

7

4

23

0

25

1

•2

'.i

17

13

7

49

1

20

1

7

10

1!

7

1

40

2

27

1

8

8

5

3

1

22

3

2s

2

2

B

•2

11

4

2'J

1

1

5

M)

1

1

1

2

i;

32

!§

::-

22

I

J16I

FIG. 33.— Correlation in buds of O. lam-
arckiana from Hilversum. Thick-
ness of hypanthium subject, thick-
ness of bud -cone relative. Coefficient
of correlation, 0.5316 ± 0.0382.

FIG. 35. — Correlation in the buds of O.
lamarckiana, No. 04113. Thickness
of hypanthium subject, thickness
of bud-cone relative. Coefficient
of correlation, 0 . 5005 ± 0 . 0398.

x»!-9

-8

-7

-G

-5

-1

~:5

-2

-1

D

1

2

3

t

5

G

7

S

0

21

•22

23

21

•2.')

20

27

28 2!l

30

31

32

33

31

SB

3G

37

3S

3D

xt

-4 85

90

1

2

1

1

5

-3|90

86

2

1

3

2

2

1

11

-2J95

KM,

1

1

2

1

1

i

1

1

2

1

2

2

16

-11001C5

1

2

1

i

1

2

1

B

2

1

2

1

1

33

o jios'iio

1

1

1

1

27

1 110J115

1

2

1

i

2

1

1

1

30

2 115120

1

1

1

2

1

1

•1

1

1

2

1

16

3 120|185

1

1

1

2

3

1

1

1

11

4 125130

1

1

1

2

3

1

9

5 130135

1

1

6 135 UO

1

1

1

2

2

3

4

S

8

13

10

13

15

20

IS

7

13

7

7

G

2

2

60

FIG. 36. — Correlation in the buds of Oenothera rubrinervis , No. 04113.
Length of ovary subject, length of hypanthium relative. Co-
efficient of correlation, 0 . 2985 ± 0 . 0486.

44 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

x

i'j

•18

-17

-Hi

•u

-1-1

i:)

-12

-11

-11)

-5

0

'

2

8

1

5

(i

7

11

a

13

14

13

Iti

17

18

111

2(1

21

22

23

21

23

21;

27

2S

21)

:',()

31

32

33

:il

35

:'.(!

:>',

x,

-590

'jr.

1

1

-4 95

100

1

1

-3100

iu:.

1

i

1

1

6

-aw

110

i

2

i

1

1

1

1

1

10

-1110

a.-.

4

4

1

2

3

1

1

2

1

23

o na

i-;»

1

1

3

5

4

4

a

7

2

1

4

1

35

1 120

rr.

1

1

:;

2

1

n

2

(i

3

2

1

34

2 125

i:;o

i

1

1

1

3

i

1

4

:;

•20

3 130

i::->

a

1

i

I

-

a

12

4 135

Hi'

i

1

5 HO

11 ,

1

i

1

3

i

a

0

0

0

0

0

0

0

0

1

U

0

1

2

7

15

15

13

11

ID

IS

19

12

7

1

1

46

FIG. 37. — Correlation in the buds of Oenothera rubrinervis, No."04137. Length of ovary subject, length
of hypanthium relative. Coefficient of correlation, 0.2211 ± 0.0531.

1 1

X.

-8

•4

-i

-s

-1

-sU-i

0

1

2

3

1

i

(i

7

8

9

10

11

12

2(i

27

28

2-J

30

31

32 S3 31

33

30

37

3^

3'J

10

11

12

13

14

15

-u;

17

Xj

-4 85

1)0

2

2

1

5

-3 90

•JO

1

1

i

1

2

2

2

1

11

-2 95

l')0

1

1

1

2

2

2

2

1

3

s

10

-1100

III.,

1

2

3

33

(i ioj

Ill;

1

1

1

1

1

27

1 110

11.-;

1

1

1

3

2

3

5

4

i

4

1

90

2 |115

r:i'

1

2

0

4

1

2

i

1

1

16

3 120

1::'.

_

2

o

i

1

11

4 125

at

l

1

2

1

2

1

1

9

5 130

i., .

1

1

6 135

110

1

1

1

1

2

i

1

7

1)

L2

ii;

21

24

111

17

S

(i

.i

2

2

(i

(i

0

1

160

FIG. 38. — Correlation in the buds of Oenothera rubrinervis, No. 04113. Length of ovary subject length
of bud-cone relative. Coefficient of correlation, 0.5976 ± 0.0343.

-

-'.

2J

21 i

2S

2<J

3IJ

31

32

33

31

35

3ti

37

:>

-ill

10

11

43

x,

-590

.1.)

1

1

-4 95

oy

1

1

-3100

0.'.

2

2

1

1

0

-2105

10

1

2

1

2

2

1

1

10

-1110

i.,

1

3

3

5

4

3

1

1

23

0 115

20

3

7

a

7

G

3

1

1

1

35

2:.

1

.->

i

10

2

a

4

2

1

34

2 125

:;ij

1

3

i

1

I

3

2

1

1

20

3 1»0

35

1

I

1

2

3

2

1

12

4 135

10

1

1

5 UO

15

1

1

1

3

.-

U

0

0

0

0

0

4

3

1'L

22

27

23

.11

11

7

?,

2

46J

FIG. 39.-Correlation in the buds of Oenothera rubrinervis. No. 04137. Length of ovary subject, length
of bud-cone relative. Coefficient of correlation, 0.2464 ± 0.0524.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 45

x.

-4

-3

-2

-1

0

i

2

3

4

:>

G

7

3

2!)

30

31

32

33

34

35

3(i

37

38

,7.1

40

41

Xj

-4 85

N

1

•2

1

1

5

-3 90

N

1

I

3

4

1

1

11

-- '•'•'

100

2

2

1

7

2

1

1

16

-1 100105

2

5

6

6

•2

8

4

2

33

0 105110

2

1

11

3

2

3

2

3

27

1 llOli:,

1

1

1

7

8

3

•2

3

1

30

2 115120

1

3

•2

B

1

1

3

16

3120J12S

I

1

1

1

2

•2

1

11

4 13.", 130

2

1

2

1

2

1

9

5 130135

1

1

6 135

110

1

1

:,

9

1-2

31

•2.:;

19

19

1'2

13

8

2

3

1

GO

X,

-4

-3

-2

-1

0

1

2

3

4

5

(i

7

32

33

34

3.-,

3(1

87

38

3'J

4d

41

42

13

xt

-5 90

;,;,

1

1

-4 95

.'K!

1

1

-3100

10.-,

1

1

t

1

1

6

-2105

10

2

2

2

2

10

-1 10

15

1

2

7

1

3

2

1

23

o ud

•id

3

2

5

B

10

1

2

35

1 120

•.':>

1

0

B

10

li

1

34

2 125

:;o

1

B

3

1

1

20

3 130

i:;.-

2

2

1

1

1

1

12

4 135

i"

1

1

5 110

r>

i

2

3

2

10

7

2(5

31

32

13

8

7

5

1

1

146

FIG. 40. — Correlation in the buds of Oenothera ru-
brinervis, No. 04113. Length of ovary subject,
thickness of ovary relative. Coefficient of correla-
tion, 0.5957 ±0.0344.

FIG. 41. — Correlation in the buds of O. rubrinervis,
No. 04137. Length of ovary subject, thickness
of ovary relative. Coefficient of correlation,
0.3357 ±0.0495.

x«

-!>

-S

-7

-6

-5

-4

-3

-2

-1

0

1

2

3

4

B

«

7

8

B

1(1

111 12

2(5

27

28

2!)

3D

31

!2

33

34

So

3li

37

3S

39

40

41

42

43

44

i:>

46

47

x,

-9

•21

2

2

-8

22

1

1

2

-7

•23

2

1

3

-6

24

1

1

1

1

4

"*

2.-,

1

2

1

2

1

1

8

-4

2(i

1

1

1

2

1

8

-3

27

1

1

1

2

1

1

1

1

I

I

2

13

-2

2^

1

1

1

1

3

]

1

1

10

-1

i'j

3

2

2

2

1

2

1

13

0

30

1

1

1

1

3

1

3

1

15

1

31

2

•2

2

2

1

3

li

2

20

2

32

1

1

2

2

3

3

2

1

2

1

18

3

33

1

1

1

2

2

7

4

34

2

1

2

1

1

3

1

1

1

13

5

3o

1

8

1

1

1

7

6

3(5

1

2

1

1

7

7

?,:

1

2

1

1

1

6

8

38

1

1

2

9

3!)

1

1

2

1

1

2

3

1

7

1

12

1(1

21

21

lil

r,

.s

li

5

2

2

0

ii

ii

I

160

FIG. 42. — Correlation in the buds of Oenothera rubrinervis. No. 04113. Length
of hypanthium subject, length of bud-cone relative. Coefficient of correla-
tion,0.3752 ±0.0458.

46 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

2 1

•s a

II

^> *-

H

-

-

*n 43

1-

?-:

-

-

-

-

-

rt

-

-

«

-

~

w »

CO

s

-

-

-

-

*'

M

«

-

-

-

-

2

3 Cl

N

s

-

n

B

-

-n

-

-

-

«

fi

c>

^

M

~

Cl

-

L:

•*-• fO

°

?i

!-7

**

*•

d

•*

•*

rt

*

•j

«

m

•°

*

*•

I

a

£3 ^

§°

-J

3

3

"~

_

„

•0

—

„

,.

(i

7

a

-

-

r°, £

x

1 •*

r,

1]

s

r;

i

-75

K

a

fi

8

i

u

i

rt

s

i

i-:

3

8

?l

-L-

^

.1.

-1-

••

|

9

9

*

-

M

rt

-

-

-

-

CO

**

-

M

c-

-

I*

n

~

-

ci

01

-

ci

S

T71

£

H

«r

"

c^

•J

;i

•<

s

-

1?

—

-

-i

^c

-

CJ

-

-

a

o

;i

a

Y

a

-

a

?

-

-

09

0

•o

eo

ca

rt

-

-

y

?

H

H

-

-

M

9

-

m

-

-r

=

0

o

s

0

V

fi

3

T

H

0

?

^7

O

a

r,

0

-

B

-

-

1

S

L:

£

~

r,

r,

?i

V,

r,

a

3

S

?i

?i

S

=

9

H

=

^5

s

S

x

z.

'-

^

-

U

-

S

~
i

j

•

1-

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 47

o

j^.^^+^^^^.

i- £• 5- | .<" S8ggS88SgB8

!.r c-r L- k L: !: U fe £ £ 5 ? 5 r £: "

:••=•=

°' ~ * "" o

g

-" J3 2 2 o M

- -

aZ

° *L £"• 0 £. ° '5 **

"

•i A

ft ;? 5- £ a «s « *• t- ww

tO M

g i

go

s -

c 33 g §• « _^ ±±^^± *

n M

g. - -?

sill ------^^^^

o

oo s; $P ^SS±fefei:S§§

—

2 _ 5- -" o

x

a 55

*

? 1 g P

« M - 83

els

sl^l 4^

to

J

S. " 5' t S

|-EE3-J«'- i- g§

o

gajg- « ,-» =

§~ g

«-

og-Sg- B «.c

ta

&. f = a 1 1 - - M "- ;

• - - - g §

M

s i

life

?;ij _¥___LL___

LrL^-!±--5L±l

ea«,|| Tr^^i^----

§^5-3.^ ^t^gSg^ggSfg:

BS588g«oS|

°2.A-S A

.x

B E I i S -

Sol

- s : ? i =

S s d,

H- ' c 3 S. M *-

s s M

0 O S". _* g ,i M

• B re -1 ""-w^

S S i

M g S <=

8 8 "

o S r4 o. M ^ M

S 8 -

S ~* $T £! "s «w-jCSgiS88««

A.OOOOOOH-

III! :

48 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

J

ix.

-1

-bj-2

-!

0

i

•2

••:

4

.">

21

25

2i i

27

2s

•J1.'

:•.')

n

32

33

x,|

35

i

i

5

_3

x

1

1

1

1

9

-J

31

!

i

2

•j

•j

u

-1

82

•J

g

t

10

2

i

i

34

0

33

!

i.1

!i

2

:;

•j

23

1

34

5

ti

i

2

i

i

19

2

u

1

l

1

r,

7

i

1

19

3

se

I

J

'i

12

4

;;:

•J

1

i

i

2

i

13

5

38

1

•j

5

8

B

39

1

1

2

7

4 I

1

i

1

3

8

I]

i

1

:

-

I'l

S2

.-;.;

17

23

13

14

i

16C

FIG. 48. — Correlation in the buds of Oenothera
rubrinervis, No. 041 13. Thicknessof ovary
subject, thickness of hypanthium relative.
Coefficient of correlation, 0 . 7026 i 0 . 0270.-

X,

-3

-2

-i

B

1

1

3

4

27

a

9

3' >

32

33

31

Xi

-4

32

1

1

2

-3

3:;

1

i

6

2

lOi

-2

31

i

4

2

7 |

-1

35

1

i

-,

I

7

2

1

2'J:

M

i

7

17

7

2

34

37

2

12

1<J

7

1

32 i

3-

3

:,

2

3

13

H

2

3

2

1

8

40

i

1

I

2

7

41

1

1

1

2

5

42

1

1

7

4:;

1

1

3

n

23

53

M

Iti

8 | 1 ! 46

FIG. 49. — Correlation in the buds of
Oenotltera rubrinen-is, No. 04137.
Thickness of ovary subject, thick-
ness of hypanthium relative. Coef-
ficient of correlation, 0.5041 ±0.0416.

xj-4

-:;

-•j

_]

0

1 | 2 I 3

1

7"

n

n

85

95 100100

IK

",

su

so

90; 90 100100,110

li:

A,

-4

2!i

2

2

i

5

-3

30

1

1

1

•2

i

9

-2

3!

!

r,

i

1

1

12

-1

3'2

4

3

'.i

13

3

2

34

0

3.:

2

2

1"

-,

1

3

23

1

Ji

1

3

i

i

2

1

1<J

2

30

3

2

;(

7

2

1U

3

3«

1

;

5 | 2

12

4)37

3

4

•J

3

1

13

5

3-

i

2

2

1

2

8

6

3:i

2

2

7

40

1

2

3

8

41

l

1

1

:,

lb

5

11

3!

•21

12

i

60

FIG. 50. — Correlation in the buds of Oeno-
. thera rubrinervis, No. 04113. Thickness
of ovary subject, thickness of cone rel-
ative. Coefficient of correlation, 0.63 10
± 0.0321.

X2'-4 ~3 --

-1|0|1 12

s !

70 75 SO 80 JSK)I 901100(105 |

75

SU

80|90

90

100 100 no

X,!

-4 32

1

1 1

2

-3: 33

1

3|6

10 i

-2 34

2 3

1

1

7

-1 35

1

8 111

3

:j

2<i

OJ36

|loll7

7

34

1

37

4

7

1218 1

32

2

38

3

7

2

1

13

3 39

5|3

S

4 40

2

4

;

7

|5 41

3

•2

5

1 6 |42

1

1

7 43

1

1

1

0

2

27

51

i;

2!)

2

ti

Fio. 51. — Correlation in the buds of Oeno-
thera rubrinervis. No. 04137. Thick-
ness of ovary subject, thickness of
bud-cone relative. Coefficient of cor-
relation, 0.5006 ±0.0418.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 49

X,

-4

-3

_2

-1

olih

3

1

70

75

SO

19

90 95 jl 00

11)5

llj

7J

SO

80

'.!',)

95

00

II.,

111)

110

X

-4

21

1

1

-3

25

2

3

2

7

-2'2G

1

S

S

3

1

1

1C

~~ 0

27

1

10

8

12

1

32

28

1

1

7

1C,

S

2

1

38

i :2'j

3

2

1

r.

1

17

2 i30

5

g

a

3

23

3 ;31

1

1

3

2

1

13

i 32

1

3

1

1

2

14

, 5j33

1

1

1

:>

Ifi

27

41

31

21

12

1

160

X;

- 1

-3

-2

-1

U

1

-

:;

70

u

so

85

'.>o

...5

UK.

105

75

XI)

SO

'.in

<ij

100

1(15

110

!X|

1-3

27

1

2

3

-2

28

2

2

2

6

-1

2V

i

12

1

23

0

.»

u

22

U

53

1

31

0

S

18

1

36

2

82

i;

B

16

3

n

1

3

1

8

4

31

1

1

0

2

J7

r,i

u

•JO

2

46

1

Fio. 52. — Correlation in buds of O. rubri-
nervis, No. 04113. Thickness of hypan-
thium subject, thickness of bud-cone
relative. Coefficient of correlation,
0.7170±.'0.0259.

FIG. 53. — Correlation in the buds of
O. rubrinervis, No. 04137. Thickness
of hypanthium subject, thickness of
bud-cone relative. Coefficient of cor-
relation, 0 .5397 ± 0.0354.

' 1 1

I 1 1

X,

-y

-S

-7

-6

-5

21

-3 -2 -1
22 23 24

0 1

«H

27 28

4

5

6

7 I 8

9

10

11

3-

13

r:s

M

•:,<<

15

10

17

1C

17

18! 19 20

25 26

29 30

31 32 33 34 35 3G

40141

-12

Xj

-3 So

'JO

2

1

!

1

2

7

-2 90

'.15

2

1

1

i

1

1

7

-1 95J100

i

1

3

r,

2

i

2

4

1

2

22

0 100:105

1

1

2

2

1

1

1

1

1

18

1 105

1(1

1

1

1

Lp

2

1

1

1

i

1

2

1

1

1

10

2 110

115

2

1

2

1

2

1

2

1

12

3 115

.•M

1

1

4 120

lij

I

1

2

4

5 1251

Hill

:

1

FIG. 54. — Correlation in the buds of Oenotliera gigas. Length of ovary subject, length
of hypanthium relative. Coefficient of correlation, 0.6237 ± 0.0432.

1

1

|

1

X.,

-fl

"S

-7

-li

-:>

- 1

-:;

-1

-1

0

1

2

"

1

5

ii

7

S

!)

10

11

30

31

32

3:1

31

35

30

37

38

S'.i

40

-11

42

43

44

-!.->

46 47

-IS

U

5(1

|X,

-385

90

2

1

2

1

1

7

-2 90

«6

1

1

1

?,

i

7

- 9;)

100

1

1

1

2

2

I

2

1

1

1

1

2

1

22

lOOilOo

1

2

1

1

1

1

1

1

2

2

1

1

1

18

lOo'llO

1

1

2

1

2

1

1

2

1

2

1

1

1

1

19

110115

2

1

2

1

1

1

1

1

1

1

12

115120

1

1

120^25

1

1

1

1

4

125130

1

1

FIG. 55.— Correlation in the buds of O. gigas. Length of ovary subject, length
of bud-cone relative. Coefficient of correlation, 0.5338 ±0.0506.

50 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

X.

-t;

-5

-4

-s

-2

-1

a

1

2

3

4

:>

(i

7

8

9

10

11

ss

3'J

10

41

12

-13

44

45

4G

47

48

4'J

50

51

52

53

-,!

-,.-,

x,

-3 83

uo

1

i

2

3

7

-2 90

'.15

i

1

3

1

1

7

-1 95

10(1

3

•4

(

S

1

3

1

22

0 JOo'lOo

i

2

2

2

2

•1

1

1

1

1

1

1

18

1 10SJ110

1

8

2

1

•2

1

1

2

1

1

19

2 110115

1

2

1

1

1

2

1

2

12

3 115120

1

4 120125

1

1

1

4

5 125

:>,(

I

i

FIG. 56. — Correlation in the buds of Oenothera gigas. Length of
ovary subject, thickness of ovary relative. Coefficient of corre-
lation, 0.63 60 ±0.0421.

| 1

X,

-9

-s

-7

-G

-5|-4 -3

-2

-1

0

1

2

3

1

5

G

7

S

9

10

11

-R

10

,51

32

S3

34

35

3fi

87

3S

:',!)

10

11

12

4:;

11

-15

4G

•17

4S

49

50

-9|16

1

1

-8

17

2

2

-7

1 .

2

1

1

1

5

-o in

1

1

1

1

4

-3 -20

1

2

1

2

1

7

-4 21

1

1

1

1

I

1

1

7

-3 22

1

2

1

1

1

6

-2 23

1

2

1

1

-1 24

1

1

1

1

1

1

1

1

8

0

25

1

1

1

1

1

2

1

1

I

1

11

1

2C

1

1

1

1

1

1

6

2

27

1

1

2

3

2.S

1

1

1

3

4

29

4

1

1

1

7

5

;;o

3

1

4

fi

31

2

1

3

7

:;2

1

1

8

:;.-,

1

1

1

3

9

::i

1

1

2

ioi

35

1

1

11

M

1

1

2

12

37

0

13

:;s

0

14

3D

1

1

15J 40

0

16 41

0

17

42

1

1

1

0

3

1

li

12

T

5

9

4

4

3

4

7

3

6

3

2

2

4

2

91

FlG. 5.7. — Correlation in the buds of Oenothera gigas. Length of hypanthium
subject, length of bud-cone relative. Coefficient of correlation, 0.5273
±0.0510.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 51

1?
p

'

s s .»

ft

t

0

•J

s

i

g

¥

i

y

¥

c_=

~

f

l/

-1

|

^

L.

11

g

§

'r

-

J?

-

1

o s 1

X

-

_

g

w

ill

o

!/

J,

0

5

i

Is: I'

M

-

-

-

U

J

§ S g.

-J

-

10

~

-

~

-

fl

i

o rt

ao

-

-

B

i

4^ 3* S*

S

M

-

-

to

-

-

-

if

to

* g* cr

M

-

-

M

>i

-

to

-

-

IS

JE

H- w o.

"

-

-

-

-

-

-

y

o

b ° o

CC

-

to

-

to

to

"

1-

-1

-

H

M

to

S?

10

-1

~

M

M

-

H

-

-

"

H

3 S.

-J

1-

M

-

to

-

-

S

»

If

c:

N

10

p-

-

-

t

Ci

-

N

~

to

ft

05 '

n *'

C

e

-»

r|

10

H

1-

t

100^

c

-

«

£

-

«

•=.

s

Pa

S1

o

n e. >o

S

1

S

i:

t

ti

±

r

ft

*

3

g

£5

i

g

y

s

-^
s

I

Q 0 1

2

§ S ?

-

«

8

o S 2.

H

-

M

N

if.

s i

-i

M

1-

M

M

-

i-

-

§

g i

§ .S- §

-

1-

M

-

M

-

to

-

M

to

»

=

8 J.

S* S* 5'

8

.0

M

-

03

M

-

05

-

0.

0=

-

-

R °

8 ff ?

jr

"

M

-

"

-

0.

M

"

0 **

|

-

*

,0

"

"

-

to

"

~

1

-S- —

sap

M

„

_

i

|

S *

" c" |.

-

^

g w

0 0*1

-

M

"=•

s «

o a

-

«

S

§33

i!

52 MUTATIONS, VARIATIONS, AND, RELATIONSHIPS OF THE OENOTHERAS.

|

X,

-8

--

-6

-5

-4

-3

-a

-1

0

1

2

a

1

5

0

7 s

9

10

2S

2'j

SO

P,l

12

S3

"A

15

.",11

,17

;s

39

10

n

-12

43

41

45

46

Xj

-G

:VS

1

1

2

-5

is:)

1

1

1

3

-1

40

i

1

3

1

1

7

-3

11

a

a

;;

9

-2

12

2

:;

1

2

2

1

11

-1

«

1

I

•1

:;

2

i

1

11

0

II

z

1

•1

1

1

7

L

15

1

•i

2

:;

1

i

10

2

H

1

2

2

1

6

3

47

1

:;

1

i

6

4

4S

1

1

i

1

4

5

41)

1

i

1

3

6

SO

1

i

1

G

7] 51

2

2

3 [52

1

1

2

8 53

1

1

10-54

0

11

55

1

1

h

0

0

•;

7

a

11

u

0

s

7

7

1

91

FIG. 60.T-Correlation in the buds of Oenoihera gigas. Thickness of ovary
subject, thickness of hypanthium relative. Coefficient of cor-
relation, 0.8385 ±0.0210.

XL,

-4

-3

-2

-ilo

1

2J3

I

8

G

3

H

tt

15

00105110

n.-.iio

i-r,

86

1.H5

g

'.'<)

.15

(HI

05[110115|120J125

ISO

U

10

m

*!

-6

;s

1

1

2

-5

yj

1

1

3

ul

10

a

2

2

7 _

-3

41

i

2

5

1

9

-2

42

2

4

1

1

11

-1

43

1

8

1

8

11

0

41

3

1

1

1

1

7

1

45

1

l

;;

2

:;

10

2

46

i

1

1

1

1

1

G

3

47

1

1

1

2

i

6

4

4*

1

1

1

1

4

5

41)

2

1

3

*

50

2

1

i

1

i

G

7

51

1

1

2

8

52

1

i

2

9

53

1

1

10

51

-

0

11

55

1

1

1

i

7

15

26

12

11

10

3

1

1

i

91

X-'

- 1

-3-2-1 Oil|2[3|4|5|6|7

86

M

95

100105: 10115J120125

130135140

'.10

li.-)

10(1

105

IK

115.120183

i:;i

135JHO

115

Xi

-8

2S

1

1

-7

2!)

0

-6

30

0

^Sl

1

1

1

3

-4J32

1

2

1

1

1

1

7

—

:;:;

2

S

1

2

8

—

34

4

4

3

1

11

-

35

3

5

2

s

13

3ii

1

•2

•2

1

6

:;7

2

1

1

1

2

1

8

3s

2

1

1

1

1

1

7

3:1

1

1

2

1

1

jr

40

4

1

1

1

7

41

1

1

1

6

42

2

1

1

4

4:1,

0

41

1

1

2

45

0

10

4(1

t

I

1

3

7

15

26

12

11

10

3

1

I

1

Oil

FIG. 61. — Correlation in the buds of OenotJiera gigas.
Thickness of ovary subject, thickness of bud-cone
relative. Coefficient of correlation, 0.4750 ±
0.0747.

FIG. 62. — Correlation in the buds of Oenothera gigas.
Thickness of hypanthium subject, thickness of
bud-cone relative- Coefficient of correlation,
0.4822 ±0.0543.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 53

A,

-1!

"11

•10

-9

-b

-(

-8

-4

-3

-2

-1

0

i

2

3

•1

1

«

7

8

g

10

11

12

13

M

15

n;

17

IS

1!)

20

21

22

23

24

2o

2t;

27

2S

2'J

30

n

.",2

;',;;

31

:;.-,

31 i

37

3H

;;•)

x.

-3 65M

;n

1

1

i

3

-2 70

75

]

3

1

1

i

1

1

1

i

1

1

1

15

-1 75

80

1

1

2

1

3

2

2

2

-1

1

2

4

i

2

i

1

30

0 sol

85

1

2

2

3

2

3

1

2

2

(i

3

2

i

33

1 83

<M

1

i

1

2

1

•1

1

3

1

1

I

1

2

i

1

25

2 00

'.).")

2

1

1

1

2

1

3

1

1

2

2

1

i

i

22

3 93100

1

3

3

1

8

4 lOOi

liM

1

2

3

5 105

llll

1

1

]

2

i |s

1

1

B

4

10

3

8

1 1

B

12

6

•1

2

1C

11

:>

a

8

2

1

i

i

140

1

FIG. 63. — Correlation in the buds of Oenolhera lata. Length of ovary subject, length of
hypanthium relative. Coefficient of correlation, 0.0944 ± 0.0565.

X.j

-11

10

-y

-8

-7

-U

-5|-4

-8

-2

-1

0

1

2

8

4

B

g

7

8

B

ID

10

20

H

22

23

21

25

26

27

2s

211

30

31

32

33

31

:;.->

3(1

37

38

3!)

10

Xi

-3 (io

7(1

2

1

3

-2 70

7.-,

i

2

2

4

2

1

2

1

IS

-1 75

M

1

2

2

t

B

3

B

2

4

4

1

30

0 80

to

1

1

2

2

r>

a

3

(i

1

1

1

!

33

1 85

M

1

4

1

1

1

r>

2

1

3

2

I

25

2 90

95

2

2

2

5

3

I

1

1

2

22

3 95

110

2

2

S

1

S

4 lOOjlOS
T 105 UO

1

1

1

3

1

1

1

0

0

1

4

(J

4

11

14

11

11

'J

10

11

11

.-)

7

1

.-.

I

1

1

401

FIG. 64. — Correlation in the buds of Oenothera lata. Length of ovary subject, length
of bud-cone relative. Coefficient of correlation, 0.6382 ± 0.0338.

x.,

-4

-a

-2

-1

0

i

2

1

•!

B

ti

34

:;.->

H

37

3,s

3'J

10

11

12

18

n

x,

-3 65

70

1

2

3

-2 70

75

2

5

S

2

2

i

15

-1 75

80

3

B

1

2

4

3

1

30

0 80

M.-,

i

:,

5

1

i

6

1

1

1

33

1 85

'.«>

1

1

1

1

7

->

3

1

25

3 95

KKI

1

2

2

2

"l

8

4 1COJ1C5

2

1

3 ^j

5 105

IK.'

1

1

2

i;

21

22

3

21 J

21

10

5

1

1

J40I

'FlO. 65. — Correlation in the buds of Oentthera
lata. Length of ovary subject, thickness
of ovary relative. Coefficient of correla-
tion, 0.3568 ± 0 . 0497 .

54 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

lit?

*

°'~ *

«r

H

-

-

« S £ §

rt

f:1

-

-

-

M

fl

3 s ~2

c,

-

-

-

-

-•'

•*

-

-

"

-

-

•M

-

~

0

•° s a o

•5 -2 - °-

0

fi

-

-

-

sill

*p

H

-

-^i

?

^T

-

-

-

-

-'

-

-

-

»

-

-

-

^ 0 S S

s

rl

-

-

-

-

H

M

= ^ " Z

g

V ~ 0 °

-

a

|

-

-

-•

r,

:'

*

*

8

I

ft

R

I

8

a

u

B

»

«

«

6

'"•

2

sill

•L

o

3 ""

.5 ^o

.= •*

ss

Ed

2

$

-

-

0

ft

-

H

-

~

»

v. 2

00

B

-

-

-

0 3

f

;?

-

-

-

-'

««

3 t

-

;-

-

*-

•«

-*

•w

S °

t^

;:

-

-

->

-

-

-

i-

i-5 "o

•«•

~

~-

-

-

-

-

L-

. a

i«

u

-

-

-

.4

n

-.1

-

~

-

-

-

-

•«

s

•S .^

«

9

-

"

-

-

3 *5

-

H

M

-

-

-

-

J O

o

~

-

-

-

-

:'

c<

-

•

§

T

-.

-

-

Cl

-

Cl

cc

-

-

O >

V

R

-

-

-i

-

;i

-

M

^

-

s

«- -g

1

17

-

-

0

-

-

M

-

;i

-

-

-

~

<« u

y

r,

-

-

-

ci

:>

M

M

-

5

3 g

f

fi

~>

-

-

-

-

u V

T

n

-

-

M

-

-

-

i

r.

-

-

H

-

-T

a •«

!!

~

-

s1*

I

r>

0

2

\

^

1 a1

S

-

-

-

0 ""

X

fig

-

-

:£

§

—

-

!

r>

?!

Ts

^

S

r,

1

$

S

S

9

5

i

s

s

«

s

2

^| 3

'"

•*-

|

o

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 55

XL

-2

-1

0

i

2

so

•

!>0

!i:>

UN

10:

111

110

l-.il

Ii-

i:;u

XO

'.Hi

y;>

UN

111."

110

11.-

l-.il

li-

i:;i

r.i-

Xj

•11

1!)

1

1

•10

20

0

-t

21

0

-8

22

1

1

-7

23

2

2

4

-6

21

I

a

i

6

-S

25

1

3

4

-4

26

6

4

1

11

-3

27

3

S

3

14

-2

2*

4

0

2

2

14

-1

2!)

1

7

3

11

0

30

2

1

1

2

9

1

31

3

3

1

10

2

32

1

1

4

1

1

1

11

3

33

8

4

2

14

4

34

1

1

5

5

35

1

1

a

2

7

6

36

i

2

1

4

7

37

1

i

3

5

8

38

2

2

4

9

3!)

i

1

1

1

4

10

40

1

1

0

24

33

30

24

15

0

0

1

0

1

40

FIG. 68. — Correlation in the buds of Oenothera
lata. Length of bud-cone subject, thickness
of bud-cone relative. Coefficient of correlation,
0.8366±0.0171.

I

x.

-3

-2

0

1

2

3

?

26

27

28

29

30

31

32

33

xt

-4

M

1

1

2

-3

::,-,

1

2

2

1

6

-2

H

1

7

i

1

2

1

24

-1

37

3

a

G

5

22

0

38

4

6

10

2

22

1

3'J

2

3

5

13

2

1

26

2

10

1

2

5

5

7

1

21

j}j

41

1

1

1

1

10

4

42

2

1

I

5

5

43

1

1

6

II

1

1

5

14

28

30

40

17

5

1

140

FIG. 69. — Correlation in the buds of Oen-
othera lata. Thickness of ovary sub-
ject, thickness of hypanthium rela-
tive. Coefficient of correlation, 0 . 6400
±0.0336

X,

-2

-1

0

1

2

•"'

4

5

a

7

8

So

!H)

•JO

100

lii:>

111)

IK)

]•:»

i-;.-,

i:;o

i:ij

Xi

-4

31

1

1

*2

H51

:;.-)

2

3

t

6

-2

31;

4

S

7

4

1

24

-1|37

5

S

3

G

22

0

38

1

a

4

3

1

22

1

39

3

i

10

1

S

26

_ 2_
3

40

1

2

3

(i

1

2

3

21

11

2

2

3

a

10

4

•12

a

2

1

5

5

43

1

1

6

11

1

1

6

21

33

30

•M

15

G

0

1

0

1

40

FIG. 70. — Correlation in the buds of Oenothera
lata. Thickness of ovary subject, thickness
of bud-cone relative. Coefficient of corre-
lation, 0 . 5839 ± 0 . 0376.

x.

-2

SO

-1
SO

0
'.10

95

100

100
110

1

110120

I "U

1-0

1-1

so

'.10

1)0

100 105

,:iO

1:11

ix.->

x.

-J

26

1

1

1

5

-2

27

a

4

3

a

1

U

a

S

10

i

1

3

28

IOJ20

•I

s

s

i;

I

2

30

"IT

30

1

r,

11

10

1

2

40

31

13

1

i

.-,

1

17

3

.52

2

1

1

1

5

4

33

i

1

6

21

33

M

21

10

6

0

1

0

i

•

FlG. 71. — Correlation in the buds of Oenothera.
lata. Thickness of hypanthium subject,
thickness of bud-cone relative. Coefficient
of correlation, 0.5549 ±0.0394.

56 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

PEDIGREE-CULTURES OF OENOTHERA GRANDIFLORA, O. BIENNIS, AND
O. CRUCIATA.

In order to test the mutative capacity of other evening-primroses, seeds
were obtained from correspondents in various parts of the world and as many
as 40 sowings were made in one year from 0. biennis. In only one strain,
however, did anything come to light which would be of interest in connection
with the chief purpose of these investigations.

Parental individuals were selected and verified by Dr. N. L. Britton in 1903,
and from the seeds furnished by them the plants were grown which furnished
material for the descriptive diagnosis published in a previous paper (Mac-
Dougal, Vail, Shull & Small, 1905). This is not the species growing wild in
Europe and cited by De Vries in his "Mutationstheorie."

Inflorescences of 4 separate individuals were inclosed in transparent paper
bags in such manner as to secure pure fertilization and the seeds preserved in
separate packages. Sowings were made in pans of sterilized soil, in accord-
ance with the usual custom, early in January, 1905.

A number of seedlings were seen in which the first foliage -leaf and all suc-
ceeding leaves were very much narrower than the parental type. Sixteen
individuals of this kind were found and preserved, appearing in the progeny
of all of the 4 parents represented. An exact count of the number was kept
only in the case of one parent, No. 8.17, in which 4 of these aberrants were
included in a progeny of 669 individuals and formed very nearly 0.6 per cent
of the entire number.

The aberrants retained their property of forming narrow leaves, which were
slightly paler in color than the parental type, and appeared to taper gradually
at the base into long petioles. One aberrant of 8. 17 gave off a lateral branch
at the base of the stem which developed broader leaves resembling those of
the parent. With the completion of development, the form of the capsules
and the entire habit of this branch were in exact duplication of O. biennis.

On July 2 one of the aberrants opened a flower which also showed some
departures from the normal, being characterized by the extremely long sta-
mens and by the elongated capsule, which was of equal diameter throughout
and did not taper to a point or toward the tip as in its parental form. The
first 4 flowers which were brought to maturity on a lateral branch were inclosed
in a bag and self -pollinated. The seeds formed were allowed to mature and
these were sown as soon as practicable in order to determine the constancy
of the new form.

The adult plant has a main stem varying from 80 to 150 cm. in height, and
branches freely from the base upward. The lower branches are long, slender
and assurgent, while the upper ones are much shorter. Numerous secondary

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 57

branches arise during the latter part of the season. The stems are nearly
terete below, but are slightly channeled above.

The stem-leaves are 10 to 16 cm. long, and not more than i cm. in width
at the middle, obscurely pubescent on both surfaces, thickish, yellowish green,
obscurely and irregularly denticulate, with a heavy, broad midvein. The
leaves are linear, and taper to the long acuminate tip and to the margined
petiole.

The bracts are narrowly linear and 3 cm. long. The conic portion of the
bud is about 12 mm. long and half that in diameter, pale yellow, and barely
tapering in the apical portion, being finely pubescent. The free erect tips are
irregularly acuminate and 2 to 4 mm. long.

The hypanthium is 25 to 35 mm. long and exceeds the calyx-lobes by half.

The filaments are 7 to 9 mm. long, stout, and the anthers 4 mm., slender.
The pistil is slightly shorter or longer than the stamens, measuring 4 to 6
mm., and this species is therefore self -pollinated like the parental form. The
stigmatic lobes are 4 to 5 mm. long and spreading.

The capsules are 25 to 35 mm. long, 4 to 5 mm. in diameter, generally
curved, sparingly pubescent, bright green and shining, scarcely tapering to the
apex, and, in general, much thinner than the parental type. The ovary is
about half or less than half the length of the mature capsule.

The sepals are 3 mm. long, the petals firm, 15 to 1 6 mm. long and 15 mm.
wide, being crenately emarginate (1905 c, p. 17).

A small package of the lot of seeds from which the mutant was obtained
originally were sent to Miss Elizabeth Billings, who sowed them in her green-
house at Woodstock, Vermont, in March, 1905, and later transferred them to
her experimental garden. The 499 plants thus obtained were kept under
observation by Miss Billings and Professor F. S. Lee, of Columbia University,
with the result that aberrant individuals were found which were identical
with those obtained during the previous September in the New York Botanical
Garden.

The development of the newly arisen form has been followed from Septem-
ber, 1904, until the present time. The first individual discovered came into
bloom on July 2, 1905, having been cultivated under glass during the preced-
ing winter. Seeds were matured 5 or 6 weeks later, and sowings have been
made from them, with the result that the aberrant type was found to produce
a progeny which contained not more than 12 per cent of the atypic form of
individuals, the remainder being of the parental type and unchanged. Hybrid-
izations between the aberrant form and the parental form gave the same
result, while the aberrant type when crossed with lamarckiana and with rub-
rinerms gave a progeny resembling in general constituency those produced
by typical biennis.

58 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

A similar aberrant of 0. cruciata has been found in many of the cultures of
that species. Much attention has been devoted to these forms, both as to
habits of growth and anatomy. The only parallel occurrences are those
which are to be found in other genera in which atypical individuals affected by
the enzymatic diseases may give a progeny in which the normal form of indi-
viduals predominates. The vigorous growth of the atypic individuals and
the comparative regularity with which it appears are well worthy of remark.
If these forms were found growing in the open, they might easily be taken as
belonging to a species apart from 0. biennis unless their progeny was tested
in pure cultures.

The pedigreed cultures of 0. grandiflora made in 1906 from purely fertilized
seeds of the previous year included about 1 500 plants. More than one season
of such cultures is necessary to secure definite results that may be depended
upon. The observations have been carried so far at the New York Botan-
ical Garden and the Desert Laboratory, however, as to warrant the assertion
that this species presents a complex progeny analogous to that of lamarckiana,
in which two well-defined mutants are readily recognizable by reason of their
striking differences from the parental form.

ORIGIN OF FIXED FORMS BY HYBRIDIZATION.

That the cross-fertilization of two forms may result in the production of a
unitypic or polytypic progeny in which the parental qualities appear in a
mosaic is well known. Components of a native flora have been suspected to
be of hybrid origin in a few instances, and by succeeding experimental tests
have been synthetized from the parents. The practice prevailing among
taxonomists of ascribing a hybrid origin to a newly discovered form, which,
in outward anatomical characters, is between two known species, is extremely
pernicious and is not justified by facts obtained in cultural work. The best
grounds for such a conclusion are to be found when two species alone occur
in a region, and the appearance of a third is attributed to hybridization ; but
even here the supposition that a mutation may have ensued is allowable.

The hybrid 0. lamarckiana X O. cruciata (form with long hypanthium and
slender bud), which was described in the previous publication by the authors,
proves to be a fixed form, and as it sets seeds freely and is self -fertilizing, it is
in every respect an independent species. Several hundreds of seedlings were
grown from purely fertilized capsules ripened in 1904, with the result that all
were seen to conform to the parental hybrid type in every particular. The
two parental forms do not meet in their native habitats, and hence this form
could not have arisen in a state of nature.

The reciprocal of this cross, that is, 0. cruciata X O. lamarckiana, consists of
three forms, one of which is indistinguishable from the 0. biennis cruciata
which has been received from various European gardens.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OP THE OENOTHERAS. 59

BUD-SPORTS—VEGETATIVE SALTATIONS.

One of the most interesting and least understood phases of heredity is that
in which the tract of embryonic tissue constituting a bud may develop and
put into external realization a set of characters wholly different from those of
the remainder of the individual. A large number of such instances have been
seen by gardeners, horticulturists, and farmers, although exact observations
on the hereditary qualities of such mutant branches are almost wholly want-
ing. In some cases the departure from the type does not affect the entire
branch, and it may not be made manifest until an advanced stage of its devel-
opment, causing some of the leaves or perhaps some of the flowers to show
atypic characters. When such partial vegetative saltation ensues it gen-
erally results in converting a lateral section of the branch or inflorescence into
the new form, and has been termed "sectorial variation" by De Vries. To
designate it beyond danger of misapprehension, however, it should be known
as sectorial bud-mutation.

The cultures carried on in the experimental garden during 1905 were char-
acterized by two remarkable bud-sports.

The hybrid O. lamarckiana X (0. lamarckiana X 0. cruciata') comprised
nearly a score of recognizable forms in the seedling stage. About two of
each type were transplanted to the experimental garden in May, 1905. Two
of the types included were apparently different from each other only in the
character that one of them bore flowers with broad petals resembling those of
lamarckiana, while the other was furnished with cruciate flowers, but with the
petals broader than the typical cruciata and quite as long as those of lamarck-
^ana. It was evident that the two forms constituted an illustration of
Mendelian combinations, alike in all particulars except as to the characters
united in the form and size of the petals.

One of the individuals with the broadly cruciate flowers bore a branch near
the base of the stem, on which were formed only flowers of the broadly-petaled
lamarckiana type. Some of these were purely fertilized and seeds preserved
for testing. These were duly sown in December, 1905, under glass, and
representative individuals brought to bloom in the open air in July, 1906.

When they came into bloom in July, 1906, it was found that the branch-
sport came entirely true to the type which it represented. With this it is
interesting to note that the purely fertilized seeds of the main stem bearing
cruciate flowers gave a progeny which contained some individuals bearing
flowers with broad petals, while the other strain, which bore only broad
petals, came entirely true to its type without deviation. No exact numerical
count was made, but it seemed evident that the cruciate petal is dominant
over the cordate, and that the plants described above exemplify the develop-
ment of a sport on a plant of the first generation in which the recessive char-
acter appeared and when extracted came true to the recessive character.

60 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

De Vries (1900, p. 86) described a case of vegetative mutation in a hybrid
between Veronica longifolia, which has a blue flower, and V. alba with a
white flower. The hybrid has a blue flower, but several cases were noted in
which the entire buds or portions of inflorescences produced white flowers.
These white sports, when self -fertilized, produced white flowers only, in the
preliminary tests cited.

Similar vegetative splitting is known in Datura hybrids, being seen by
Naudin in 3 individuals only, while sports of this character of Brassica, Raph-
anus, Abies, Anagallis, Helianthemum, Zea, and Cytisus are well known.
Nearly all other sports of hybrids are clearly of an atavistic nature, being
more or less direct reversions, sometimes rehearsing the juvenile characters
of one of the parents of the hybrids.

A second instance was offered by a form known as 0. ammophila. This
material was grown from seeds obtained from Professor W. O. Focke, which
had been collected on the coast near Bremen, Germany.

A large number of seeds were sown in sterilized soil in the propagating
house on February 10, 1905. A month later the seedlings obtained were
examined and a dozen representing the widest visible range of variation were
transferred to pots, while the remainder of the culture was discarded. With
successive repottings the number of individuals was reduced to 7, which were
transplanted to the experimental garden in the latter part of May, at which
time no differences of moment were noticed. During the summer 3 of the
specimens sent up shoots, while the other 4 formed dense rosettes with a few
short lateral branches.

About the middle of August the more advanced individuals came into
bloom. No careful examination of them had been made up to this time,
but it was now seen that one of them had formed a lateral branch at the
base of the main stem, which, by the spread of its branches, the shape and
expansion of its leaves, and its general vigorous growth had become the larger
member of the shoot. This branch, in the form and behavior of all of its
organs, including flowers and fruits, was an exact reproduction of the 0.
biennis which has formed the basis of cultures under that name in the New
York Botanical Garden. Branches of the sport and of typical biennis were
submitted to several botanists, with the result that they were found to be
indistinguishable by anatomical characters.

The main stem, which had branched in the usual manner, had been crowded
from its natural upright position and had assumed a half -recumbent position,
having the appearance of a lateral branch.

The suggestion lies near at hand, that 0. ammophila is a hybrid derivative
of 0. biennis, and that the vegetative mutation is simply one of reversion, after
the manner of examples cited above. No positive evidence upon the origin of
ammophila is at hand, however.

Oenothera aminophila with bud-sport. The narrow-leaved decumbent stock represents the typical parent,
while the more vigorous upright branch with broader leaves is the mutant branch which represents
Oenothera hiennis.

HEUOTYPE CO., BOST

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 6 1

Purely fertilized seeds of the mutant branch and of the main stem were
obtained, and these were sown late in September, 1905. Those of the main
stem came true to the type of the main stem, that is, to 0. ammophila. All
of the progeny obtained from the purely guarded seeds of the sport conformed
strictly to the biennis type (plate 9).

THE INDUCTION OF MUTATIONS.

One of the most important problems confronting the investigation in this
subject is that of the localization of the mutations in the life-history of the
plant. If we could fix upon the exact stage in which the direct changes
occurred, we then would be in a position to examine the protoplasts con-
cerned with a view to ascertaining what changes ensue in the chromosomes in
connection with saltations in inheritance. Furthermore, opportunity would
also be offered for attempts to determine the factors which operate as stim-
uli, or loosing agents in setting the mutations free.

A theoretical consideration of the subject seemed to indicate that the
changes constituting the mutations which give rise to atypic seedlings took
place in a stage previous to the reduction divisions in the embryo-sac or in
the pollen mother-cell. It was planned, therefore, to subject these structures
to the action of chemical agents at a time before fertilization had occurred.
The attempt was made to secure two forms of stimulation by using some solu-
tions of high osmotic value, and other mineral compounds which are stimu-
lative in low concentrations. A desideratum in such experimentation was to
use plants in which a large number of ovules were to be found in one ovarial
cavity. The solutions were injected into the ovaries by means of a physician's
hypodermic syringe. Operations of this character were carried out with
Begonia rotundifolia and with a species of Cleome, but with negative results,
and injections were also made into the central placenta of a number of flowers
of Abutilon abutilon, but in all cases at a very late stage, and so far no results
have been obtained with this plant. It is being used in a new series of experi-
ments, however, which, with improved technique, may be expected to furnish
some interesting material.

Attention was next turned to several species of Oenothera which were
being cultivated in pure pedigreed strains in the experimental garden.
Solutions of copper sulphate of i to 400,000 were injected into ovaries of
lamarckiana immediately previous to pollination and the pistils were purely
pollinated. The capsules formed from ovaries treated in this way were vari-
ously distorted as a result of the wounding, but a large crop of perfect seeds
was matured, which were sowed in September, 1905. A careful census of the
seedlings late in November made it appear that in the small number of mutants
present the usual mutants of this species occurred in normal proportions.

62 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE) OENOTHERAS.

A number of radium pencils were taken from a series used by Dr. Gager in
his investigations, and were fastened in an inflorescence in such manner that a
deleterious influence must have been exerted on some ovules, while in other
cases the effect must have been stimulative if the reactions of seedlings of
other plants may be taken to offer a fair analogy. The crop of seeds matured
in the treated ovaries was sown in September, 1905, and late in November a
census showed that only 20 normal specimens of lamarckiana were present
and that 2 mutants were also included. The rate of mortality was thus seen
to be twice as great in the type as in the mutants in this single test.

Similar tests were made with 0. biennis. Solutions of magnesium sulphate
in distilled water were used without any noticeable departure in the composi-
tion or behavior of the progeny arising from the seeds. Poisonous solutions
containing i part of zinc sulphate in 500 of distilled water were also used.
The seeds produced in the ovaries which had received this treatment contained
atypic forms of the kind seen in ordinary cultures in apparently the cus-
tomary proportion of about one in 200. In addition however, a single rosette
was found which differed widely from any known type, and of this form, which
was recognizably different from the parental type in many qualities, some of
the differences were plainly apparent even in the earliest leaves of the seedlings.
These differences have become accentuated in the adult plant.

The parental form has been under observation for five years in cultures
and in a wild condition. An aberrant form, which appears to be eversporting,
has been previously figured, and while this form appeared in the injected
or treated seeds in a normal proportion, yet the newest aberrant has not been
seen elsewhere. The probability must be taken into account that it may be
a mutant of rare occurrence, the cycle of which came within the experiments ;
but in either case it is plainly a mutant, and it only remains to be seen whether
or not it was induced by the action of the zinc solution. The presumption
seems to favor such a conclusion. Seeds of this form were harvested in
August, 1906, and sown immediately, with the result that the characters of
the new form were found to be fully transmissible, the first generation of the
progeny being duplicates of the parent within the limits of fluctuating varia-
bility. The entire plant is characterized by a much deeper green color than
the parental form and the leaves are slightly curled and twisted, owing to
inequalities of growth, and it reaches maturity quite early in the season.

In addition to the crop of guarded seeds harvested from some of the
branches, the available remainder were collected and sown in the greenhouse of
the Desert Laboratory, with the result that this second generation of the deriv-
ative was found to conform exactly in every individual to the derivative type.
Either the mutant does not intercross with the parent, although the branches
were in contact, or if a cross has occurred the derivative qualities are dominant.
This matter may be determined by the time this paper finds its way through

A. Raimannia and induced derivatives. The three rows of plants in the background include some large
normal individuals and small mutants grown from seed taken from a capsule which had been treated
with calcium nitrate. The remainder are mutants of the second generation.

R. Individuals grown from seeds taken from one capsule. The large rosette is of the parental type, the
smaller are derivatives two of which have come into bloom at the same age and are scarcely so high
as the normal rosette.

HELIOTYPE CO.. BOST

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 63

the press. Among the large progeny grown from guarded seeds was only one
which showed any marked variation from the blunt rounded leaves of the
rosette of the derivative, but this variation did not extend to the foliar organs
developed later.

The radium pencils which were later used on 0. lamarckiana were affixed to
an inflorescence in such a manner that the radium coatings were 15 to 25 mm.
from developing flower-buds. The corollas of many were so retarded that
they failed to open and fell off prematurely. At greater distances develop-
ment of the ovary proceeded but slowly and normal size was not reached.
Perfect seeds were formed in many of them, however, and these when sown
gave the normal frequency of the aberrant mentioned above, and which is not
to be confused with the one induced by the stimulative action of zinc sulphate.

Some decisive results were also obtained from Raimannia odorata, a member
of a separate genus of the evening-primrose family from Patagonia. During
the first season (1905) injections of the ovaries were made with several sub-
stances, with the result that an atypic form identical in all cases was found in
seeds from ovules that had been treated in various ways. Two such mutants
were secured from seeds of an ovary that had been treated with a 10 per cent
sugar solution, 10 from one that had been injected with a solution of calcium
nitrate i part to 1000 of distilled water, and one was also found in the
progeny from seeds taken from a capsule which had been exposed to the
action of a radium pencil.

In all of these injections ordinary jointed metal syringes were used, and the
water was from a single distillation in a copper still, so that no special reliance
may be placed upon the purity of the solution injected into the ovary.

Seeds from capsules which had been treated were harvested in June, 1905,
and were sown early in July. The atypic derivatives could be recognized as
soon as the cotyledons were fully expanded, and no skill was needed for their
detection. The parental form bore leaves villous hairy, or with ciliate mar-
gins, while the mutant was entirely and absolutely glabrous. The leaves of
the parent generally have an unbalanced linear growth of the margins by
which the blades become fluted. The excess of growth in the mutant lies
along the midrib and the margins become re volute. The leaves of the mutant
show much less expansion than those of the parent, the lower ones being
narrowly linear, 5 to 6 cm. long and 3 to 4 mm. wide, while the upper ones are
lanceolate linear, 3 to 4 cm. long or even longer, the mutant being much the
narrower. The parental type is of a marked biennial habit and near the close
of the season the internodes formed are extremely short, which results in a
dense rosette. The mutant does not make a rosette, by reason of the fact
that its stem does not alter its rate of elongation, but proceeds at a uniformly
rapid rate, thus presenting a leaf-stem which appears to be perennial in the
climate in which the experiments were performed. So rapidly did develop-

64 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

ment proceed that one of the atypic individuals opened a flower-bud on No-
vember 27, 1905, and seeds were harvested before the beginning of the year.
Other lots of seeds were obtained throughout the year and several sowings
were made, with the result that the atypic form was found to transmit its con-
stellation of characters to its offspring. Among the fully atypic individuals,
of which a few hundred have been grown, one was found with wide leaves like
the parental form, another with ciliate leaves, and another with undulate
leaves. This activation of parental characters is one fully illustrated by
lamarckiana and its derivatives and shows that the treatment has simply
thrown certain parental characters into a state of latency and awakened
others with which the parental characters are mutually exclusive as to external
manifestation. These individuals did not afford a transition or intergrading
series between the derivative and the parental form, however (plate 10).

In the summer of 1906 a second series of injections was made, with the result
that capsules treated with zinc sulphate, i part to 2000 parts of distilled
water, yielded seeds which produced progeny inclusive of the atypic form
described above, and also some other combinations which it has not been
possible to follow in successive generations. The group, however, bears a
general resemblance in relationship to that of the lamarckiana mutants.

An injection with calcium nitrate, i part to 1000 of distilled water, was
without effect and the progeny were all of the parental type. It was remark-
able, however, that a treatment with this substance in the previous season
had secured some atypic forms, probably due to some opportune condition in
the experiment not yet understood. Furthermore, the plants of the progeny
of the first treatment which were apparently normal yielded seeds which gave
a few atypic forms, indicating that the effect of the first treatment had been
more or less permanent. Confirmation of this important matter has not yet
been obtained, however. Then, again, an injection of ovaries in 1906 with the
distilled water subject to impurity from the still and from the syringe, as men-
tioned above, also resulted in a progeny in which some atypic forms were
found to occur. So far, then, as experience with this plant is afforded it is to
be seen that a variety of agents act in inducing discontinuous variation in the
progeny and that in one instance the variability was carried to the third
generation, as far as that part of the test has been extended. The atypic
forms transmit their qualities perfectly from generation to generation, and
the third generation now in hand are like the first from which they came
originally.

(M W-. ;'•?

A. Ocnothera grandiflora, seedling about two months after germir

B. Ocnotbera grandiflora, rosette five months old.

HEUOTYPE CO., BOSTON.

IDENTITY OF EVENING-PRIMROSES.

PREPARED BY ANNA MURRAY VAIL.

In view of the extent to which the evening-primroses are being used to
obtain experimental evidence upon questions of heredity at the present time,
it has been deemed important to make a systematic study of the group. To
this end seeds have been obtained from correspondents in various parts of
America and from European botanical gardens in which these plants appear
to be extensively cultivated.

In the few years in which the cultures have been under way it has become
plainly apparent that a much larger number of species are native to America
than has been supposed hitherto. The series of pedigree-cultures made from
them, using carefully guarded seeds, shows that botanists have customarily
grouped many elementary species under the name of 0. biennis. Furthermore,
as has been noted on page 8, none of these appear to coincide with the form
cultivated in Europe under this name.

Several of these elementary species occur on Ivong Island, a number of them
collected by Mr. E. P. Bicknell deserving mention, and requiring special study
to determine their relationship to the common 0. biennis.

Another group of closely related but apparently distinct elementary species
came from the west shore of Lake Champlain. There, in waste ground at the
roadsides about Plattsburg, fairly typical O. biennis is to be found. A small-
flowered and very characteristic plant abounds on inland sand-dunes, in
sandy clearings along the Ausable River, and on the railroad embankments
near Bluff Point. In the Ausable River woods it grows associated with (9.
cruciata, and a supposed hybrid between the two was observed. Occasionally
a larger-flowered form on the edge of woods was observed and still another on
the gravelly shores of the lake. These types, with the exception of the hybrid
of which no mature seed was secured, were grown in the New York Botanical
Garden, where they reproduced in every particular the characters that distin-
guished them in the field. They will require more extended study for the
determination of their specific relationships.

The more important facts concerning the anatomy and distribution of a
few forms have been obtained and descriptions of 0. grandiftora, 0. simsiana,
O. oakesiana, O. parviflora, and O. muricata are here presented. It is proposed
to reserve discussion of the large number of other distinct forms until their
life-history and hereditary qualities shall have been more definitely ascer-
tained.

65

66 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

The use here of the name Oenothera for species belonging to the genus
Onagra (Tournefort, Adanson, Spach) is according to the decision of Dr. J. N.
Rose in a recent paper (1905), where he points out, and with good reason, that
Oenothera biennis should be considered as the type of the genus. The descrip-
tions in each case have been made from living plants.

OENOTHERA GRANDIFLORA AITON

Oenothera grandiflcra Ait. Hort. Kew., 8: 2, 1789.

Onosuris acuminzta, Raf. Fl. Ind., 96, 1817 (?).

Oenothera grandiflora ft Sims. Curtis's Bot. Mag. 46: pi. 2068, 1819.

Oenothera biennis var. grandiflora Lindl. Bot. Regist. 19: pi. 160*, 1833 (?).

Onagra grandiflora, (Ait.) Vail. Torreya, 5: 9, 1905.

Onagra vulgaris Spach. Nouv. ann. mus. Paris, 4: 353, 1835. (Reprint 33, 1835.)

In part.

Seedling about 2 months old. — Rosette loose, spreading, 7 to 12 cm. in diameter; outer
leaves 4 to 7 cm. long, 2 to 3 cm. wide; blades oblong, or oblong-spatulate, nearly glabrous,
irregularly blotched with dull red, broader above the middle, rounding or acutish and
shallowly undulate toothed at the apex, tapering at the more deeply toothed base into the
margined petiole ; the inner blades oblong, acutish, and soon becoming more deeply toothed
at the base (plate 1 1 , A) .

Seedling about 5 months old. — Rosette rather loose, spreading flat on the ground, about
30 cm. in diameter; outer leaves 12 to 15 cm. long, 3 to 4 cm. wide; blades oblong-spatulate,
broadest above the middle, acutish at the apex, tapering at the irregularly pinnatifid-
toothed base into a long, broad petiole, obscurely puberulent near the outer margin above
and beneath on the white midvein and veins, bright dark blue-green, irregularly blotched
throughout with red, which color also appears sometimes on the wide white midvein
(plate n, B).

Adult plant. — Plant 1.5 to 3 m. in height, branching from the base upward, the terminal
portion of the main stem bearing a cluster of short branches with a profusion of flowers.
Stems stout, often reaching a diameter of 6 cm. at base, nearly terete below, slightly angled
and channeled above, or sulcate, the epidermal tissues becoming detached on the lower
part of the main stem, which is hard and woody, clothed in the upper portion with short,
spreading, somewhat varyingly muricate pubescence; leaves 10 to 20 cm. long; blades ovate-
lanceolate, glabrate, with appressed scattered hairs on the veins, denticulate, being more
deeply and irregularly toothed in the basal portion, tapering to a short margined petiole,
bright green and shining above, paler beneath; terminal rosette-like cluster of the inflores-
cence symmetrical; flowers very abundant and fragrant; bracts lanceolate, 3 to 5 cm. long,
acuminate, abruptly tapering into a short petiole, or sessile ; conic portion of bud 3 to 4 cm . long,
6 to 7 mm. in diameter at base, very slender, tapering from base to apex, sparingly pubes-
cent, thin, dotted with small red spots, the erect free tips very slender, setaceous, 8 to 10 mm.
long; hypanthium 4.5 to 5.5 cm. long, very slender, much longer than the reflexed calyx-
lobes that usually cohere in pairs at the tips; ovary i cm. long, or less, slender, petals firm,
3.5 to 4 cm. long, and about 4 cm. wide, more or less emarginate, wedge-shaped at base;
filaments 2.5 cm long, very slender; anthers slender, i cm. long; pistil much longer than
stamens, and projecting from the flower and from the unopened buds late in the season;
stigmatic lobes 10 to 12 mm. long; capsule 3 to 3.5 cm. long, 8 mm. in diameter in thickest
portion, green and shining, glabrous or with a few scattered hairs, four-angled and tapering
from base (plate 12).

Alabama. — Earle's Landing and Dixie Landing near Tensaw, Tracy, No. 8001 in the
herbarium of the New York Botanical Garden.

Oenothera grandiflora. 1, stem-leaf; 2, leaf from young rosette; 3, bract; 4, bud; 5, flower
with petals removed; 6, capsule and bract; 7, petal (minimum size).

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 67

Texas. — Wright, without locality, in the Gray Herbarium of Harvard University (?).

Kentucky. — Neighborhood of Lexington, Short, without date, in the herbaria of Columbia
University and Academy of Natural Sciences, Philadelphia. The sheet in the Columbia
University herbarium bears the note "Oenothera muricata. Vespertine, 3 to 5 feet high."
This is the sheet referred to (MacDougal, Vail, Shull & Small, 1905, p. 6) as belonging to
O. lamarckiana, but which in the writer's opinion should belong here. The Philadelphia
specimen gives the exact locality "Neighborhood of Lexn fl. Augt. & after, i mile from town
left of Coles road to Frankfort. " It may be an escape from cultivation.

The rediscovery of the original type locality of this species is recorded in a
previous paper (MacDougal, Vail, Shull & Small, 1905, pp. 7 and 8). The
seeds from which the plants described above were raised were sent to the New
York Botanical Garden from Tensaw, Alabama, by Mrs. J. F. Davis at the
request of Professor Tracy, absolutely ripe capsules not having been found at
the time of his visit late in August, 1904. In cultivation, 0. grandiflora has
the appearance described by Bartram in 1793 and again by Professor Tracy in
1904. (MacDougal, Vail, Shull & Small, 1905, pp. 7 and 8).

It is very fragrant and showy, and flowers abundantly. Most conspicuous
and characteristic are the tapering pale-yellow buds with long, slender, seta-
ceous tips to the calyx-lobes. They have none of the heavy, swollen appear-
ance of those of 0. lamarckiana and of some of its derivatives. For 24 hours
or so before the opening of the flower the closed stigma crowded into the apex
of the bud is distinctly seen through the thin epidermis of the closed calyx-
lobes. The calyx-lobes when expanded are split open in twos (but very rarely
in fours), the petals only attaining their full size during expansion. This is
also the case in Oenothera argillicola (Onagra argillicola McKenzie). (Mac-
Dougal, Vail, Shull & Small, 1905, p. 12.)

A further comparison of supposed plates of 0. grandiflora with living speci-
mens of the species would indicate the following: The plate of "Oenothera
lamarckiana" in Lemaire (Illustration horticole, 9, p. 318, 1862) appears to
be that species, but the description of the 600 flowers, buds, and capsules on
one single plant would appear to refer to 0. grandiflora, as 0. lamarckiana does
not (in cultivation) flower so abundantly. The plate in Edward's Botanical
Register No. 1604 of "Oenothera biennis var. grandiflora" has much stouter
bud-tips, bracts that are broader at the base, and much broader petals than
the typical 0. grandiflora.

The plate Oenothera grandiflora Sims (Curtis's Bot. Mag., pi. 2068) is quite
typical of the plant grown in the New York Botanical Garden as to the shape
of the petals, but the bud-tips are much too heavy and the bracts too broad
and clasping at the base.

Large-flowered evening-primroses have appeared from time to time in the
eastern seaboard States; but none that have been examined so far can be
determined as being certainly indigenous there. In the Gray Herbarium the
following specimens are noted.

68 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

New Jersey. — Pine barrens, Miss Treat of Vineland, 1871, "wild." This has a flower
that when expanded must have measured nearly 10 cm. in diameter, the petals being 5 cm.
long and over, but the bud is stouter and heavier than O. grandiflora and has more the char-
acter of that of O. lamarckiana. It might be interesting to know whether it still survives in
the pine barrens.

Maine. — Orono. M. L. Fernald, July 17, 1892. A specimen with very much the appear-
ance of Oenothera rubrinervis, having the very characteristic bud of that derivative of O.
lamarckiana. It is probably not indigenous. Large-flowered specimens labeled Oenothera
biennis var. grandiflora and Oenothera grandiflora from California appear to be Oenothera
hookeri Torrey&Gray.

OENOTHERA SIMSIANA SERINGE.

Oenothera corymbosa Sims. Curtis's Bot. Mag., 45: pi. 1974, 1818. Not Lamarck.
Onayra spectabilis Spach. N'ouv. ann. mus. Paris, 4: 352, 1835. (Repr. 32, 1836.)
Oenothera simsiana Seringe, in DC. Prodr., 3: 47, 1828.

Seedling about 6 -weeks old. — Leaves glabrate ; blades oblong, those of the later leaves oval,
20 to 35 mm. long, 10 to 15 mm. wide, obtuse at the apex, gradually tapering into the
margined petiole, light yellow-green (fig. 72).

Seedling about 5 months
old. — Rosettes rather loose,
40 to 45 cm. in diameter,
spreading; outer leaves 15 to
20 cm. long, 4 to 6 cm. wide ;
blades oblong to obovate,
broadest somewhat above the
middle, tapering to the obtuse
apex or in the younger leaves
to an acutish apex, gradu-
ally tapering into the broad
petiole, light yellow-green,
obscurely puberulent, with
widely scattered hairs on
both surfaces (under a lens),
approximately repand-
denticulate with rather dis-
tinct shallow teeth; petiole Fi(, 72._<?ewo/A€raJtw«ana seedling about 6 weeks after germination,
white, 8 mm. or more wide.

Mature rosette. — Leaves loosely spreading, finely but obscurely pubescent all over, the
larger ones 20 to 28 cm. long, 5 to 6 cm. or more wide; blades oblong to oblong-obovate,
shallowly repand-denticulate, tapering into a very wide white-margined petiole (plate 13).

Adult plant. — Plant rather straggling in appearance, appressed-pubescent and hispidu-
lous, especially in the upper portions. Main stem about 2 m. in height, greenish-white,
becoming reddish, stout, terete, or slightly angled and sulcate above, irregularly branched
at the base, the long, rather weak, virgate branches ascending; leaves minutely pubescent
above, somewhat less so beneath, 7 to 10 cm. long, blades obscurely repand-denticulate,
oblong-lanceolate to lanceolate, acute or acuminate at the apex, abruptly narrowed into a
very short margined petiole or the uppermost sessile, coriaceous, very brittle, bright yellow-
green, soon turning red, twisted, the upper ones more markedly so; terminal rosette-like
cluster of the inflorescence symmetrical, the floral bracts divaricate and somewhat reflexed ;
bracts lanceolate, acuminate, subsessile and subcordate ; conic portions of the bud 2 cm.
long, 6 to 7 mm. in diameter at the base, canescently appressed-pubescent and hirsute,

PLATE 13.

Cc nothera simsiana. . l.leaf from mature'Vosette; 2, stem-leaf ; 3, bud; 4,
petals removed and bract; 5, capsule and bract; 6, petal.

Adult rosette of Oenothera simsiana (upper left-hand corner). Mature plant of Oenothera simsiana in
experimental grounds in New York, hearing flowers and capsules.

HELIOTYPE CO., BOSTON.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 69

tapering in terminal portion only, the erect, free tips of the sepals, 3 to 4 mm. long, mostly
unequal; hypanthium 4 to 5 cm. long, very slender, densely appressed-pubescent ; ovary
12 to 13 mm. long, slender, densely appressed-pubescent; sepals 3 to 4 cm. long, shorter than
the tubular portion of the hypanthium; petals thin, bright golden-yellow, fading saffron-
yellow, 4 cm. long, 4 to 5 cm. wide, deeply emarginate; filaments 25 mm. long, very slender;
anthers slender, 12 to 15 mm. long; pistil as long or slightly longer than the stamens; lobes
of the stigma 6 to 7 mm. long, divaricate; capsule about 3 cm. long, 6 to 7 mm. in diameter
at the widest portion, 4-angled, appressed-pubescent, rather abruptly contracted at the
apex (plate 14).

Type locality of the plants described above, along International Railroad,
City of Mexico. Collected by J. N. Rose and Jos. H. Painter, September 20,
1903. No. 7219.

Seeds of a late-flowering plant were sent to the New York Botanical Garden,
unnamed, by Dr. Rose late in 1904, and the plants from which the above
description was compiled were grown under glass for the first 6 months and
then transferred to the experimental grounds. The flowers produced at the
height of maturity are somewhat larger than those received with the seeds,
but the late summer flowers exactly resembled those from the original locality.
Remarkable for the form and color of the light green rosette and for the very
characteristic assurgent or erect habit of the slender virgate stems and the
acuminate, curved, and twisted leaves.

This species is referred to Oenothera simsiana, perhaps somewhat doubt-
fully. The plate (1974) in Botanical Magazine agrees with it, except for the
corymbose inflorescence, which may occur readily enough in the plant in its
native habitat, as many related species vary in that regard. The mature and
normal capsules do not show the strongly reflected, white valve-tips that are
seen in the illustration, but the plant as cultivated in the New York Botanical
Garden was subject to the sting of an insect that caused a malformation of the
ovary or young capsule, which then had very much the appearance of the cap-
sules in Sims's plate, though when malformed they never quite reached the
state of maturity depicted there. This is what may have occurred at Long-
leats in 1816.

0. simsiana, with the synonymy as given above, is listed in Hemsley (Biol.
Centr. Am., 1 : 454) without any indication of definite locality or collection.

An unnamed Oenothera collected at the city of Durango and vicinity, by
Dr. Edward Palmer (No. 293, 1896, in the herbarium of the New York Botani-
cal Garden) can be referred here with a fair degree of certainty.

O. simsiana is closely related to Oenothera hookeri Torrey & Gray, with
which it has doubtless been confounded in herbaria. The growing plants are,
however, very dissimilar from that species in general aspect and habit from
the rosette stage to maturity, and the flower appears to retain its yellow color
very much longer than does 0. hookeri.

70 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

OENOTHERA OAKESIANA (ROBBINS) S. WATSON.

Oenothera biennis var. oakesiana Robbins. A. Gray, Man., ed. 5, 178, 1867.
Oenothera oakesiana S. Watson. Biblio. Ind. N. Am. bot., 383, 1878.
Onagra oakesiana Britton. Mem. Tor. bot. cl., 5: 233. 1894.

Seedling about 2 months old. — Leaves obscurely puberulent, with few scattered hairs;
blades oblong or oblong-obovate, 6 to 8 mm. wide, rounded or obtuse at the apex, tapering
to a slender margined petiole (plate 15, A).

Seedling about 5 months old. — Rosette crowded, symmetrical, 7 to 8 cm. in diameter; outer
leaves 3 to 4 cm. long, 8 to 10 mm. wide; blades oblong-ovate, broadest above the middle,
obtuse at the apex, tapering to the white, margined, entire petiole, dull or pale blue-green,
rather fleshy, approximately denticulate above, glabrous, except for a few obscure hairs on
the margins.

Seedling 8 months old. — Rosette 9 to 10 cm. in diameter, very symmetrical and crowded,
raised above the ground with a few of the old leaves adhering to the rootstock; blades approx-
imately and more prominently denticulate.

Mature rosette. — Leaves more or less strigose-pubescent all over, the central ones quite
densely so, the larger ones from 15 to 20 cm. long; blades narrowly lanceolate, approxi-
mately and shallowly toothed at the acute apex, more deeply toothed at the slender, taper-
ing base (plate 15, B).

Adult plant. — Plant i to 1.5 m. high, rather slender, clothed nearly throughout with a fine,
close, strigose, appressed, and somewhat cinereous pubescence and few scattering longer
spreading hairs, becoming almost glabrate and glaucous with age. Stem light or whitish
in color, angled and channeled above, branched to the middle; leaves 10 to 15 cm. long;
blades shallowly and remotely toothed, often more deeply so at base, narrowly lanceolate,
acute and tapering at each end, sessile or nearly so, gray-green and shining above, paler
beneath, rather thick and somewhat brittle ; terminal rosette-like cluster of the inflorescence
symmetrical, the floral bracts very small, divaricately spreading; bracts narrowly lanceolate,
more deeply and regularly toothed than the leaves, acuminate at the apex, tapering to the
base, sessile, becoming 2.5 times as long as the mature capsule; conic portion of bud 10 to
13 mm. long, rather prominently 4-angled, 4 to 5 mm. in diameter at the base, the free
separate divaricate tips 5 mm. long, finely but inconspicuously appressed pubescent; hypan-
thium 2.5 mm. to 2.7 mm. long, slender, finely but sparingly pubescent; calyx-lobes half as
long as hypanthium; petals firm, 13 to 15 mrn. long, 12 to 14 mm. wide, deeply emarginate,
not opening widely; filaments 12 mm. long; anthers 7 mm. long; pistil as long or slightly
shorter than the exserted stamens, the lobes erect, 4 to 5 mm. long; capsules 3 to 3.5 mm.
or more long, 7 to 9 mm. in diameter, angled and rounded, finely appressed-pubescent,
abruptly constricted near the apex. Remarkable for its large seeds (plates 16 and 17).

Massachusetts. — Uxbridge, J. W. Robbins, August, 1878.

Rhode Island. — Providence, E. P. Bicknell, August, 1896; September, 1899.

New York. — New York Botanical Garden, D. T. MacDougal, September, 1904; Cold
Spring Harbor, Long Island, G. H. Shull, September, 1904.

The above specimens are the only ones in the herbaria of the New York
Botanical Garden and Columbia University that can with any degree of
certainty be referrred to O. oakesiana. The Robbins specimen is from the
herbarium of the late Rev. Thomas Morong and attached to its sheet is the
following note in the writing of its discoverer.

A. Seedling of Ounotheru oakusiana about t\

B. Adull rosette of Oenothera oakesiana.

HELIOTYPE CO., BOSTON.

PLATE 16.

Oenolhera oakesiana. 1, leaf from mature rosette; 2, stem-leaf; 3, capsule and bract;
4, flower with petals removed; 5, buds; 6, petal (minimum size).

Oenothera oakesiuna grown wild in waste land in New York Botanical Garden.

HELIOTYPE CO., BOSTON.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 7 1

OENOTHERA OAKESIANA mihi.

This plant was presented to Professor Gray many years ago as a well -characterized and
distinct species, which I still consider it. I discovered it in the medical colleges in New
Haven, Connecticut, in 1827. In 1831 I sowed its seed with those of the ordinary variety.
The differences are as follows:

(1) If raised from seeds sown in the spring it is annual, the common variety by its
side being biennial.

(2) The leaves of the first year, when biennial, are much narrower.

(3) The pubescence throughout is soft-appressed, that of the other being coarser and
spreading.

(4) The points of the calyx are always spreading, in the other appressed.

(5) The capsule is longer and more taper.

(6) The ripe seeds are larger.

I have found but two other localities, namely, on sides of the railroad north of Norton,
Massachusetts, and another, same situation near Apponong Depot, Rhode Island. Also, it
is well naturalized in Uxbridge, Massachusetts, from my seeds.

UXBRIDSE, sth Sept., 1878. J. W. ROBBINS.

The plants from the New York Botanical Garden were first noticed by Dr.
MacDougal late in the summer of 1904. They grew not far from the embank-
ment of the Harlem Division of the New York Central Railroad and were
quite abundant, though not covering a very widespread area. With them
were associated a number of 0. biennis, the latter in that particular spot, how-
ever, being not as abundant as elsewhere in the garden. Rosettes were
obtained of the 0. oakesiana and the plants raised from them the following
summer were in every way similar to the wild plants, and formed the basis of
these notes on the species. Plate 17 was photographed from a specimen
growing wild in the New York Botanical Garden.

OENOTHERA PARVIFLORA LINN^US.

Oenothera parviflora Linnaeus. Syst., ed. 10: 998, 1759.
Onagra parviflora Moench. Meth. Supp!., 287, 1802.

Seedling about 6 weeks old. — Rosette 2.5 to 3 cm. in diameter, rather loose and irregular,
blades of the leaves ovate or ovate-spatulate, glabrate, ciliate.

Rosette about 3 months old. — Leaves finely pubescent on both surfaces, less so above;
blades various, those of the earlier leaves narrowly oblong or oblong-spatulate, varying to
oblong-obovate, 6 to 10 cm. long, the larger ones 2.5 to 3 cm. wide, approximately and
shallowly denticulate, obtuse or nearly so at the apex, gradually narrowed to the base of the
margined petiole, which is very early streaked with red or pink.

Mature rosette.— Leaves long and slender, spreading flat on the ground ; blades lanceolate,
or oblong-lanceolate, acutish, irregularly and strongly denticulate, more deeply toothed at
the long, tapering base, dark green and shiny, sparingly mottled with red, sparingly puber-
ulent above, more so and paler beneath, 15 to 30 cm. long, 3 to 5 cm. wide; petiole white,
8 mm. wide, margined to base, distinctly pink or red ringed most of its length (plates 18
and 20).

Adult plant. — Central stem not always maturing the first season when forced. Lateral
branches of annual rosette stout and angled, pubescent with spreading hairs; leaves
resembling those of the rosette in color, texture, and pubescence; blades oblong to oblong-

72 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

lanceolate, tapering to the short, margined petiole, shallowly and approximately denticulate
above, some of them crinkled close to the midvein; terminal rosette-like cluster of the inflo-
rescence symmetrical and spreading; bracts lanceolate, acute, sessile or nearly so, divaricate,
crowded on short internodes; flowers numerous; bud club-shaped, the conic portion 8 mm.
long, 3 to 4 mm. in diameter, rarely angled, the free tips separate, quite slender, erect, 2 mm.
long or less, mostly tipped with red; hypanthium 3 to 3.5 mm. long, slender, puberulent;
ovary 8 to 9 mm. long; calyx lobes one-third as long as the hypanthium; petals firm, 8 mm.
long, 9 mm. wide, deeply emarginate, cuneate; filaments 7 mm. long; anthers 4 to 5 mm.
long, pistil shorter than the slightly exserted stamens, the lobes divaricate, 3 to 4 mm. long,
capsules abundant, 2.5 cm. long, 6 mm. in diameter at the widest portion, bluntly 4-angled,
pubescent, bright green, much thinner in apical portion. (The apex of some of the valves
are occasionally more or less distinctly bifid.) (Plate 19.)
Illustration. — Meerburg, Plantas rariores. pi. 34, 1789.

Early in 1906 a package of seeds distributed as Oenothera parmflora was
received from the Botanic Garden of Madrid, Spain. The plants raised from
these seeds differed essentially from any of the various so-called 0. biennis of
European origin in cultivation in the experimental grounds of the New York
Botanical Garden. The plants were very persistently biennial, and only one
or two of those placed in the open threw out lateral branches, the main branch
only starting its growth too late in the autumn to mature. Specimens of any
Oenothera closely resembling it are not to be found in the herbarium of the
garden or in that of Columbia University. As 0. parmflora Linnaeus was
credited to "Canada to Virginia" by Pursh (Fl. Am., Sept., 261, 1814), with the
further note that it was rare, the development of the Madrid plants was fol-
lowed and recorded with the hope that eventually the species might still be
found a component part of the North American flora.

Late in the fall of 1905, Miss N. M. Stevens sent in some rosettes of what she
considered as two distinct species of Oenothera collected in the neighborhood of
South Harpswell, Maine. The rosettes were placed in the experimental
greenhouse for the winter and later Miss Stevens sent seeds of both species
from the same locality. These latter germinated in the greenhouse and
though somewhat similar in the early rosette stage, one of them, the "pre-
vailing species" of the locality from whence the seeds came, was identified
with the plant determined as 0. muricata Linnaeus and the other, the "rarer
species" growing on sandy wastes, was identical with the species grown from
the Madrid garden seeds under the name of O. parmflora, and it is from the
plants raised from Miss Stevens's seeds that this description has been drawn
(plate 20).

The species shows distinct characteristics in color and habit. The flowers
are but rarely exserted beyond the bracts, which are so crowded on short
internodes that the ends of the branches have a closely-tufted appearance.

The Linnaean description under Oenothera reads: "Parviflor. A. OE. fol.
ovate-lanceolatis planis, caule laevi subvelloso. Alar go coronans fructum, non
uti praecedentis quadrifidus, sed octofidus. "

Oenothera parvirtora grown from seeds obtained from the Botanical Garden of Madrid.

MELIOTYPE CO., BOSTON.

2 1

Oenolhera pamiflora. 1, seedling two months after germination; 2, leaf of young rosette;
3. stem-leaf; 4, bud and bract; 5, flower with petals removed; 6, flower and
bract; 7, petal; 8, capsule and full-grown bract (from seeds from Maine).

Oenothera parviflora grown from seeds of plants native near South Harpswell, Maine.

HEUOTYPE CO., BOSTON.

Oenothera muricata. 1, leaf from mature rosette; 2, stem-leaf; 3 and 4, young bracts and
buds; 5, flower with petals of maximum size; 6, petal; 7, flower with petals
removed; 8, capsule with bract; 9, capsule.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 73

The last character would appear to be an accidental one, that is, if the name
belongs at all to the plant in question here, and it is extremely doubtful if
that point can ever be definitely settled.

Some of the capsules of the Madrid plant showed a distinct bifid termination
of the valves of the capsules, but in the Maine plant that character was less dis-
tinct or altogether lacking. The Spanish plants and those from Maine, grow-
ing side by side, were not distinguishable, but the Maine plants matured more
rapidly.

OENOTHERA MURICATA LINNAEUS.

Oenothera muricata Linnaeus. Syst , ed. 12, 263, 1767.
Onagra muricata Moench. Meth., 675, 1794.

Seedling about 6 -weeks old. — Rosette crowded and a little irregular, 2 to 3 cm. in diameter,
blades of the leaves oblong, glabrate, slightly ciliate on the margins.

Seedling about 3 months old. — Rosettes crowded, 7 to 8 cm. in diameter; leaves 3 to 3.5 cm.
long, 7 to 8 mm. wide, blades oblong-spatulate, broadest above the middle, obtuse at the
apex, tapering into the margined petioles, blue-green, fleshy, glabrate or becoming appressed-
pubescent as the leaves increase in size, obscurely denticulate, the apex of the denticulations
mostly reddish.

Mature rosettes about 5 months old. — Rosetts 15 to 17 cm. in
diameter, rootstock thick raised 2 to 3 cm. above the ground,
bearing a few old leaves under the later green ones; leaves
crowded, the outer 9 to 14 cm. long, 1.5 to 2 cm. wide; blades
approximately denticulate, irregularly and more deeply so towards
the base above the broad, white-margined petiole, appressed-
pubescent or hirsute, without any trace of red in the leaves,
except at the apex of the denticulations.

Adult plant. — Plant i .5 m. high, pubescent with short appressed
as well as longer spreading hairs throughout. Stems erect, stout,

1 to 1.5 m. high or more, single or generally branched above,
angled, turning red and the outer bark splitting into shreds at
maturity; stem-leaves rather crowded, loto 15 cm. long, 1.5 to

2 cm. wide; blades narrowly lanceolate or linear-oblong, acute,
approximately denticulate, more deeply so at the base, appressed-

pubescent with short strigose hairs more or less on each surface, FlG- 73.— Seedling of Oeno-

, , thera muricala about 2

light green, paler beneath; terminal rosette-like cluster 01 the months old
inflorescence barely symmetrical; midvein broad, whitish; inflo-
rescence crowded; bracts lanceolate, acute at the apex, tapering to the sessile base,
finally two or more times as long as the mature capsule; conic portion of bud 1 1 to 12 mm.
long, 4 mm. in diameter, pubescent with appressed and spreading hairs, the free tips 4 mm.
long; hypanthium 2.5 to 3 cm. long, pubescent; calyx-segments about half as long as the
hypanthium; ovary i cm. long; petals thick, 10 to 12 mm. long, n to 14 mm. wide, deeply
emarginate, not opening widely; filaments about i cm. long; anthers 4 to 5 mm. long; pistil
shorter than stamens; stigmatic lobes erect, thickish, 3 to 4 mm. long (plate 21).

This description is based on plants raised from seeds secured from Long
Island (Bicknell, 1904, and Vail, 1904) and grown two successive years in the
New York Botanical Garden.

74 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

The Linnaean description reads as follows: "OE. fol. lanceolatis planis,
caule purpurascente muricato. Similis parviflorae, sed Fructus os non S-fidis.
Caulis puncta rubra sparsa. Canada." (L- Syst., ed. 12, 263. 1767.)

Based on this description and without reference to any later interpretation
of the species, the following specimens are placed here :

Maine. — York, Bicknell, 1896; Cape Neddick, Bicknell, 1896; South Harpswell, Stevens,

1905-

Massachusetts. — Ipswich, Morong, 1872; Nantucket, Bicknell, 1899.

New York. — Long Island, Edgemere, Bicknell, 1902 and 1904; Coney Island, Vail, 1904;
Staten Island, New Dorp, Kearney, 1894.

O. muricata L. raised from seed received from Professor De Vries from the
Holland sand-dunes resembled these American plants, but were not abso-
lutely identical. This should also be said for some plants that were raised
from seed collected by Professor De Vries near Chicago, Illinois, in 1904.
Undoubtedly 0. muricata has a wider distribution in the northeastern States
than can be noted here.

In addition to these quite a number of plants that apparently are referable
to this species are preserved in the herbaria of the New York Botanical
Garden and of Columbia University. They are as follows :

Canada. — Province of Quebec, New Carlisle, Williams & Fernald, 1902; Rimouski
County, Collins & Fernald, 1904; River Ste. Anne des Monts, Collins & Fernald, 1905;
Anticosti, Jupiter River, Macoun, 1883.

Massachusetts. — Provincetown, Hollick, 1901; Nantucket, MacDougal, 1905.

Rhode Island. — Block Island, Hollick, 1897.

The Nantucket, Provincetown, and Block Island plants are probably
stunted specimens merely, but the Canada specimens are quite remarkable
in that the mature plant still preserved the remains of the rosette of the pre-
vious year at the base of the stem and in general appearance ; even to the bent
habit of the upper portion of the stem they can be easily identified with the
plate and description of 0. muricata in Flora Danica, pi. 1757, 1823.

An attempt was made to raise some plants from seed taken from Messrs.
Collins & Fernald's specimen from River Ste. Anne des Monts, 1905, but it
was not very successful. The plants were quite persistently biennial and in no
case was a normal central stem secured, and the rosettes in the garden were
very much larger than those on the herbarium specimens. The Canada plants
are evidently a more northern form of the IvOng Island plant (plate 22).

The habit of the rosette of the previous year remaining at the base of the
flowering stem of the second year's growth is one that is also claimed for his
species O. ammophila by O. Focke (1904).

The plate of O. muricata published by Murray (Nov. comm. soc. reg. Sci.
Goett., 6: 24, pi. i, 1776) appears to have broader leaves than the American
plants enumerated here.

Ocnothera muricata grown near the northern limits of the species in Quebec.

HELIOTYPE CO., BOSTON.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 75

GENERAL DISCUSSION.

HISTORICAL CONSIDERATIONS.

With the increasing attention being enlisted in the subject of mutations in
organisms, numerous records of the saltatory activation and latency of char-
acters in lines of descent are being disclosed by workers in all branches of bio-
logical science. Some observations published by Dr. Arthur Hollick in March,
1879 (Hollick, 1879) are among the most suggestive. The views expressed by
Dr. Hollick over 26 years ago harmonize so well with the sum of informa-
tion available at the present time that it will be profitable to give them in
full in this place. He says:

The object is more to call attention to a few facts which seem to have been generally
passed over in botanical researches as devoid of interest. The whole subject arranges
itself under two heads. The first will comprise true "albinos" or such plants as have
spontaneously, in a state of nature, lost their colors and become white flowered. The second
relates to those plants in which the colors have been more or less eliminated by artificial
means.

First, then, we have to consider those "sports" of nature where there has been a sudden
change, without any intermediate steps, from a plant with colored flowers to a pure white
variety; which change,for want of a better term, we maycall "spontaneous." Such may be
aptly termed "negative" varieties, since their pecularity is due rather to an absence of their
normal color than to the presence of white.

It is nothing uncommon to see, in many species, a gradual change from a brightly colored
individual, through successive lighter and lighter ones, until a pure white is reached. This
may be very well seen in Hepatica triloba, which comprises individuals of every shade from
dark purple to white; or in Polygala sanguined, in which we find the same gradual change
from a dark red. * * *

The following list of "albinos" is made from specimens collected during the last three
years. Vernonia novaboracensis Willd., Lappa officinalis var. major, Lobelia syphilitica
Linn., Epiphegus virginiana, Verbena hastata Linn., Asclepias incarnata Linn., Trifolium
pratense Linn., and Brunella vulgaris Linn. Both Gentiana crinita Froel. and Lobelia
cardinalis Linn, have been reported to me as having produced at times albino forms, but I
have never seen them personally. * * *

Now, in the first place, not only does the flower show the characteristic absence of color
but the leaves, stem, and, in fact, the entire plant, are invariably of a lighter green; and if
any red or green should be normal to the stem (which is often the case) this will also be of a
lighter shade. * * *

Secondly, if we have under consideration a plant which commonly is known to have
juice of an acid or peculiar taste, this is generally more or less absent in the albino form, and
sometimes is eliminated entirely. Darwin has noted this fact, and, in commenting upon it,
says that honey bees evidently are aware of it, for they perforate the calyx and corolla of
the white Aconitum napellus, to get at the nectaries, but will not so do with the colored
ones. * * *

It has often been urged that these albinos are mere "sports" of Nature, with nothing
constant about them; their peculiarities due, often, to growing. in the shade, etc. In fact,
that it is a condition due to bleaching or insufficient sunlight, and that there is nothing

76 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

inherent in the constitution of the plant. Fortunately I have been able to test this. In
the case of Lobelia syphilitica I first found the plants in the shade of some rather thick
underbrush, in the month of September. This growth was cut down the following spring,
and the place opened to the full glare of the sun. This was done three years ago, yet, every
autumn since, the plants have either reappeared, or else left offspring which have inherited
their albino nature. This shows them not only constant in their peculiarities, but also that
these are bred in the plant and capable of inheritance. Epiphegus -uirginiana and Brunella
vulgaris offer the same proofs. Nor has the influence of locality much or anything to do
with it, for a plant of the Lobelia syphilitica (with white flowers), which originally grew in a
swamp, was transplanted to a dry garden a mile or more away, yet came up and blossomed
white the next year.

Let us now see what the experience gained in cultivation of white varieties can tell us.
Perhaps the Japanese have brought the art of eliminating color from plants to the greatest
perfection. Scores of species and genera have been by them variegated in the most peculiar
manner. But this is never constant in this country, but after a while always tends to
revert to the primitive color again. * * *

As there seems to be no tendency to reversion in these natural albinos, they might per-
haps be made permanent varieties and be valuable on that account. No doubt this per-
manence is due to the change being sudden, leaving no trace of color by intermediate steps,
while in cultivated examples the white has generally been obtained by a gradual selection of
less and less darkly colored ones, and hence there would be a greater tendency to reversion
back through these steps again.

From the foregoing it is clear that retrogressive white-flowered derivatives
were seen to arise from colored parental forms by saltations in which the color
of the flower behaved as a unit-character, and furthermore that the new color-
less forms were constant and were not affected by environmental conditions.
These ideas in fact underlie an important part of the mutation theory, and the
statements by Dr. Hollick constitute the unexpanded thesis of the saltatory
behavior of one set of characters. Not the least interesting part of the dis-
cussion is that which in effect recognizes the fact that many white-flowered
and white-leaved forms in cultivation represent the extremes of fluctuating
variations, and that when artificial selection lapses these forms gradually
return to a normal mean ; also that some species, such as Hepatica triloba, may
also have an extremely wide range of fluctuating variability as to the flower-
colors.

A dim realization of the importance of sports and of the necessity for some
other hypothesis besides natural selection to account for the existence of all
living plants seemed to be widespread among botanists about the time De Vries
began the investigations which have led to his conclusions as to the influence
of mutations. An interesting forecast as to the matter was made by the late
Thomas Meehan (Meehan, 1882). He said:

The conclusion I have been forced to is that the odd forms we often find in nature are not
necessarily hybrids, but are as likely, if not more likely, to be the outgrowth of some internal
law of form with which we are as yet unacquainted. That they do not often perpetuate
themselves is [not] remarkable when we remember that of thousands of seeds produced on
anyone tree but a small percentage ever gets a chance to form, and of those which do sprout,

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 77

again but a small percentage survives to become bearing trees. As the number of trees
reproducing the general features of the original may be a hundred to one of the more strik-
ingly aberrant forms, we may see that though individual instances may be common, we are
never likely to meet many trees of one stamp. Once in a while an individual tree may find
itself in a situation favorable to the preservation of a number of seedlings, which might
endure until again reproductive; in such cases a marked variety may originate and make
its way over the earth.

I have often thought it probable that in time a few individuals of these suddenly intro-
duced forms might again leap into new features, and then if they should be able to sustain
themselves we should have new species quite independently of any principle of natural
selection.

The negative implied by the context is supplied in brackets and indicated
by italics.

The above are by no means to be the earliest formulations of the hypothet-
ical conclusion as to the importance of sports in phylogeny. Among others
which might be mentioned the statement by Kerner in 1869 is worthy of
citation. He says:

Die Arten die uns gegeniibertreten, sind nur Stadien and haben als solche zwar fur eine
gewisse Zeit Konstanz, konnen sich aber friiher oder spater in andersgeformte Arten auflo-
sen. Thatsache ist es, dass alle Pflanzen friiher oder spater einmal vereinzelte Abarten,
d. i. Sprosslinge zu erzeugen in Stande sind, welche in ihren Merkmalen von der Mutterart
abweichen, und zahlreichen Erscheinungen drangen uns zu der Annahme, dass unter dem
zusammentreffen giinstiger Bedingungen solche Abarten die Ausgangspunkte neuer Arten
werden.

Among other recent observations is that of Schaffner (1906) concerning
a derivative of Verbena stricta which bore pinkish-white instead of the deep
purple flowers of the species. The aberrant form was present in many thou-
sands of individuals among the typical specimens, with which it seemed to be
in successful competition. Tests have not yet been made to determine
whether or not it does not readily cross with the parental form, or whether it
is dominant or recessive with regard to the more important differentiating
character. In any case, however, its mutative origin is based upon evidence
similar to that which zoologists seem willing to accept as proof of mutation in
animals, and its successful maintenance against parental competition is an
established fact.

Many of the observations brought forward by various authors concern
species which have been long under cultivation, and so many facts reported
from horticultural material have been found to be unguarded against errors
affecting purity of lineage that much prejudice exists with respect to data
obtained from domesticated forms. So far as the effects of actual tillage or
domestication are concerned, these prejudices are without foundation, since
at the present time no well-founded evidence exists to show that cultural con-
ditions have ever produced alterations which were strictly and continuously
inheritable under environic conditions other than those by which the altera-

7& MUTATIONS, VARIATIONS, AND RELATIONSHIPS Otf THE OENOTHERAS.

tions were induced, although vague statements and erroneous generalizations
to the contrary are current. The possibility of such results is to be by no
means denied, but if this view is to be upheld it must be supported by some
evidence based upon actual experimentation. Vicinism, the somatic multipli-
cation of bud-sports and extreme variants, and the confusion of closely related
elementary species form the basis for the greater number of positive assertions
as to the effects of cultivation, and it is necessary to examine all facts bearing
upon the lineage of supposedly new forms with the greatest care before their
aspect or behavior may be taken as evidence upon phylogenetic problems.

An extended consideration of the phenomena of mutability has led the
authors of this paper to entertain grave doubts as to the practical value of
the conception of periodicity. De Vries throughout his writings has favored
the assumption, on purely theoretical grounds, that a species must, at certain
times in its history, be in a mutable condition, while in other periods of great
length all of the progeny come true to the type, subject only to fluctuating
variability.

It is quite possible that any given form may give off derivatives steadily
for centuries, but none of these being fitted to survive, no trace of this activity
remains. A change in the character of the mutants originated, or newly
found conditions met with in comparatively short migrations, might readily
give an appearance of the beginning of a mutative period, unless the matter
had been tested by pedigree-cultures.

With regard to the actual succession of generations two aspects of the case
present themselves. In one case it may be supposed that an annual species
might have a limited distribution, and, occupying small areas, might be repre-
sented by only a few individuals every year. Combined with these circum-
stances, the proportion of mutants produced by the species might be so small
that it would be possible to examine every specimen for a hundred years
without meeting a single atypic individual. The conclusion would then be
reached that this was included in the immutable period. Then if the number
of individuals grown and examined in the next ten years numbered as many
millions, and some mutants were found, the line of reasoning followed would
lead to the fallacious assumption that a mutable period had begun, or rather
that some physiological change had ensued in consequence of which mutations
were occurring. These might as readily be seen during the first year of
observation, however, if enough individuals had been grown.

It appears, therefore, that the real state of affairs is better represented by
the phrase "frequency of mutation," by which is expressed the number of
individuals which must be grown to furnish one mutant, and which is nearly
identical with "the coefficient" of mutability. If the parental type pro-
duces more than one mutant, the frequency of these may vary widely from
each other.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OP THE OENOTHERAS. 79

Thus of the great number of mutant individuals originated by 0. lamarck-
iana but few have been rubrinervis, still a smaller number represented gigas,
while the combined observations at New York and Amsterdam have failed to
bring brevistylis within the range of the cultures, this form having arisen at
Hilversum, where it is still in existence. To secure mutants, or to allow a
species to display its possible derivatives, the long stretch of centuries may be
shortened and as many individuals examined in a seedling stage in a single
season as might naturally mature in twenty decades. The real point to be
attained is to secure sufficient individuals to include the proportionate num-
ber of all the derivatives.

ADDITIONAL EVIDENCE AS TO THE DISTRIBUTION AND OCCURRENCE
OF LAMARCK'S EVENING-PRIMROSE.

The records and material cited and examined show that a large-flowered
evening-primrose, which can be identified with nothing so closely as with
0. lamarckiana, had found its way into the gardens near Haarlem, Holland, as
early as 1756, several years before 0. grandiftora was brought from its habitat
in Alabama to England in 1778. Also that an evening-primrose with large
flowers was seen on the sand banks north of Liverpool, England, in 1806, on
the coast of Somerset in 1837, and has been cited and described by various
authors in 1845, 1851, 1855, 1860, and continuously since 1892. Plants
grown from seeds obtained from the region in question proved to beO. lamarck-
iana with two of its mutants, rubrinerms and lata. Not only did this con-
firm the supposition as to the long history of lamarckiana, but it also showed
that rubrinerms had successfully maintained itself in competition with the
parental form, while lata had become capable of self-fertilization, the only
instance on record.

THE MUTANTS AND MUTABILITY OF LAMARCK'S EVENING-PRIMROSE.

The observations begun by De Vries upon 0. lamarckiana in 1884 and
continued upon purely pedigreed material by himself until the present time,
and the cultural experiments described in this paper, demonstrated a very
high frequency of mutation in this species and also that the mutative depart-
ures show a great diversity when compared with other forms. In addition to
the derivatives which constitute the greater proportion of the atypic offspring
gigas has recently reappeared in an example in New York, being one of the
rarer forms. Then the sudden appearance of a totally new type among the
mutants leads at once to the confirmation of the suggestion that certain
mutants may appear so infrequently that they might not be expected more
than once in a million individuals and consequently would not be seen except
at long intervals.

Of the known mutants albida, oblonga, gigas, nanella, lata, and scintillans
were found in the pedigreed cultures described in this paper, and with greatly

8o MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

varying frequency. In one culture the coefficient of mutability of oblonga
was pushed slightly beyond the limit hitherto assigned to it by De Vries and
it was found to constitute 54 per cent of the atypic derivatives. A survey of
all of the results obtained in the New York Botanical Garden since 1902 does
not justify the assertion, however, that any change in the frequency of mutation
has been induced by cultural conditions, and the only really new feature
noticed consisted in the appearance of a new mutant, which, however, was so
unsuited to the climatic conditions that it perished before reaching maturity,
an example of the eliminative action of selection in the determination of what
mutants shall survive and what shall perish.

Cultures made from seeds obtained from merchants did not show a fre-
quency of the mutants above 1.5 per cent in any instance, which is much less
than that observed in the strain growing at Hilversum. The sources from
which the seeds were obtained indicate that the species is in a state of muta-
bility in various parts of the world, and that it has probably been so for some
time. The low frequency of mutation may have been due in part to unsuit-
able methods of cultivation, in consequence of which the parental individuals
were badly nourished.

O. brevistylis was brought to maturity and found to agree in all characters
with specimens grown in Amsterdam. This form differs from the parental
type chiefly by the retrogressive character of a shortened style, and by gen-
eral observation this organ shows a wider range of fluctuation about its average
than the same organ in the parent, although no exact examination was made.
Despite the fact that this type produces comparatively few seeds it was found
by De Vries among the individuals of the parental type in the original locality
in which mutation in lamarckiana was observed in 1 886, and it has since held
its own in competition with the parental type, in a manner demonstrative of
the fact that it is possible for a mutant to survive in competition with existing
forms.

O. lata grown from seeds obtained from Professor De Vries and from indi-
viduals arising in the pedigreed strains in New York was seen to remain con-
stant to its characters. A few tests had been made both in New York and
Amsterdam, but no fertilization was secured with its own pollen, although some
of it appeared structurally perfect in a microscopic examination. O. lata
appearing in a culture from seeds of 0. lamarckiana from England, however,
included some individuals which likewise produced pollen, and also showed
the hitherto unknown capacity of being fertilized by it. The seeds obtained
in this way gave rise to a progeny which showed only the constituents usually
found in a progeny of this plant when fertilized by lamarckiana.

O. scintillans grown directly from its own seeds, and also as it appeared as a
new mutant in the cultures, was constant in its ever-sporting character. Seeds
of this species purely fertilized gave rise, in addition to the type, to a few

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 8 1

individuals of oblonga and lamarckiana, in addition to an unrecognizable form,
the representative of which did not reach maturity. Purely fertilized seeds of
this form have also been seen to give rise to lata and nanella.

0. oblonga was grown directly from its own seeds and also appeared in all
of the cultures of the parental type which were spread sufficiently to include
its coefficient of mutability. It appears to be constant in its characters and
is strikingly differentiated at a very early stage of its development.

0. albida was also grown directly from its own seed and appeared in the
pedigreed cultures of lamarckiana. It appeared to coincide in all of its
characters with the descriptions given by De Vries.

The single test which was made of the effect of crossing the parental type,
lamarckiana, with one of its mutants, rubrinerms, gave results in accordance
with the conclusions of De Vries in this matter. Rubrinerms is a mutant of
supposedly low frequency of occurrence, since it has not been found in any
pedigreed culture for several years. It may be regarded as a progressive
derivative of the parental type and presumably carries all of the lamarckiana
qualities in a latent condition. On the other hand, the strain of lamarckiana
with which it was crossed, and from which it was derived several generations
back, must bear the characters which gave rise to this form as a mutant. The
two plants must therefore carry the same characters, but some are latent in
one and some in the other.

The origin of species by hybridization, which has been a well -recognized
phenomenon for half a century, receives further exemplifications in crosses
between 0. cruciata and O. lamarckiana. The fixed forms which may be pro-
duced by this cross have been cultivated in European gardens for some time
under the names of 0. cruciata, 0. cruciata varia, and 0. biennis cruciata. The
list of species which have originated by hybridization is a long one and need
not be recounted here. Singularly enough, this fact seems to be unknown to
workers who carry on extensive operations in plant breeding. In the instances
cited in this paragraph, the most remarkable feature is that a fixed form or
new species may be produced by the hybridization of two parental types of
comparatively high frequency of mutability. It seems also pertinent to again
sound a warning against the continued practice of taxonomists of ascribing
hybrid origin to species, with characters apparently intermediate between two
neighboring and known forms. Numerous hybrids of known origin in the
New York Botanical Garden have been submitted to competent taxonomists,
and in no single instance has a correct diagnosis of the parentage been made.

The fertilization of the eggs of lamarckiana from the pollen of rubrinervis
resulted in a single observation in a progeny which contained 71 per cent of
the latter, or mutant, type. This is not a universal result, however, since the
proportion has been found to vary from 19 to 74 per cent. The pollination of
0. biennis by the mutant described in this paper gave a progeny in the first

82 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

generation which conformed to the biennis type. The same result is true of
the reciprocal cross. It remains to be seen whether the mutant characters are
merely recessive in this combination, or whether a condition of latency has
been assumed in which the mutant appears only in its customary frequency.

The evening-primroses, then, afford the following exact evidence as to the
effects of intercrossing : Many of the mutants when crossed with the parental
form give progeny composed of the parental form and the mutant in the first
generation. In the case of gigas and rubrinerms the proportion of the mutant
in the progeny is generally much higher than the parental type. Rubrinerms
when grown in close contiguity to the parental form showed only a small pro-
portion of crosses, and the newly discovered mutant of 0. biennis also gave
but a small proportion of crosses when exposed to danger of hybridization
with the parent. Lastly it is to be pointed out that 0. brevistylis, which pro-
duces but few seeds, and which is recessive when crossed with 0. lamarckiana,
has been in continuous existence for 20 years, perhaps much longer, at Hil-
versum, where it is in direct competition and danger of being hybridized with
the parental form. Experiences in the guarded cultures seem therefore to be
confirmed by an analysis of natural conditions. Swamping of new forms, by
intercrossing, is a specter which looms vast and shadowy across the visions of
writers of a speculative habit, but at the present moment it is believed that no
evidence carefully tested by modern methods is available to justify the hypo th-
esis, so far as plants are concerned.

The demonstration of the greater variability of the mutants as compared
with the parent form seems to be complete, and this relation is one of consid-
erable practical and theoretical importance. The greater variability of phylo-
genetically new characters as compared with older ones is not wholly new and
unique, though we believe our studies in 1904 gave the first statistical support
to the thesis when stated in this form. Darwin's (1859, Ch. II) view that
characters which distinguish species are more variable than those that sepa-
rate genera, and that varietal are more variable than specific characteristics,
gives the first recognition of this principle, if we take into account that accord-
ing to his conception of the relation of these several form-groups the variety is
an incipient species and the species an incipient genus. Field (1898) worked
out the relation between a varietal and a specific characteristic in the wings
of a lepidopterous insect, giving a confirmation of Darwin's rule, but he did
not relate his facts in any way to the relative age of the characters studied.
It is probable of course that the varietal characteristic was a more recent
acquisition than the specific, and this would make his results accord well with
the observations of the authors on Oenothera.

To the student of evolution who wishes to institute pedigree-cultures, the
relation here demonstrated should prove of much practical value, for it matters
not whether he desires to study the possibility of fixing extreme fluctuations,

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 83

or whether he is engaged in tracing the character, frequency, and causes of
mutation a preliminary statistical investigation will aid in the choice of mate-
rial best suited to his purpose. It is obvious that the great variability of any
elementary species would disfavor it as research material for the study of
mutations rather than indicate its desirability, as the wide range of its fluctua-
tions would suggest the likelihood of its being a recent mutant rather than a
mutating species. A more promising procedure with respect to any variable
form would seem to be to use the most nearly related or contiguous species
having a low degree of variability in the hope of thus hitting upon the parental
type from which the variable form originated. The student of pedigree-
cultures frequently receives letters from the kindly disposed, calling his atten-
tion to strikingly variable forms which are, on the ground of their variability,
conceived to be especially valuable. It is to be hoped that in the future more
attention will be given to constant forms having one or more variable near
relatives.

Whether the relative variability of old and new structures could likewise be
used by the morphologist as a new criterion for working out lines of descent it
would be somewhat reckless to say at present, but an important field for
research is indicated. Are forms which are obviously more recent at the same
time more variable than the older? Are the structures characteristic of the
Orchidaceae, for instance, more variable than those of more generalized mono-
cotyledons, and those of the Composite more variable than those of other
dicotyledons? Are monocots on the whole less variable than the dicots?
While not wishing to press the application of the principle too far, it may be
pointed out that one feature of the investigation here reported seems to bear
directly upon the phylogenetic significance of the greater variability of new
characters. The characteristic feature of the Oenothera bud is the hypan-
thium. It doubtless represents one of the most fundamental mutations in the
history of the genus. It accords well with the hypothesis here offered to find
that the length of the hypanthium is in every species the most variable part
of the bud. The great excess in its variability over that of other parts of the
buds is made strikingly apparent by comparing the average variability of the
several parts of the buds in all the species studied. The average coefficients of
variability are as follows, in order of their magnitude : Length of hypanthium,
14.97; length of cone, 9.56; length of ovary, 8.66; thickness of cone, 7.85;
thickness of ovary, 6.30 ; thickness of the hypanthium, 5.83. The consistency
of these results make it certain that the great variability of the hypanthium
originated with the origin of the hypanthium.

The fact that a new species is more variable than an old one should also
give to the phytogeographer a method of investigating the exact status of an
endemic species. If it presents a case of initial endemism it should be more
variable than the nearly related, wide-ranging, and hence presumably older

84 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

species. If on the other hand it is a local relict of an old species, which once
had a wide distribution, it should present less variability than its wide ranging,
presumably newer relative. The only satisfactory evidence heretofore avail-
able for determining a point of this kind has been drawn from paleontological
sources.

To the metaphysician who is striving to picture to himself the laws of vital
motion, the results of these statistical studies will give an altogether new view
of the relation between fluctuation and mutation. Rosa (1899) has published
a treatise in which he contends that there is a progressive lessening of varia-
bility during the progress of evolution, and he sees as the certain result of this
process the ultimate extinction of life. Plate (1904) takes issue with Rosa on
this point and maintains that there is no such law. The demonstration that
there is greater variability in new species, coupled presumably with a decrease
in variability as the species grows older, indicates that both Rosa and Plate
are partly right. It is now seen that although the individual species may
decrease in variability as it grows older, this decrease is compensated for every
time a new species springs into existence. Instead of mutations being the
cumulative results of ever-increasing fluctuation, they appear now as an initial
process of which fluctuation is in part an after effect.

A detailed description of but a few of the species of Oenothera have been
given. Seeds and material have been brought in from all parts of America
east of the Rocky Mountains, and a consideration of this material gives
foundation for the conclusion that under the name biennis is included a swarm
of elementary species very closely related, but easily distinguishable when
grown side by side, having no intergrading forms.

Even the imperfect knowledge we possess of the species, owing to the
scarcity of collections made of them, shows that they overlap in distribution
in the most haphazard manner. In the same way, the elementary species
comprised within cruciata interlock. In discussing the principles that govern
distribution, it is to be admitted that in accordance with their usual practice
many systematists would not allow specific rank to the forms in question, so
closely are they related. It will be in order, therefore, to take up the next
grouping of interest, that of biennis, and the allied species of oakesiana, parvi-
flora, muricata, and cruciata, which were accepted as varieties of biennis 30
years ago, but which are accorded specific rank at the present time. It is
found that biennis blankets nearly the entire range of these forms, and also
that the ranges of these forms overlap among themselves. It is to be seen,
therefore, that no matter what range of opinion may enter into the discussion
as to the rank of the forms of the evening-primroses of the eastern part of
North America, the thesis that the most nearly related species do not overlap
in distribution is unsupported.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OP THE OENOTHERAS. 85

A condition similar to the above has been found in Crataegus by Sargent
& Peck, who say (1906) :

The peculiar tendency of Crataegus to flock together is strikingly illustrated in our terri-
tory. It is rare to find any large area occupied by a single species. Where many thorn
trees and bushes grow together there are usually many species. A remarkable example of
this kind is found in a narrow strip of pasture land bordering the Erie Canal near Menands.
Here 10 species are growing in an area of about i acre. The closest condensation of num-
erous species that I have seen anywhere is near Albia, where 9 of our native species are
growing in a kind of irregular row along the west bank of the Wyantskill Creek. The
length of the row is about 100 feet. It is worthy of remark that 3 of these species, Cratae-
gus ferentaria, C. rhombifolia, and C. succulenta, belong to the group Tomentosae. Such
close associations of members of a single group as this one at Lansingburg are very signif-
icant and when more fully understood may possibly throw some light on the interesting
problem of the development of species.

Other students of the thorns have found similar conditions with strains
which were found to be distinct and hereditary in pedigreed cultures. Here,
again, if a conservative view is upheld and specific rank is not accorded these
clearly separable elementary species, a similar association of the closely
related group species may be found. A similar condition among the opuntias
has been noted by the author. These facts make it conclusive that the
assumption that the nearly related or most nearly related species of a group
do not occupy the same habitat is untenable.

Two bud-sports of different types as to origin were found during the
summer of 1905 and tested as to their hereditary qualities.

The vegetative mutant branch of ammophila was seen to give rise to a
progeny which came true to the characters of the branch producing the seeds
from which they were grown, and the reversion, if it is such, may be consid-
ered complete. If, as suggested, ammophila is a hybrid and the sport is a rever-
sion to one of the parents, then this case may be taken to agree with that noted
by De Vries (1900, p. 86), in which a hybrid of two forms of Veronica gave off
bud-sports, which in purely fertilized seeds came wholly true to one of the
parental types. In this case the sport was a reversion to the ancestral qualities
which had become latent when the hybridization was effected.

The second bud-sport was one in which the recessive characters appeared on
a branch in the first generation of a cross, and when this was purely fertilized
showed only a strict inheritance of the recessive characters.

The facts offered by bud-sports have much significance as to the localization
of mutations, and as to the nature of the stimuli which set the mutatory pro-
cesses in action. Primary buds usually arise laterally from the tip of the
growing point and within a millimeter or two from its apex. The saltation,
of whatever character it may be, occurred in a cell or cells in the embryonic
tissue while the growing tip was moving upward a distance of a millimeter or
more. Much depends upon the place where this occurs. If it is possible for a

86 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

single cell near the apex of the growing point to undergo such changes during
the rapid division that characterizes this region that it gives rise to a mass of
cells from which the bud arises, this would favor the probability that the salta-
tory changes are due entirely to internal causes and might not be affected at all
by circumstances.

Examples of sports are known in which a section of the shoot involving
several branches are of an atypic character. De Vries offers an illustration in
his description of a green branch on a colored oak in which the greening not
only affected the entire branch, but also a section of the member from which it
arose, showing that the mutated protoplast was one which gave rise to more
than the atypic branch. In other instances the mutation may involve only a
longitudinal section of a branch or of a member or limited growth such as
inflorescence, and this may be carried to the extent that even half of a flower
may be of an atypic character, and likewise similarly complex fruits may be
produced. It is, of course, admissible that in reversionary sports, as most of
these sectional variations are, the qualities that appear in the sport are latent
in the entire plant and need but a slight stimulus to awaken them, or some
force to weaken the dominancy of the main stock. So far as the saltation here
is concerned it is one in which any disturbance of the equilibrium would in all
probability have but one result.

In the case of the appearance of characters not inherited in a latent, reces-
sive, or active condition, but which constitute a progression or acquisition, the
case is different, and if more than one protoplast is concerned in the alterations
ensuing preliminary to the organization of a bud-sport, then the suggestion
that the alterations were the result of stimulation from factors external to the
cell in which the changes ensued would have great force.

Bud-sports usually occur on the lower part of the main stem, where many
of them lie dormant in sleeping buds, and may be awakened by decapitation
of the shoot. It is thus to be seen that they belong to a comparatively early
stage in the life of the sporophyte. The mutations which give rise to atypical
seedlings, on the other hand, seem on theoretical grounds to be due to changes
ensuing in the very closing stages of the sporophyte, and previous to the quali-
tative or reducing divisions which form the egg, or in the early stages of the
pollen mother-cells. Here is encountered a feature not yet recognized in bud-
mutations — that of a fairly constant frequency.

No evidence has yet been secured to show that the pressure or direct action
of environmental factors upon the vegetative organs has produced any per-
manent inheritable changes in elementary strains of plants. The record of the
experiments in which solutions of various kinds were injected into ovaries
immediately previous to fertilization seems to point definitely to the con-
clusion that the bearers of the hereditary characters may be definitely and
directly affected by forces external to the cell, however. The alterations thus

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 87

induced consist in the suppression of some qualities of the parental form and
the substitution therefor of other characters with which the suppressed
features are mutually exclusive. The changes are expressed in anatomical
alterations and modifications of attendant functions.

Atypic derivativesjiave been secured in this manner from two species,
Raimannia odorata and Oenothera biennis, by the use of more than one reagent,
as described in the preceding text. The induced mutants have been tested to
the second and third generation, with the result that both are found constant
in the sense that they do not intergrade back to the parental form.

The injection of the solutions into the ovaries results in the total destruction
of some eggs by the mechanical action of the needle and of the solution itself,
while in a few the stimulative or other action is of such nature that the changes
are induced which result in the modification of the hereditary characters and
of the adult sporophytes which carry them, while doubtless in most of the
cases no action at all resulted. Of course in some of the unsuccessful tests all
of the ovules were destroyed and fell off before maturity. While it seems
more natural to assume that the injection results in direct action of the fluids
on the egg-cells, yet the possibility is by no means excluded that the saturation
of the tissues of the ovary with the solution results in the induced effect being
due directly to the action of the reagent upon the advancing pollen-tube and
its contents.

The wide variety of substances used as reagents in securing the alterations in
question makes it impossible to suggest the nature of the changes induced in
the reproductive elements, since the actual effect may be osmotic or chemical.
Neither may it be said whether the changes consist in disturbances of the pro-
cessions of enzymatic action, or actually modify the structure and composition
of the cytoplasm or nuclear constituents.

Mutants have thus been produced by conditions brought about in an exper-
imental way, and it can not be denied that similar action of external forces
might occur unaided by man. The radioactivity of spring and rain water, or
the effects of corrosive or other gases which are being set free in numberless
places on the earth's surface, might produce effects similar to these obtained
in the pedigreed cultures. Undue or abnormal secretions from the walls of the
ovary, from neighboring tissues, or from' intruding foreign pollen incapable
of accomplishing fertilization might exert a similar influence. The stings and
lacerations of insects and animals and the action of parasitic fungi are known
to be accompanied by the liberation of certain substances, the activity of
which results in profound changes in the somatic tissues. It is not unrea-
sonable to suppose that the effect of these substances upon the ovary might
exert equally important changes in the egg apparatus or developing ovules.
In its present state this investigation appears to be of great potential impor-

88 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

tance, since it offers the suggestion that we may become able, more or less at
will, to modify existing streams of heredity and cause new organisms to arise.
The point not to be lost sight of is that the action of external agents as
described and deemed possible is not through the somatic tissues to the egg or
germ-plasm, but is exerted directly upon the reproductive elements themselves.
To assume that such action is stimulative, as has been done by Tower in dis-
cussing the earlier announcement of these results (Tower, 1906), is to adhere to
one .of a number of allowable suppositions— one, however, which seems most
applicable to the results secured by him with beetles, but one to which we are
by no means confined in the consideration of the induction of the atypic forms
in plants. The determination of the character of this action constitutes one of
the most difficult problems in connection with the entire matter.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS. 89

RECAPITULATION.

The principal features of the foregoing paper may be stated in the following
brief generalizations:

(1) The probability of the origination of new elementary strains of plants
arising by saltations or sporting is one which has received the support of many
authors, from inferential conclusions based upon the examination of living
material. It has also been found by more than one worker that characters
acquired in a saltatory manner are not swamped by intercrossings. If such
characters behave as units they may well survive, whether dominant or reces-
sive with respect to the corresponding character of the parental stock, the prob-
abilities for survival being somewhat greater with recessive characters.

(2) Frequency of mutation is put forward as a better expression than
periodicity of mutation for the proportion of salts occurring in the course of
generations.

(3) An evening-primrose inseparable from 0. lamarckiana occurred around
Haarlem, Holland, as early as 1756, and near Liverpool, England, in 1806.
It has been noted in this and other districts at various times, and has been
under more or less continuous observation since 1892. Apparently estab-
lished and growing with it are 0. rubrinervis and 0. lata, two of its mutants.
0. brevistylis, another mutant which is recessive when crossed with 0. la-
marckiana, has been in existence in competition with the parental form in
Holland for twenty years.

(4) The coefficient of mutability of Oenothera lamarckiana has not been
increased in the cultures in America, although a form which has hitherto
escaped observation was secured. Seeds of this parental form, from various
sources outside of Amsterdam, furnished a smaller proportion of mutants than
the material furnished by Professor De Vries.

(5) O. oblonga has been found to constitute as much as 54 per cent of the
atypic derivatives of 0. lamarckiana. 0. lata, from material grown in England,
was found capable of self-fertilization and gave rise to a progeny containing
the same elements as when fertilized by the parental form.

(6) Fixed hybrids constituting species were secured in combinations of
0. lamarckiana and 0. cruciata.

(7) The greater variability of phylogenetically new characters as compared
with older ones, which was supported by statistical evidence brought forward
in a previous paper, is confirmed by the evidence presented by further studies.

(8) The forms which would appear to promise most of importance in a
pedigree-culture would be species with a low degree of variability, having
nearly related forms with a high degree of variability.

(9) The hypanthium of Oenothera, which seems to be a recent development,
is more variable than any other feature of the bud.

90 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

(10) The inference seems justifiable that individual species become less
variable as they grow older, and that this decrease is compensated for by the
fact that when a new species springs into existence, it has a greater degree of
fluctuability than its parent. Such saltations are the starting-points from
which series with decreasing fluctuation follow.

(n) O. parviflora, which has been known in Europe since 1759, has been
found in its native habitat in America in the same manner that 0. grandiflora
was traced to its original habitat. This leads to the hope that 0. lamarckiana
may be found in a wild state.

(12) The species of evening-primrose of eastern America are distributed
in such manner that the most closely related species overlap in distribution.
This conclusion is true, no matter upon what taxonomic basis the forms
are classified. Similar conditions in Crataegus and Opuntia are cited.

(13) Two bud-sports have been followed through two generations and
found to be constant. The sport in one case embodied the recessive char-
acters of a hybrid combination showing alternative inheritance, the recessive
qualities being thus activated and extracted in the first generation, coming
true thereafter.

(14) The action of reagents having an osmotic and a chemical effect has
resulted in the induction of mutants in the progeny of Raimannia odorata
and Oenothera biennis. The mutants thus induced have been tested to the
second and third generation and found to come true to their newly assumed
characters.

(15) The induction of mutants by the action of reagents is a conclusive
demonstration of the fact that hereditary characters may be altered by
external forces acting directly upon the reproductive mechanism. The
action of the reagents used experimentally is simulated by many conditions
occurring in nature.

MUTATIONS, VARIATIONS, AND RELATIONSHIPS otf THE OENOTHERAS. 91

BIBLIOGRAPHY.

1789. AITON, WILLIAM. Hortus Kewensis: or, A catalogue of the plants cultivated in the
Royal Botanic Garden at Kew, 2, 2. London, 1789.

1902. BAILEY, CHARLES. On the adventitious vegetation of the sandhills of St. Anne's-
on-the-Sea, North Lancashire. Mem. and Proc. Manchester Lit. and Phil.
Soc., 47: pt. i, 1902.

1 793. BARTRAM, WILLIAM. Travels through North and South Carolina, Georgia, East and
West Florida, the Cherokee country, the extensive territories of the Musco-
gulges or Creek Confederacy, and the country of the Chactaws. Dublin,
1793. (Reprinted from the Philadelphia edition of 1791.)

1839. BAXTER, W. British phanerogamous botany, 4: pi. 257, 1839.

1859. DARWIN, C. Origin of species.

1900. DE VRIES, H. Das Spaltungsgesetz der Bastarde. Ber. d. deut. bot. Gesell., 18:

83-90, 1900.

1901. DE VRIES, H. Die Mutationstheorie. Versuche und Beobachtungen ueber die

Entstehung von Arten im Pflanzenreich. Erster Band. Die Entstehung der
Arten durch Mutation. 648 pp., 8 pis. Leipzig, 1901.

1902. DE VRIES, H. Die Mutationstheorie. Versuche und Beobachtungen ueber die

Entstehung der Arten im Pflanzenreich. Zweiter Band. Die Bastardirung.
Erste Lieferung. pp. 1-240. Leipzig, 1902.

1903. DE VRIES, H. Die Mutationstheorie. Versuche und Beobachtungen ueber die

Entstehung der Arten im Pflanzenreich. Zweiter Band. Elementare Bastard-

lehre. pp. 241-751, 4 pis. Leipzig, 1903.
1905. DE VRIES, H. Species and varieties: Their origin by mutation. Edited by

D. T. MacDougal. xvm -f- 847 pp. Chicago and London, 1905.
1905. DE VRIES, H. Ueber die Dauer der Mutationsperiode bei Oenothera lamarckiana.

Ber. d. deut. bot. Gesell., 23: 382-387, 1905.
1898. FiELD, W. L. W. A contribution to the study of individual variation in the \vings

of Lepidoptera. Proc. A. A. A. S., 33: 389-396, 1898.

1904. FOCKE, O. Oenothera ammophila. Abh. Nat. Ver. Bremen, 18: 182-186. 190^.
1907. GATES, R. R. Pollen development iu hybrids of Oenothera lata x O. lamarckiana,

and its relations to mutation. Bot. Gaz., 43: 81-115,07.

1902. GREEN, C. T. The flora of the Liverpool District, p. 58, 1902.
1845. HALL, T. B. Flora of Liverpool, 37, 1845.

1879. HOLLICK, ARTHUR. A few notes on the abnormal absence of color in plants. Bull.
Torr. Bot. Club, 6: 293-296, 1879.

1869. KERNER. Die Abhangigkeit der Pflanzengestalt von Klima und Boden. Inns-
bruck, 1869.

1903. MACDOUGAL, D. T. Mutations in plants. American Naturalist, 37: 737-770,

November, 1903.

1905. MACDOUGAL, D. T. Discontinuous variation and the origin of species. Science,

21: 20-23, I9°5- Also Torreya, 5: Jan. 1-5, 1905.

1905. MACDOUGAL, VAIL, SHULL, and SMALL. Mutants and hybrids of the Oenotheras.
Publication 24, Carnegie Institution of Washington, 1905.

1905. MACDOUGAL, D. T. The evolution of plants by mutation. Lecture before the

Barnard Botanical Club, New York, December 18, 1905. Chicago, 1905.

1906. MACDOUGAL, D. T. Discontinuous variation in pedigree-cultures. Pop. Sci. M'thly,

September, 1906.

92 MUTATIONS, VARIATIONS, AND RELATIONSHIPS OF THE OENOTHERAS.

1882. MEEHAN, T. Hybrid oaks. Bull. Torr. Bot. Club, 9: 55, 1882.

1 760. MILLER. Figures of the most useful and uncommon plants described in the Garden-
er's Dictionary exhibited on three hundred copper plates accurately engraven
after drawings taken from nature, vol. u, 1760.

1904. PLATE, I/. Giebt esein Gesetz der progressiven Reduktion derVariabilitat? Arch.

f. Rassen u. Gesells. Biol., 1: 641-655, 1904.
1895. POHL, J. Ueber Variationsweite der Oenothera lamarckiana. Oesterr. Bot. Zeitschr.,

45; 166-169, 205-212, 1895.
1899. ROSA, D. La riduzione progressiva della variabilita, pp. 135. Torino: Carlo

Clausen, 1899. (French Abstract, Arch. Ital. de Biol., 33: 314 et seq., 1900.

German translation by Dr. H. Bosshard.)
1903. ROSA, D. Die progressive Reduktion der Variabilitat und ihre Beziehungen zum

Aussterben und zur Entstehung der Arten, pp. 105. Jena: Gustav Fischer,

1903.

1905. RosE, J. N. Studies of Mexican and Central American plants. 4. Contributions

U. S. Nat. Herb., 8: 330, 1905.

1906. SARGENT, C. S., and PECK, C. H. Species of Crataegus found within 20 miles of

Albany. N. Y. State Museum Bulletin No. 105. Report of the State Botanist,

1905. 1906.
1906. SCHAFFNER, J. H. A successful mutant of Verbena without external isolation.

Ohio Naturalist, 2: 31, 1906.

1806. SMITH, J. E. English botany, pi. 1534, 1806.
1906. TOWER, W. L. An investigation of the evolution of the chrysomelid beetles of the

genus Leptinotarsa. Publication 48, Carnegie Institution of Washington, 1906.

UNIVERSITY OF CALIFORNIA LIBRARY

Los Angeles
This book is DUE on the last date stamped below.

891

i L9-50»i-ll,'50 (2554)444

UNIVERSITY of CALIFOBMU

AT

LOS ANGELES
LIBRARY

UC SOUTHERN REGIONAL LIBRARY FACILITY