Qi: ''^1

/-A^t) • o i^O

JUN 15 190
O.E..-.DS CONNECTICUT n^.

ilGRICULTURAl EXPERIMEIT STATION

NE\\r HAVEN, CONN.

BULLETIN 150, JUNE, 1905.

The Selection of Tobacco Seed Plants.

CONTENTS.

Page.

Possibility of Improving the Quality of Tobacco 3

Lack of Uniformity of Plants in the Average Tobacco Crop 5

Self-Fertilization and Cross-Fertilization of Tobacco 8

Advantages of Using Seed from Self-Fertilized Flowers 9

Method of Selecting Plants and Securing Self-Fertilization 12

The Bulletins of this Station are mailed free to citizens of Con-
necticut who apply for them, and to others as far as the editions
permit.

CONNECTICUT A&EICULTDRAL EIPERIMENT STATION.

o:f':fice]i?,s -A-dstid st.a.:f:f_

BOARD OF CONTROL.
His Excellency, Henry Roberts^ Ex officio, President.

Prof. W. O. Atwater Middletown.

Prof. W. H. Brewer, Secretary New Haven.

B. W. Collins Meriden.

T. S. Gold West Cornwall.

Edwin Hoyt New Canaan.

J. H. Webb Hamden.

E. H. JenkinSj Director and Treasurer New Haven.

STATION STAFF.

Chemists.

Analytical Laboratory.

A. L. WiNTON, Ph.D., Chemist in charge.

E. Monroe Bailey, Ph.B. I. A. Andrews, Ph.B.

Kate G. Barber, B.S.

Laboratory for the Study of Proteids.
T. B. Osborne, Ph.D., Chemist in charge. I. F. Harris, M.S.

Botanist.
G. P. Clinton, S.D.

Entomologist.
W. E. Britton, Ph.D.

Assistant to the Entomologist.
B. H. Walden, B.Agr.

Forester.
Austin F. Hawes, M.F.

Grass Gardener.
James B. Olcott^ South Manchester.

Stenographers and Clerks.

Miss V. E. Cole.
Miss L. M. Brautlecht.

In charge of Buildings and Grounds.
William Veitch.

Laboratory Helper.
Hugo Lange.

Sampling Agent.
V. L. Churchill, New Haven.

The Selection of Tobacco Seed Plants.'

By a. D. Shamel,

Of the Bureau of Plant Industry, U. S. Department of Agriculture, in

Cooperation with the Connecticut Agricultural Experiment

Station, Nezv Haven.

Every grower is familiar with the fact that no two tobacco
plants are alike. Frequentl}^ the grower finds a plant in the
field that almost or quite fulfils his ideal of a perfect plant. It
would be very desirable to produce uniform crops of the type
of these best plants, because they would be much more pro-
ductive and of better quality. As a result of the investigations
which have been carried on in the Connecticut valley and else-
where, it has been found that it is possible for every grower to
select the type of plant he desires to grow as seed plants, and
by protecting the flowxrs produced by these plants from cross-
fertilization, to secure uniform types or strains like the parent
seed plants. This important improvement can be made by the
grower in a simple and practical manner, as described in the
following article, without any extra expense and with very little
trouble or work. In view of the fact that any improvement in
the 3deld or quality of the crop secured by seed selection is pure
profit, because it costs no more to grow good plants than poor
ones, there is no farmer who can afford to grow a poor type
of tobacco when it is possible to produce improved strains.

Possibility of Improving the Quality of Tobacco.

There is no lack of either practical experience or experimental
evidence to prove that improved strains of tobacco may be pro-
duced by systematic seed selection and breeding. In the Con-
necticut valley there are many strains of the broadleaf or seed
leaf variety, which are adapted to certain market require-
ments, or local soil and climatic conditions. The broadleaf
variety is commonly believed to have been produced from
the Maryland smoking tobacco variety, the seed of which
was brought into the valley over a half -century ago. The

* For a more complete discussion see Year Book U. S. Dept. Agric. 1904.

4 CONNECTICUT EXPERIMENT STATION, BULLETIN 150.

local strains, as the Halliday, the Willow brook, and many
others, were produced by certain growers who selected a par-
ticular type of seed plant that suited their fancy for several
years in the usual manner until the type became fairly uniform
and constant. The popular Havana seed variety is reported to
have been produced from seed originally imported from Cuba,
and grown in Connecticut for a number of years in small
patches, until the variety had become adapted to the soil and
climatic conditions of this section. There are at present a
number of strains of this variety, distinguished by peculiar
shape of leaves or other characteristics, which have been secured
by many years of careful seed selection by the growers. One
of the most striking illustrations of the possibilities of seed
selection is the origin of the White Burley variety. This
variety, as is well known, was produced by a grower in southern
Ohio, who noticed in his field of Red Burley these plants having
the peculiarities of the now well-known White Burley type, and
saved them for seed. From this beginning a great industry
has been developed in southern Ohio, in Kentucky and other
states. Numberless other illustrations might be cited in this
connection, but on account of the fact that they are well known
to almost every grower, and are matters of history, it is
unnecessary to further comment on the work that has been
done, in order to prove the fact that it is possible to greatly
modify and improve this crop by an application of the laws of
plant breeding.

In the season of 1903 the writer began a series of experiments
in the Connecticut valley with tobacco in order to determine
the possibilities of improvement by the application of the laws
of plant breeding, as worked out in the cases of corn, cotton
and other farm crops, and to find the methods by which this
improvement could most rapidly be effected. It is not oppor-
tune in this short bulletin to go into details as to the results
of the experiments in that and the succeeding season, which
will be presented in a later and more extended bulletin, but the
general results will be briefly discussed here and the practical
results and methods presented for the help of the growers for
this season. This matter is not theoretical or impractical and
of scientific interest only, but it is intensely practical and can
be applied by every grower of tobacco to the advantage of his
crop, and with increased profits in his business.

IMPROVEMENT OF TOBACCO BY BREEDING. 5

Lack of Uniformity of Plants in the Average Tobacco Crop.

The extent of the variation or lack of uniformity among
plants in fields of tobacco in the Connecticut valley is best illus-
trated in the crops grown from seed originally imported from
Sumatra and Cuba, but grown in Connecticut for several years.
In a careful and systematic study of such crops by the writer,
it was found that the difference in plants was so great that it
was possible to separate out and classify several distinct types.
The differences between these types were just as marked as the
distinguishing characteristics of the broadleaf and the Havana
seed varieties. These variations were not due to local soil or
fertilizer conditions, as generally supposed, but to the seed, as
shown by the fact that they occurred in all parts of the fields
'investigated, in many cases two plants of radically different
types growing side by side under as'nearly uniform conditions
as is ever secured. The type characteristics were very marked,
and easily distinguished, even by the most casual observer.
Some of these are shown in Plate I. In one of the types
of Connecticut Sumatra the leaf was very long, narrow,
heavy, with peculiar angular veins, unusual light greenish
gray color, and of very peculiar and characteristic habit of
growth. From its fancied resemblance to the so-called Bel-
gian tobacco, it was called and is recorded as the Belgian
type. Another of the types produced nicely rounded leaves,
from eighteen to thirty on every plant, with very peculiar wavy
or crumpled edges, comparable in many ways to the crumpled
leaves of some varieties of lettuce. The leaves of this type
were very thin and in most cases lacked strength and elasticity,
so that when wrapped on cigars a considerable percentage was
likely to break or be injured in some manner. From the char-
acteristic wavy margin of the leaves, these plants of tobacco
were classified as the crumpled type. Another type found in
these crops, called the Sumatra, produced leaves that were nearly
round and were borne in a drooping position on the stalks.
This type of leaf in the field was pronounced by one of the
extensive growers as ideal for cigar wrapper purposes, so far
as the shape of leaf and other qualities that can be determined
in the field were concerned. The leaves had good body, and
extremely regular and fine veins, which stood out almost at
right angles with the mid ribs. One of the most striking types

6 CONNECTICUT EXPERIMENT STATION, BULLETIN I50.

produced very large leaves, in many cases larger than the
largest strains of the broadleaf variety, and a characteris-
tically large number of them. While the growth of the
plants was very slow in the first part of the season, they
produced ripe tobacco as soon as the early maturing types.
The rate of growth of the plants of this type in twenty-
four hours, at the middle of .the season, was nearly double
that of corn at the most favorable time for rapid growth.
Another type produced characteristically small leaves, called the
greenleaf type, because of the deep shade of green color of the
leaves. The average length of the leaves of this type was not
more than one-quarter the average length of the leaves of the
broadleaf type. The leaves were very uniform in size as well
as shape from the top to the bottom of the plants. The number
of leaves per plant was very large, in many cases plants pro-
ducing forty desirable leaves. 'This increase in number of leaves
in this type was not accompanied by a corresponding increase
in the height of the plants, but was correlated with short inter-
nodes or spaces between the leaves on the stalk.

In all of the crops grown from freshly imported seed, there
appeared a large proportion of abnormally early, small, heavy
leaved types, commonly called freaks. These plants had in
most cases the branching habit of the wild or unimproved
varieties of tobacco, and for this and other well-founded
reasons were supposed to be reversions to some of the unim-
proved varieties from which the present varieties have been
developed. These freaks were particularly noticeable in the
crops grown from freshly imported Cuban seed. In one typical
field of this kind, the writer found about thirty-three per cent,
of the plants to be of this freak type. The irregular stand can
be seen in Plate II, prepared from a photograph of this
field. Inasmuch as the leaves produced by the freak plants
are almost wholly worthless for wrapper purposes, it can readily
be seen that such crops are not profitable to the growers.
After a most careful study of this subject it can be safely
said that in no case should growers depend on freshly imported
seed for their main crops of tobacco. If it is desirable
for the grower to test foreign, imported or new seed of any
kind, it should be done on a small scale until it is determined
that the variety is adapted to local soil and climatic conditions.

IMPROVEMENT OF TOBACCO BY BREEDING. /,

or acclimated strains are produced by seed selection and
breeding.

The variation in type in the broadleaf, Havana seed and
other so-called native varieties is not so marked as in the case of
varieties grown from recently imported seed. The individual
variation or differences between different plants as regards the
number, shape, and size of leaves, the number of suckers, the
number of seed pods and other characters are nearly as
great as among plants grown from imported seed. In
every field examined some plants were found with nearly
double the number of leaves found on the average plants
in the field. Many plants produce rounded leaves, while others
bear long, narrow and pointed leaves. Some plants have a
large number of suckers, while others have comparatively few
of them. Some plants produce from one hundred and fifty to
two hundred seed pods, while others bear from twenty-five to
one hundred pods. In many instances the writer has found
plants that produce ripened leaves two weeks earlier than the
remainder of the crop. Many plants produce leaves that ripen
uniformly from the top to the bottom of the plants, while others
produce bottom leaves and lower leaves that ripen from ten days
to two weeks before the upper or top leaves are ready to harvest.
Some plants produce leaves with large and coarse veins, while
others in the same field bear leaves with fine and small veins.
Individual plants were found in all fields bearing leaves having
more desirable body, stretch or elasticity, color or appearance,
and other characteristics, than the other plants in the same field.
In fact the tobacco crops present the greatest amount of vari-
ation in type and individual characteristics of any farm crop
known to the writer. The large number of grades of color,
sizes, and qualities of leaves, which the growers find necessary
to assort after the tobacco is cured or fermented, is definite
evidence of the lack of uniformity in this crop. This lack of
uniformity in the crop is detrimental to its value, not only
because some of the grades are poor and of very little value,
but also on account of the cost of assorting these variations into
their respective grades.

Increase in the uniformity of the tobacco crop means
increased yield and a more valuable crop. Inasmuch as it

8 CONNECTICUT EXPERIMENT STATION, BULLETIN I50.

costs as much, if not more, to grow a crop lacking in uni-
formity, as to grow a uniform crop of desirable plants, this
increase in value is pure profit to the grower. The methods
of seed selection and breeding presented in this bulletin have
been found both experimentally and in farm practice to pro-
duce more uniform crops of tobacco, with no extra expense to
the grower.

Self -Fertilization and Cross-Fertilisation of Tobacco.

In order to present the methods of seed selection and breed-
ing, it is necessary to refer to the tobacco flowers and the
natural method of fertilization in tobacco plants. The appear-
ance and structure of the flower may be seen in Plates
III and VI. The tobacco flower consists of a brightly col-
ored tube-like corolla, enclosing five, more or less, stamens,
i. e., slender filam.ents bearing at their ends plump bodies,
called anthers, in shape somewhat resembling a cigar. When
the flower is just about to open, these anthers contain a large
amount of fine dust-like particles, the pollen grains, or the male
fertilizing element of the flower. At the base of the flower is
a small pod-like receptacle, which later forms the seed pod and
which bears a long, slender, rod-like projection, having a knob-
like swelling at the tip, the whole being called the pistil. This
knob-like portion is the receptive part of the female portion,
or stigma of the flower, on which the pollen falls which ferti-
lizes the flower. In the early stage of the development of the
flower, this stigma projects above the cluster of anthers, but
later the anthers gradually grow past it, at the same time
opening and discharging the pollen so that it falls on and
fertilizes the stigma. At this time the stigma will be found to
be covered with a sticky substance to which the pollen adheres.
Gradually the anthers push out of the opened flower and pro-
ject beyond the stigma. The different stages in the develop-
ment of the. flower and seed are shown in Plate VI. At the
base of the tube-like corolla will be found a secretion of
sweetish honey-like substance, which can be shaken out of the
flowers and falls like drops of rain on the plant. Many species
of bees and other small insects crawl down into the opened
flowers and feed upon this honey-like substance. The writer has
observed humming birds feeding on this honey, similar to the

IMPROVEMENT OF TOBACCO BY BREEDING. 9

manner in which they visit the fiov^ers of the common honey-
suckle. In passing to and from these flowers, they naturally
carry the pollen from flower to flower, and from plant to plant
and rub it off on the stigmas, and in this way effect "cross-
fertilization" among the plants in the field. Cross-fertilization
must be very general among tobacco seed plants in the field
under the ordinary conditions.

A flower is "self-fertilized" when the seeds are produced by
the impregnation or fertilization of its pistil by pollen from its
own stamens. It is "cross-fertilized" when seeds are produced'
by impregnation with pollen from another plant. In some
species of plants cross-fertilization is absolutely essential to seed
production. In other species it produces more and better seed
than self-fertilization. But tobacco belongs to still another
class of plants which are abundantly self-fertile and in which
self-fertilization is apparently more effective for seed pro-
duction than cross-fertilization.

Charles Darwin found in greenhouse experiments that "in
six trials with crossed and self-fertilized" (tobacco) "plants,
belonging to three successive generations, in one alone did the
crossed show any marked superiority in height over the self-
fertilized ; in four of the trials they were approximately equal ;
and in one (i. e. in the first generation) the self-fertilized plants
were greatly superior to the crossed. In no case ■ did the cap-
sules from flowers fertilized with pollen from a distinct plant
yield many more and sometimes they yielded much fewer seeds
than the capsules from self-fertilized flowers."

Advantages of Using Seed from Self -Fertilised Flozvers.

In the last two years, extensive experiments have been made
in the Connecticut valley on the relative value of seed produced
by exclusive self-fertilization and of seed produced by natural
cross- fertilization. These experiments have been made with
the Connecticut Havana, Broadleaf, Sumatra and Cuban types
of tobacco, under cloth and in the open, and on a commercial
scale; in all cases with like results. These results we consider
are of the very greatest importance to growers and are here
very briefly summarized.

Seed produced by exclusive self-fertilization for the two
generations covered by our observations has been lighter in

lO CONNECTICUT EXPERIMENT STATION, BULLETIN I50,

color, heavier, freer from mold or fungous disease, and in all
ways more valuable, as will appear in what follows.

The plants grown from self-fertilized seed reproduced exactly
the character of the mother plant from which they came. If
the parent had large leaves, so did all the progeny. If the
parent had small leaves, so did all the progeny; and in both
cases the average size was the same as the average size of the
parent's leaves. This uniformity and exact reproduction of
the characters of the mother plant are well shown in Plate IV.
Figure A shows two rows of plants from the self-fertilized
seed of a single mother plant. The uniformity in size of the
plants as well in the shape, size, veining and number of leaves
is very striking. The adjoining rows of tobacco are of a differ-
ent type. Figure B, on the same plate, shows a like uniformity
in two rows of plants from seed of a single mother plant of the
so-called Belgian type.

Plants from a parent with few suckers had correspondingly
few suckers. The shape of the leaves of the offspring was
closely like the shape of the parent's leaves.

The same correspondence appeared in the number of leaves.
If the parent had thirty-five leaves, the offspring averaged
about thirty-five leaves. If the parent bore ten, the offspring
averaged ten.

In a word, the individual characters, such as shape and color
of leaves, numbers of leaves and suckers, body or texture, size
of veins, time of maturity, and all other observed characters
were transmitted from the parent seed plants to their offspring
with marvelous uniformity.

Two particular experiments emphasize this accurate trans-
mission of individual characters by the seed of self-fertilized
flowers.

In a field of Connecticut Sumatra, several hundred parent
plants of different types were selected and the seed saved
under bag; hence it was exclusively from self-fertilization.
The progeny of these plants were set out in separate rows in
the same field, under as uniform conditions as it is possible
to secure, the following season. One of these types produced
leaves which would not burn, even with all of the tests that
were applied. Growing side by side with this type, in several
parts of the field, was another type, which, tested in the same

IMPROVEMENT OF TOBACCO BY BREEDING. II

ways, had the most perfect burn of any kind of tobacco grown
in any part of the world. Here were two types, from the same
original lot of seed, the parents grown in the same field, the
progeny grown under as uniform condition as is ever secured
in the field, the tobacco cured and fermented absolutely alike,
one of which would not burn, and the other burned perfectly.

One more instance is presented because of its specially inter-
esting nature and importance. A field was set out with plants
grown from imported seed, which were attacked by a fungous
root disease, and all died with the exception of a few
plants. These resistant or immune plants were found irregu-
larly over the field, and produced ripe tobacco of excellent
quality. All the other plants were completely destroyed with
the exception of one or two semi-resistant plants that produced
a large amount of seed, but very few and extremely small
leaves. The seed was saved separately from the resistant
and semi-resistant plants, and sowed in separate sections of
the seed beds. The resistant seed produced perfectly resistant
plants, both in the seed bed, and in the field where the plants
were destroyed the previous year. These are shown in Plate
V. Most of the seedlings from the semi-resistant seed died
in the seed bed, but enough were finally secured to set out one
or two rows in the field. These plants grew slowly, some died,
and none reached maturity, all having the characteristics of
the diseased plants in root, stalk and leaves. Some of the
resistant seed was sowed on the seed beds where the diseased
seedlings had been destroyed by the disease, and this immune
seed produced perfectly resistant plants under these circum-
stances.

These extensive observations and experiments show that the
transmitting power, as defined by Dr. H. J. Webber, or the
"prepotency" of the seed of tobacco from self-fertilized flowers
is wonderful, and we believe of very great practical importance.
With an understanding of its value, with keen observation of
the characters of individual tobacco plants which it is desirable
to retain and reproduce, and with very little labor, the grower
can modify, change and improve his strain of tobacco and can
produce crops which are uniform in all respects and like the
parent plants which he selects for self-fertilized seed production.

12 CONNECTICUT EXPERIMENT STATION, BULLETIN I50.

Method of Selecting Plants and Securing Self -Fertilisation.

We may now briefly describe how this is to be done. In the
first place, the field from which the selections of seed plants are
to be made should be gone over carefully, before the plants are
topped, and the plants which come nearest the ideal of the
grower should be selected for seed plants. This selection can
be made at any time before topping, during the cultivation of
the field as well as by special visits for this purpose, and during
the topping process. The grower should, at these times, carry
a handful of tags in his pocket and tie one on each of the
plants which suits his fancy. No good selections are ever made,
or permanent progress accomplished, by hit or miss and careless
methods. The seed heads of the selected seed plants must be
covered before any of the flowers open in order to prevent any
possible cross-fertilization. If, by accident, some flowers have
opened before the seed heads are bagged, they should be care-
fully picked off and thrown away. It is a good plan to pick off
all of the top leaves down to where the plant would have been
topped if it had not been saved for seed. All of the large lower
seed branches should be broken off, and only the central cluster
of seed-bearing branches left for seed purposes. As a rule,
each mature pod contains from three to seven thousand seed,
so it can be readily seen that a few pods will produce a large
amount of seed. Of course, the grower should save many
times the amount of seed that he expects to use.

The common manila twelve-pound grocery bag, procurable
at any good grocery, hardware, or general store, is admirably
suited for this purpose. The bag with a roof-shaped bottom
is better than the square-bottom shape because it sheds the rain
better. The bag should be carefully fitted over the seed head,
and the open end tied around the stalk below the seed head, not
so tightly as to injure the plant, or so loosely as to allow the
bag to be slipped or blown off by the wind, or allow the visits
of insects which might crawl up the stem inside the bag. The
adjustment of the bag is shown in Plate VII.

As the plant grows in height, the bag should be moved up
the plant so as to accommodate the extra growth and prevent
the seed head from pressing against the bottom of the bag, which
might injure or break off the flowers and seed pods. Late in
the season, after all or most of the pods have set, the bags

IMPROVEMENT OF TOBACCO BY BREEDING. 1 3

should be opened and all of the loose flowers and other debris
shaken out, after which it should be immediately retied and
allowed to remain in this condition until the pods have turned
brown, indicating maturity. Some growers advocate picking
off the leaves from the seed plants as they ripen, while others
allow them to remain until the seed heads are harvested. If
the suckers are carefully kept removed, it will be perfectly safe
to allow the leaves to remain on the seed plants. On the other
hand, if it is desirable to save the leaves, they can be picked off
as they ripen without injury to the seed.

We would advise those who wish to make the most careful
selection and who have the facilities for it, to pick the leaves
from the separate seed plants, when they are ripe, cure them as
usual in the barn with the rest of the crop, keeping the leaves
of each plant by themselves, suitably labeled. If opportunity
offers, let them be fermented with other tobacco during the
winter. They can then be judged quite fairly as to burn, colors
and texture, and the seed of the very best of them saved for
the following crops.

When the pods are mature, the seed heads, bags and all,
should be cut off, and hung up in a dry, airy place, where
they can thoroughly dry out with a free circulation of air.
After drying it is a good plan to keep them where the tempera-
ture does not fall much below zero. 'In the spring, when the
time comes for the preparation of the seed for sowing, the
largest and best pods should be picked off by hand, and thor-
oughly threshed out to remove all seed. The seed should then
be separated with the seed separator in the way described in
Bulletin 149 of this Station.

PLATE I.

Differences in characters of leaves from plants of the same variety of tobacco.

PLATE IL

Tobacco grown in Connecticut from Seed just imported from Cuba, showing
freak plants and irregular stand.

PLATE III.

A Flower-Head of Tobacco read}' to cover. The opened flowers show position
of anthers and stiemas at time of fertilization.

PLATE IV.

Fig. A. Uniformit}- in Tobacco Plants from Self-fertilized Seed. The two central
rows of one strain ; adjoining rows frotn a different strain of same variety.

Fig. B. Uniformity in Tobacco Plants from Self-fertilized Seed of Belgian type.

PLATE V.

Resistant and semi-resistant Plants growing on infected Soil.

PLATE VL

Different stages in the Development of the Tobacco Flower and Seed.

PLATE VII.

Flower Head protected from foreign pollen ("cross-fertilization"), by a

manila bag.

I

University of
Connecticut

Libraries

39153028850925