54 
AMERICAN POMOLOGICAL SOCIETY. 
view, and bending all efforts toward its accomplishment. The fruit or cereal 
with which the breeder is working must be thoroughly familiar to him. The 
characteristics of the varieties grown must be understood, in order that the 
most desirable improvements may be comprehended and the best means of 
obtaining them recognized. The wild species related to the plant which he 
is studying to improve must be known, so that the operator can judge whether 
or not the desirable feature exists in a wild species in a greater degree than 
in any cultivated sort. Should a wild plant show the desired character, it 
probably can be introduced into a cultivated sort by hybridization. 
The speaker emphasized the necessity of working with large numbers, 
both in selection and in hybridization experiments, insuring thereby a greater 
opportunity of securing the feature desired in the most marked degree. The 
choice of parents in hybridization and selection experiments is also of 
primary importance. If endeavoring to increase the size or yield, the in- 
dividuals showing these features in the very highest degree should be selected 
for the parents. The larger the numbr of individuals one has to select from 
the greater will be the probability of securing the desired result. 
As illustrating the desirability of having a definite aim in view and the 
general plan of work, the speaker described the experiments in orange breed- 
ing now being conducted by the Agricultural Department. The principal 
improvements here sought are, first, to obtain hardy varieties of the orange 
and lemon; second, to obtain a common sweet orange with the loose, easily 
removable skin of the Mandarin or Tangerine; third, to obtain new types of 
fruit by crossing distinct species like the orange and the pomelo; fourth, to 
improve the general quality and extend the season of ripening by crossing 
:the best varieties. 
No problem in orange and lemon culture in this country is of such impor- 
tance as the securing of hardy sorts which will not be injured by the 
occasional freezes which so frequently cause great losses in the sub-tropical 
regions of the United States and the most attention has thus been directed 
toward this end. The plan followed has been to use the Japanese Trifoliate 
orange ( Citrus trifoliata), which is deciduous and perfectly hardy as far 
north as Philadelphia, in crossing with the various varieties of the orange 
and lemon. A number of these hybrids have been produced and some of 
them have very interesting intermediate characters. Where the Trifoliate 
orange was used as the mother parent, the hybrids all have trifoliolate leaves, 
but some of them have the central leaflets much larger than in the typical 
Trifoliate orange and are apparently evergreen in habit instead of deciduous. 
Quite a number of the hybrids of this combination do not show any inter- 
mediate characters, and it seems probable that such seedlings may be devel- 
oped from some of the so called adventive embryos which are developed 
from the tissue of the mother parent, and thus could not be expected to show 
any influence of the hybridization. The fact that the common orange and 
many other species of the genus Citrus are polyembryonic is well known. 
A single seed of the common orange has been known to produce as high as 
thirteen different seedlings, although it is seldom that more than three 
of the embryos are capable of development. Strasburger in his study of the 
polyembryony of this group found that the embryos other than that developed 
from the egg cell proper, are developed from certain cells of the nucellus lying 
near the embryo sac wall, which become specialized, develop rapidly and 
form a tissue which pushes out into the embryo sac and forms an embryo 
v similar to that formed in the normal way from the egg cell. The embryos 
