a3 TRANSACTIONS OF SECTION D. 433 
dispermic eggs, in which a quadripolar spindie may result, and ultimately the egg 
segments into four cells, Lach of these cells may contain more than half the 
normal number of chromosomes, but they are irregularly distributed, and in one 
or more cells one of the chromosomes necessary for further development is 
absent, and the cell remains inactive. All theories of this kind postulate the 
individuality of the chromosomes. That this hypothesis is true is indicated by 
their constant number and by the fact that in some species (e.y., Brachystola, 
described by Sutton) they differ from one another in size and shape, and these 
differences are constant. Further, when for any reason the number becomes 
abnormal, as when an egg develops parthenogenetically, the abnormal number 
persists in the later cell-divisions (Boveri, &c ). Moenkhaus found that in two 
fishes, Fundulus and Menidia, the chromosomes differed in form, and when the egg 
of one was fertilised by sperm of the other the two kinds of chromosomes could 
be recognised in the mitoses of the embryo as long as it continued to develop. 
We may therefore take the hypothesis of chromosome individuality as probable, 
if not certain, and proceed to consider its bearings on theories of heredity. 
In most of the recent work on gametogenesis it has been established that the 
spireme thread of the primary spermatocyte splits longitudinally and then 
segments into half the number of chromosomes characteristic of the species, 
These chromosomes then split transversely, giving an equal number of tetrads. 
The two maturation divisions then separate the four quarters of each tetrad into 
the four daughter-cells which give rise to the gametes. Farmer and Moore have 
described a slightly different, but exactly equivalent, process in the vertebrates 
and higher plants, by which the original spireme segments are divided by one 
transverse and one longitudinal division among the four daughter-cells. It may 
therefore be assumed that this process is universal in gamete-formation, the differ- 
ences being only in detail. Montgomery, Gross, and others have found in certain 
insects and other forms that the spireme of the primary spermatocyte breaks up 
into the normal number of segments, instead of the half number; these split 
longitudinally, and then become paired together end to end. In this way a 
number of tetrads is formed which is half that characteristic of the somatic cells 
of the species, as in the cases referred to above. This gives us definite evidence 
that the spireme segments which appear in half the normal number in the more 
common cases are bivalent, and are equivalent to two chromosomes attached end 
to end, 
It has been found by Gross, Montgomery, and others that in some insects 
where this pairing of chromosomes takes place they differ from one another in size, 
and those two which pair together are always alike. In Brachystola the chromo- 
somes are of very various sizes and shapes, and in all the cells, from the fertilised 
egg up to the primary spermatocyte, there are two of each kind. - In the primary 
spermatocyte each pair of similar ones becomes coupled together, and also longitu- 
dinally split, and by the two maturation divisions the four quarters of the tetrad 
so formed are distributed into the four cells which develop into spermatozoa. As 
this happens with all the tetrads, it is evident that in each spermatid or ovum 
there is one chromosome of each kind, and when the spermatozoon conjugates 
with the ovum each introduces one of each kind, and so the double number 
characteristic of the zygote is produced. 
From this it appears that of the couples of chromosomes which pair in the 
primary spermatocyte, one member of each couple is derived from the father, and the 
other from the mother, and these are separated into different gametes at the 
reduction division. But since the arrangement of the pairs in the equatorial 
plate is a matter of chance, any gamete may receive a paternal chromosome from 
one pair and a maternal from another. For example, if the reduced number be 
four and the paternal chromosomes be represented by the letters A, B, C, D, the 
maternal by a, 6, c, d, then the pairs will be Aa, Bd, Ce, Dd, and the possible 
combinations in any gamete will be A, B, C, D, A 4, C, D, a, B, C, D, &c., the total 
number of combinations being sixteen. If now, using the Mendelian terminology, 
we regard the chromosomes as each bearing an allelomorphie character, exactly 
the results will follow which are found in Mendelian cases of inheritance, Using 
1905, FF 
