656 



SCIENCE 



[N. S. Vol. XXVI. No. 672 



numbers prove that the coupling between 

 the blue factor and the long pollen char- 

 acter is altered and becomes far more com- 

 plete. When the hood standard is segre- 

 gating from the upright standard at the 

 same time as the blue is segregating from 

 the red (viz., non-blue), and the long pol- 

 len from the round pollen, the gametic 

 series is no longer 7 blue ; long + 1 blue ; 

 round -\- 1 red ; long + 7 red ; round, but 

 is evidently 15 + 1 + 1 -f 15, unless, as is 

 still possible, the actual numbers are 16 -{■ 

 1 + 1 + 16. 



A second case of this peculiar distribu- 

 tion exists in regard to the two characters, 

 sterility of anthers and absence of colors 

 in the axil; there the association is 

 15 (or 16) fertile c?; colored axil + 1 fer- 

 tile c? ; green axil + 1 sterile J* ; colored 

 axil + 15 (or 16) sterile c^; green axil. 



The F^ numbers resulting from the re- 

 combinations of two pairs of allelomorphs 

 distribvited independently, and according 

 to various simple systems of partial gam- 

 etic coupling may be tabulated as follows. 

 In each pair one of the factors is taken to 

 be dominant over the other. 



and so forth. 



Curiously enough, we have as yet no 

 certain case of the coupling in a series of 

 8, viz., 3 + 1 + 1 + 3, though we can 

 scarcely' doubt that the system exists. 

 There are, however, clear indications that 

 couplings of a still closer order exist and 

 we may reasonably expect them to fall into 

 systems corresponding with the series of 

 powers of 2. This evidence will, in all 

 probability, be of great assistance in the 

 attempt to close in on the question of the 

 moment at which the segregation of char- 



acters is effected and must be taken into 

 account in any discussion of the nature of 

 the dichotomies themselves. It becomes 

 very difficult to suppose in these cases of 

 close though still incomplete coupling that 

 all the segregations occur at the reduction 

 division— or indeed at any single division 

 — and we await with some interest the 

 result of cytological studies of the ante- 

 cedent stages in maturation. The diffi- 

 culty reaches its maximum when we at- 

 tempt to conceive the process of character 

 distribution among the egg cells of plants. 

 The male cells in plants and animals are 

 so numerous that their numbers supply 

 sufficient scope for the formation even of 

 very long series of couplings. The egg 

 cells, on the contrary, are few, and very 

 often definitely grouped in special organs 

 which again are arranged on a definite 

 geometrical plan relatively to the gross 

 anatomy of the plant. Even if the various 

 accessory cells of the plant ovary are reck- 

 oned as belonging to the gametic series, the 

 number seems still insufficient to allow for 

 the development of a coupling which de- 

 mands a long series for its expression. Is 

 there, then, any organized system of dif- 

 ferentiation connecting the several ovaries 

 into a common plan? In maize and peas, 

 where indications of this system might be 

 expected to be found if they existed, the 

 evidence is entirely negative, and that is 

 all which can be positively asserted. 



Turning now to another aspect of the 

 problem, we have to look for facts which 

 may help us to limit our search for causes 

 of variation. We may, as I have said, 

 assume that a vast number of variations 

 are due to the addition or removal of 

 definite factors. We begin, therefore, to 

 have some dim conception of the nature 

 of this class of variations, and at all events 

 to appreciate that they must occur as defi- 

 nite and specific events. As to the causa- 

 tion of these events, there is almost no 



