IN THE TROPICS. 
349 
Bateson and Saunders pointed out the probable analogy of 
the phenomenon of sex with that of Mendelian segregation. 
And Castle (9) has considered at some length the theory that 
the inheritance of sex among dioecious animals and plants 
follows Men delian principles. In the case of hermaphrodite 
animals and plants Castle believes that a mosaic inheritance 
of sex takes place ; and he discusses the bearing of a number 
of known facts upon these suppositions. 
The exact time and method of separation of the germinal 
representatives (Anlagen) of a pair of allelomorphs is another 
point of great theoretical interest. Strasburger (56) supposes 
this process to occur during the visible reducing division 
which takes place in the nuclei of the embryo-sac- and pollen- 
mother cells. And this indeed appears to be the most obvious 
view. Correns (15), however, believes that he has found 
evidence to disprove this supposition in the behaviour of the 
colour of the pollen grains in the offspring of a cross between 
two races of Epilobium angustifolium. The pollen grains do 
not show segregation in respect of colour,and Correns conse¬ 
quently believes that the separation of the “ Anlagen ” must 
take place in the pollen grain itself. 
There can be no doubt that Mendel’s theory of the separa¬ 
tion of the pure “Anlagen” of certain characters in'the 
gametes of hybrids is one of the very greatest importance, 
both theoretically and in practice. Bateson and Saunders 
would extend this principle to the formation of the gametes 
in zygotes arising by normal fertilization. In this connection 
we may recall Weismann’s view(75) that, the possible number 
of “Anlagen” contained in a gamete being limited, there 
must of necessity be a reduction in this number every time 
the gametes are formed in organisms which undergo normal 
fertilization. This is undoubtedly the case if we accept any 
theory of inheritance by definite germinal representatives. 
Mendel’s discovery not only appears to render the acceptance* 
of such a theory inevitable, but also shows that the reduction 
8(10)04 /oo\ 
