to do so until all the embryos have been determined. This occurs during the 

 adult life of the mother. In addition, there is a further element in the system 

 of control, namely the competence of the embryos to respond to the maternal 

 stimulus. The evidence suggests that the competent stage is reached when the 

 embryos are about haH-grown. 



The determination of sex in Megoura appears to involve an entirely different 

 maternal system for although it is reactive to temperature, photosensitivity 

 is absent. It is very striking, however, that even under the most favourable 

 environmental conditions (a medium temperature range) male-production 

 is always deficient, the ratio of males to females rarely exceeding 1 : 4. Нолу 

 this sex ratio is imposed by the maternal reproductive system is at present 

 obscure. We can only add that if the formation of the male parthenogenetic 

 egg involves chromosome loss, this process must be accessible to environmental 

 conditions (temperature). 



A third maternal system directs the development of virginoparae towards 

 the alate or apterous condition. In this instance neither the physical environ- 

 ment nor the nutritional condition of the plant appear to play a major role. 

 Instead, crowding experiments suggest that wing-production is a response 

 to mutual stimulation — an ,, effet de groupe" in the sense of Bonnemaison 

 (1951). This stimulus is effective during the adult life of the parent and influen- 

 ces only the largest embryos. 



It foUows that the developing eggs and embryos of Megoura are subjected 

 to a series of maternal mechanisms which can switch development into dif- 

 ferent pathways. The sex of the egg is probably determined first. Then female 

 embryos, when about haK-grown, are determined as either virginoparae or 

 oviparae. The fate of virginoparous embryos as future winged or wingless 

 individuals is decided last, just before their birth. 



I have emphasised at the outset that the different parental morphs of J/e- 

 goura vary in their capacities as form-determiners. Alate virginoparae, for 

 example, produce alate daughters sparingly or not at aU, whereas apterae 

 do so readily. And differences in degreee are apparent in every other environ- 

 mental response. These parental forms are therefore differentiated both ana- 

 tomically and physiologically. This is not so in the immediate descendants 

 of a fundatrix which are indistinquishable morphologically from virginoparae 

 taken from a longestabhshed clone; yet these individuals are unresponsive to 

 conditions which normally lead to the production of sexual forms in the next 

 generation. 



Observations on the behaviour of young clones of Megoura under short day 

 conditions have shown that the production of males and oviparae is regulated 

 by a timing mechanism which at first prevents their appearance. The most 

 remarkable property of this "interval timer" is its independence of generation 

 number. This can be demonstrated by experiments of the type originally em- 



306 



