Genie Control of Development 41 1 



Goldschmidt, unpublished ) . Differential growth alone would, however, 

 not explain the specificity of the transformation, and therefore Vogt is 

 forced to assume the presence of a "leg inductor" which is always 

 available and comes into action when growth changes. Vogt assumes 

 that the antennal disc always contains a leg inductor. In the antenna 

 of the wild-type fly the antennal inductor (for arista) is above the 

 threshold quantity, while in ss" the leg inductor is made to surpass the 

 threshold concentration (which could be expressed also in terms of 

 control of competition for a substrate ) . An alternative interpretation is 

 that in ss** the mutant locus produces the extra growth of the end 

 segment, as reported, which in turn results in an earlier or higher 

 reactivity to the leg inductor. 



At this point we should discuss how the different grades of 

 penetrance and expressivity fit into this scheme. Exposure to cold in 

 the temperature-effective period increases the higher, tarsus-like classes 

 of expression, which means that at this time the tissue of the end knob 

 still has the potency to react at a different rate and (or) strength to 

 the leg inductor, or to produce or use varying amounts of it. The rate 

 of growth, slowed by cold, has something to do with the effect, while 

 stoppage of growth by colchicine has the opposite effect. The later 

 the action of cold sets in, the lower the expressivity of the tarsus 

 effect; this must mean that the lability of determination is narrowed 

 down progressively during the labile period or, expressed positively, 

 that irreversible determination increases with time (labile determina- 

 tion), which again means some kind of threshold phenomenon. Since, 

 with low expressivity, the basal part of the arista shows a tarsus-like 

 structure for the longest time, while the distal part is arista-like, one 

 might think of some gradient of determination. But the distal end can 

 still be made to become tarsus-like at the end of the labile period, 

 which is not in favor of a simple gradient concept. One might say 

 there is a gradient of determination (determination stream) but not 

 of irreversible determination. (This might be considered an example 

 for Harrison's 1937 contention that there is no real irreversible deter- 

 mination. ) Finally, the fact that high classes of tarsus formation can- 

 not be induced after a certain time links the gradient of determination 

 either with a gradient of growth from base to tip, or with a super- 

 imposed gradient of more or less irreversible determination in the 

 same direction. 



This analysis shows how a mutant locus may interfere with pri- 

 mary processes of embryonic determination by changing one variable 

 in a complicated system, each cog of which is determined inde- 



