Genie Control of Development 345 



the mutant locus is smaller than normal, but the Bar dosage series 

 permits a more direct approach to genie action. It can be established 

 that the inhibition of facet number by the diflFerent Bar dosages follows 

 definite laws. Another Bar allele, infrabar, has an ejffect intermediate 

 between + and B and also has a dosage effect. We are not concerned 

 here with the problem of what infrabar is in terms of duphcation, 

 position effect, or point mutation, since we know it does not affect the 

 dosage problem. Therefore we include infrabar in the dosage series, 

 though it is known to have a completely different reaction upon 

 temperature changes (Luce, 1926; complete references in Gold- 

 schmidt, 1938; Steinberg, 1941; Goldschmidt, 1945Z?; DeMarinis, 1952). 

 My original approach to the dosage problem (1927) was this. If the 

 facet-forming reaction (whatever this means) is expressed as a series 

 of curves for the different dosages, and the actual number of facets 

 in the different combinations is plotted on a line parallel to the ab- 

 scissa, representing the threshold at which the result is decided, the 

 heterozygotes should have curves with an angle intermediate between 

 those of the homozygotes, if simple proportionality to dosage obtains. 

 Actually, the curves plotted from Sturtevant's data of facet number at 

 25° showed that some exponential relation must be involved. This 

 I assumed to be the number n of cell divisions of the anlagen cells 

 which produce 2^ primordial cells and subsequent facets. Thus the 

 exponent n was calculated from the actual data (e.g., 9.6 for normal 

 eyes). As the number of primary cells decreases with dosage of B, 

 the potency of action of each combination would be inversely pro- 

 portional to the number of cell divisions, or, in terms of an inhibiting 

 effect, would be directly proportional to it. Calculating in terms of 

 potency of action, we can equate the relative dosage involved to the 

 number of cell divisions needed to reach the actual facet number, in 

 relation to a normal point assumed to be ten divisions. Since n = 9.6 

 for + eyes the calculated dosage (or potency effect) is 10/9.6. Normal 

 is, of course, homozygous +. Therefore the relative potency (dose) of 

 one locus is 10/9.6 X 1/2. A table can be compiled for this ideal dose 

 in all the combinations counted by Sturtevant, adding the individual 

 doses for each locus, and this calculated dosage of each combination 

 can be compared with the effect actually found in terms of total 

 dosage, that is, 10/n. Table 3 shows the results. 



With few exceptions (9, 13) the order of the two columns is 

 identical. The table shows the simple proportionality between dosage 

 and effect as well as the action of the alleles upon the number of 

 primary cell divisions in the eye anlage. 



