Genie Control of Development 383 



c. Dosage via chromosomes 



As I mentioned earlier (III 5 C c), the least reliable dosage 

 experiments are those which change dosage by means of trisomy, 

 monosomy, and so on, or by polyploidy. In all these cases, different 

 doses of a mutant locus are obtained, but not on one and the same 

 genetical background. This might be irrelevant in some cases, for 

 example, when the tiny fourth chromosome of Drosophila is involved, 

 or when the mutant action is so direct that there is little if any inter- 

 action with different chains of reaction. The latter might be true, 

 especially when dosage effects upon simple chemical end products 

 are involved. Clear dosage effects in such experiments would lead to 

 the same conclusions as those drawn in the last chapter. The best 

 examples might be expected in polyploids because here the parallel 

 dosage changes for the entire genome might provide a genetic back- 

 ground which is not very different; though this is not necessarily so, 

 as the variable effects of polyploidy upon cell size and physiology of 

 plants demonstrate. The least reliable results are expected for number 

 changes of single chromosomes which may change the entire develop- 

 mental system considerably. 



In view of all this it is surprising that simple dosage effects can 

 sometimes be observed in such cases. Actually, Correns (1900) had 

 already noticed intensification of mutant action in the endosperm of 

 maize (the endosperm is triploid). The best example of this kind was 

 found by Mangelsdorf and Fraps (1931), who studied vitamin A 

 content in corn, using the triploid endosperm. Yellow has a higher 

 content than white, and triploid yellow has the highest. The case is 

 remarkable, because here, simultaneously, dosage by chromosome 

 number increase, by mutation, and by heterozygosis could be checked 

 in the same experiment. If Y = yellow and y = white, the number of 

 Y loci will be in a triploid endosperm 0, 1, 2, 3 in the compositions 

 yyy, yyY, yYY, YYY. The resulting effect on vitamin A units per gram 

 was in the same order, 0.05, 2.25. 5.00, 7.50, an astonishingly exact 

 dosage relation. Obviously, the single genie action here is of the simple 

 type expressed in the one gene — one synthetic step idea, which would 

 permit such exact dosage results in a triploid without interference 

 from other reactions. The same is true when it is found that a homo- 

 zygote produces twice as much of an enzyme as a heterozygote does. 

 But there are similar examples in which morphological characters are 

 involved, some of which are reviewed in my book of 1938. Here, also, 

 other examples of such dosage relations concerning chemical products 



