GENETICS: W. E. CASTLE 
505 
With reference to the reconstruction for D. virilis, which is very 
clearly three-dimensional, the authors have nothing whatever to say. 
It would not be known that they had seen the paper^ in which that 
reconstruction is described except for a footnote at the very end of 
their paper, which reads ''Castle concluded that the white-forked 
value of 45.7 used by him, is somewhat too high, which is 'true; but 
there were available to him more than 40,000 flies (Bridges, Genetics, 
1, 1916 and Weinstein, Ibid, 3, 1918) giving the lower value, in con- 
trast to the less than 4000 flies on which the high value was based." 
There seems to be here a veiled rebuke that I did not look outside 
Table 65 of Morgan and Bridges for additional data. I saw no reason 
for doing, so since Morgan and Bridges had based their linear diagram 
exclusively on ''the data summarized in Table 65," and it seemed to 
me fair to test an alternative hypothesis by the same data. More- 
over I was not at that time aware that additional cases had been 
recorded bearing on the white-forked linkage relation. In comparing 
the linkage system of D. ampelophila with that of D. virilis, I mentioned 
quite incidentally the longer distance recorded for the former species 
but that my model indicated this value to be probably too high, 
"since the wire joining white with forked in the model is too long to 
harmonize fully with other linkage values given by Morgan and Bridges, 
the wire being curved." If my model, merely by a curved wire, was 
able to detect an error which would otherwise have been overlooked 
by me, and which it required 40,000 additional flies to prove, the 
reconstruction method certainly has utility. Now I have made certain 
predictions as to what some undetermined linkage relationships in 
D. virilis will be found to be, based on the reconstruction figured in 
my former paper. An experimental test of their correctness is in- 
vited. Let us taste this pudding to prove or disprove its edibility. 
With reference to point (3), Morgan and his associates consider 
cross-over values greater than 50 possible, though not as yet observed. 
They are led to this conclusion by the consideration that values 
greater than 50 are obtained by summation of short distances, on 
the hypothesis of linear arrangement. Forced to a choice between 
non-linear arrangement and map distances greater than 50, they 
choose the latter, and then contend that distances greater than 50 
are possible and would actually be observed except for double cross- 
ing-over. For my part, I can not conceive of a mechanism which 
would tie two genes together in such a way that they will subsequently 
separate from each other oftener than they will remain together, yet 
this is what the idea of cross-overs in excess of 50 per cent amounts to. 
