264 Genetical Studies in Moths 



as those of the preceding cross and therefore display a very contracted 

 range of variation. Once again a single back cross has apparently 

 sufficed to restore the autumnata facies, and once again the evidence 

 gained points to permanent gametic contamination. 



I'he Back Cross between the F^ % and filigraramaria </ . 



Here a barred filigrammaria male such as was employed originally 

 was paired with an Fi female. The early stages need no special treat- 

 ment but one must not omit to state that the larva leant very distinctly 

 toward filigrammaria in the strength of its stripes. 



The imagines obtained, if they had been captured in nature, would 

 undoubtedly have been regarded as filigrammaria, but nevertheless 

 direct comparison with both autumnata and filigrammaria shows that 

 they present clearly enough points of divergence which we shall proceed 

 to point out. 



Before doing so, it must be noted that their range of variation is, 

 unlike that of the last cross, of the same extent as in the parents but 

 emphatically not greater. 



The first great difference between it and filigrammaria is the supe- 

 rior size, the mean being in both sexes slightly higher than that of the 

 ^1 insect instead of less, a circumstance brought about by the undue 

 weight of one of the superior size classes which two facts prove to be of 

 no definite import : (1) in the female one of the lower classes is similarly 

 overweighted but not enough to counterbalance the former; (2) the 

 reciprocal cross shows no such features. The only other sign of the 

 influence of autumnata worthy of special note is that, whilst every phase 

 of variation of pure filigrammaria is encountered it is always toned 

 down, no matter how small the degree, by the indecision introduced 

 from autumnata. 



To sum up, the insect comes very near to filigrammaria but shows 

 very obviously its autumnata blood and is thus a good intermediate to 

 the Fi lot axid filigrammaria. Thus we perceive that one back cross in 

 this direction does not suffice to cause a reversion to the parent type. 

 This being so, and the insect being perfectly fertile, it was once more 

 crossed back on filigrammaria, females of this stock being paired with 

 males oi filigrammaria. 



The Back Gross between the preceding and filigrammaria cT . 



The larvae of this cross were exactly the same as their cousins of 

 T^ure filigrammaria blood which were reared in the adjoining cage. And 

 the same remark is applicable to the imagines ; of all the large number 



