J. W. H. Harrison 261 



feature lies in the fact that the outermost line of all, where it intersects 

 the sixth nervure, projects along it and causes the whole band to jut out 

 at that point. Outward from the central band is a broadish white bar 

 traversed by a wavy line; this, again, is much- clearer than in any form 

 previously considered. It is succeeded by an almost solid blackish 

 suffusion in which the pale subterminal line is wholly or partially buried. 

 Within the limits of this suffusion the nervures are heavily outlined in 

 a perfect manner quite foreign to filigrammaria or autumnata or their 

 other hybrids. A similar encroachment of the black scales outlining the 

 nervures of the cell and those issuing from it is made along other nerves, 

 so that in extreme cases the whole nervure system is lined in in black. At 

 the same time a change in the relative position of the discoidal point is 

 seen. Normally it lies in the angle of the elbowed line of the outer 

 band. Here, in most examples, it lies on that line and even on others 

 beyond it, a displacement never encountered in autumnata and filigram- 

 maria. The total effect of this change in " band formula " is to throw 

 the band into a condition recalling the narrowed complete carpet band 

 observable in so many genera of the Larentiad group — even those re- 

 moved fairly widely from Oporahia — and possibly, therefore, indicating 

 that the completion of the band is reversional. 



On the hindwings the terminal bands and suffusions are obsolescent 

 but the increased strength of the black scaling outlining the nervures is 

 very marked : in particular, the black V at the base of veins 3 and 4 in 

 the fore wings, just at their origin in the cell, is reproduced as one never 

 sees it elsewhere. 



What is the cause of this postponed or pseudo-segregation ? Had it 

 occurred in such forms as Oenothera Lam,arckiana it would have been 

 heralded as a genuine example of mutation. This explanation, however, 

 is here excluded because the combinations which were used to synthesise 

 the form are known. Although two generations removed, the original 

 parents were hybrid in origin and the phenomenon, whatever it is, is 

 one of hybridity, and its explanation must lie in some germinal dis- 

 turbance brought about by that — and germinal irregularities were met 

 with early. As will appear in my subsequent paper on the gameto- 

 genesis of these hybrids the chromosome complement of filigrammaria 

 is 37 and that of autumnata 38 — these being the haploid numbers. So 

 nearly homologous are they that, in the gametogenesis of the ^i hybrids, 

 in many cases all of the filigrammaria chromosomes find mates; in 

 other cases up to four do not, with the consequence that certain oocytes 

 and spermatocytes are endowed with 38, 39, 40, 41, 42 or 43 chromo- 



17—2 



