290 MESSRS. C. SHEARER, W. DE MORGAN, AND H. M. l<nTCHS 



skeleton in the postoral arms, when they have become unhealthy ; this abnormal 

 condition being accompanied' with considerable stunting of the larva. We might be 

 easily lead astray, therefore, in crossing Echinus with Sph(sr echinus, by the appearance 

 of this lattice skeleton in the offspring, in tliinking it represented the influence of one 

 of the parents, whereas, although normally absent, it sometimes appears in the other 

 parent, under unusual conditions. 



As the resulting hybrids of a cross are more frequently inclined to be unhealthy, 

 and their mortality is always much higher than the pure-bred forms, it is clear that 

 results obtained from the investigation of skeletal characters can only be accepted 

 when it has been shown that they rigorously fulfil the two following conditions : — 



1. The variation of the characters must never overlap in the pure parent 



species ; it must be first determined clearly whether they show any 

 convergent variation under normal or pathological conditions. This rules 

 out the use of a " lattice skeleton " in crosses into which Echinus enters, 

 for in the members of this genus a lattice skeleton sometimes appears in 

 the pure forms, although only in a small percentage of cases. 



2. The hybrid larvge should be equally as healthy as the pure-bred ones. 



Provided these two conditions are always fulfilled, there is no reason why the 

 evidence of the skeleton should not be accepted. The difficulty is that in the past no 

 investigators seem to have observed these conditions satisfactorily. In recent work 

 Tennent (90) may have observed i, but not 2 ; Loeb, King, and MooRE (54) do not 

 seem to have observed either. 



These last authors, working at Pacific Grove, California, made reciprocal crosses 

 between Strongylocentrotus franciscanus and S. purpuratus. They came to the 

 conclusion that the larval skeleton could not be treated as a whole, but that it is 

 composed of a number of factors. These behave as allelomorphic pairs, one character 

 of the pair being dominant over the other in both reciprocals of a cross, independently 

 of which parent it comes from. Thus in their crosses they consider the larvae of both 

 reciprocals to develop the dominant members of the follov\ing pairs of allelomorphs : — 



Clubbed ends of skeleton rods are dominant over arched ends. The round dome- 

 shaped form of larva is dominant over the pyramidal form. The early development 

 of the pluteus arms is dominant over the late development of the arms. The rough 

 spinous condition of the skeleton is dominant over the smooth condition. A well- 

 developed " Mittelstab" is dominant over a rudimentary one, and an " Aufsteigender 

 Ast" is dominant to the absence of the same. 



A careful examination of their figures, it would seem to us, bears out ordy the 

 following conditions : — 



In the cross *S'. purpuratus $ X »S. franciscanus $ , their figures 24-35, representing 

 six-day larvae, seem to show that the clubbed condition of the skeleton is dominant 

 over the arched form, as they claim, but that tlie round form of the larvse is not 



