320 MESSES. C. SHEARER, W. DE MORGAN, AND II. M. FTTCIIS 



as in the liybrid E. acutus % X E. mdiaris ^ , otherwise the mitoses are as regular as 

 in the normal eggs of the female parent. 



(b) E. miliaris ^ X E. esculentus ^ (1912). — Only 2 per cent, of the eggs from this 

 cross were found to be fertilized. The mitotic figures were exactly similar to those 

 described above for the hybrid E. miliaris ^ X E. acutus $ , there being an 

 elimination of one or two chromosomes at the first segmentation division, but no 

 vesicles were ever formed. 



On Baltzer's theory these re.sults would doubtless be explained somewhat as 

 follows : — 



1. In the hybrid E. esculentus ^ X E. miliaris $ the paternal chromosomes remain 

 latent as in Eclmuis $ X Antedon S , and so the larvae exhibit maternal characters. 



2. In the hybrid E. acutus ^ X E. miliaris S the chromosomes which are 

 eliminated are not those which are concerned with the characters of larvse which are 

 being used as a criterion of inheritance, while an application of his "incompatibility 

 theory " would necessitate the assumption that the eliminated chromosomes are those 

 derived from the male parent. 



3. In the hybrid E. miliaris ^ X E. acutus $ or E. miliaris '^ X E. escidentus $ 

 it is necessary to suppose that eggs which develop normally without elimination, and 

 those in which elimination occurs, all give rise to the same form of larvse. As in the 

 cross E. acutus ^ X E. miliaris $ , the eliminated chromosomes are not connected with 

 the larval characters. 



It will be observed that all the explanations are entirely negative, each is based 

 upon an assumption for which at present there is no prospect of proof; in fact, it 

 must be admitted that these hybrids afford us no evidence in favour of the correlation 

 of definite larval characters with certain of the chromosomes. 



Again, Baltzer's explanation of the Antedon hybrid, and the application of his 

 principle to explain the hybrid E. miliaris ^ X E. acutus $ and E. miliaris $ 

 X E. esculentus $ , nullifies his suggestions concerning the eliminated chromosomes in 

 the hybrid Strongylocentrotus % X Sphcerechinus $ . For if paternal chromosomes 

 can behave apparently normally in a hybrid and yet be so influenced by the foreign 

 cytoplasm as to be "latent " — as in Echinus^ X Antedon $ — then it is impossible 

 to decide whether those chromosomes which are not eliminated from the hybrid 

 Strongylocentrotus $ X Sph(xrechinus $ contain factors for larval inheritance but 

 which are " latent " ; or whether these factor's are located in those chromosomes 

 which are eliminated. It is obvious that this question can only be settled by 

 obtaining the Fg generation of the hybrids, for if Baltzer's explanation is correct, 

 then the skeletal characters, etc. , of this generation and all others will be exactly the 

 same as in F^. On the other hand, if the characters of Fj are due simply to dominance 

 of certain characters over others, then one-quarter of the individuals in the Fo 

 generation wiU be of the recessive paternal type. 



In connection with Baltzer's thesis that the maternal characters in the larviB are 



