1914] JEFFREY— SPORE CONDITIONS 333 



in fact the least characterized by shriveled microspores of any 



species which I have yet had the opportunity to examine. It would 



seem consequently that it has either largely recovered from the 



effects of previous hybridization or else that the hybridizing 



ancestors have been less incompatible than those of, for example, 

 O. biennis. 



It is obvious that in the Onagraceae we have a peculiar con- 

 dition, which is likewise present in other families of the angiosperms. 

 Taking the Rosaceae as an illustration, there are in the various 

 genera of this family numerous forms which are generally recog- 

 nized on external characters as hybrids. Some of these conditions 

 have recently been described by Brainerd. In addition to ad- 

 mitted hybrids in the Rosaceae, there are a number of forms which 

 very generally are admitted as good species on account of their 

 relative constancy and the absence of observed intergrading types. 

 In many of these recognized species, however, the morphological 

 conditions, found in connection with the formation of the repro- 

 ductive elements, are clearly those of hybrids. In view of the 

 common occurrence of admitted hybridism in the family, these 

 may properly be regarded as concealed or crypthybrids. The value 

 of internal structure in the case of hybrids, as a key to the condi- 

 tions involved, has not been sufficiently regarded in the past. It 

 is obviously of the greatest importance in the Rosaceae. A further 

 group of forms is found in the Rosaceae, where the reproductive 

 structures are those of normal species elsewhere. That is to say, 

 the pollen, the more readily investigated of the reproductive ele- 

 ments, is quite sound. This condition, for example, is character- 

 is tic of monotypic genera or genera with few well marked species 

 (for example, Kerria, Rhodotypus, Cydonia, etc.). In genera with 

 numerous species, contamination is very likely to be present, 

 except in cases where the species are very distinct, or are isolated 

 phenologically or geographically. For example, Rosa or Sorbus, 

 geographically isolated in a mountainous or insular habitat, is 

 genetically pure; while Rosa or Sorbus of continental or lowland 

 origin is very likely to be genetically impure. 



The principles cited above in the case of the Rosaceae have 

 an obvious bearing on the conditions found in the Onagraceae. 



