492 



THE AMERICAN NATURALIST [Vol. LII 



withstanding the difficulty of interpreting the mutants as 

 direct recombinations, the inference can scarcely be 

 avoided that mutation in this group is conditioned by 

 heterozygosity. 



In an endeavor to obtain more definite information on 

 this point it was decided to make simple and back-crossed 

 hybrids between two varieties of Egyptian cotton and 

 to study these hybrids in comparison with line-bred 

 progenies of the parent varieties. Mutants in this group 

 are of comparatively rare occurrence, nothing analogous 

 to the "mass mutation" observed by Bartlett (191.5) in 

 Oenothera and by De Vries (1918) in Zea having been 

 observed. Very large numbers of plants of the hybrid 

 and parental stocks will therefore need to be examined 

 before we may hope to obtain reliable statistics as to the 

 production of mutants. In the meantime, it is believed 

 that what has been learned in regard to the behavior of 

 these hybrids in the first three generations is of sufficient 

 interest to warrant preliminary publication. 



Most previous studies of hybrids in the genus Gos- 

 si/pium have been made upon interspecific crosses, such 

 as Sea Island cotton (G. barbadense) X Upland cotton 

 (G. hirsutum) and Egyptian cotton 2 X Upland cotton. 

 In these cases there is very great variability in the F 2 

 and later generations. Cotton breeders have found it 

 to be practically impossible to "fix" such hybrids, even 

 after selection continued during six or seven genera- 

 tions. On the other hand, little is known of the be- 

 havior of crosses between varieties within the same 

 species. Are such hybrids less variable and less diffi- 

 cult to fix by selection, and, if so, can not stable and uni- 

 form new varieties be obtained by recombination? It 

 is believed that these questions are partly answered by 

 the data presented in this paper. 



The investigation was conducted at the Cooperative 

 Testing Garden, Sacaton, Arizona, which is conducted 



bri/origl^BalL 1912°' "T* COnimonl - v opposed to be of hy- 



d?cL n a d nitudf erenCeS fr0m ^ m6riCan Upland cott « n °"e c^rtahdy of spe- 



