'20 The Rural Library. 



as variable offspring in 1891 as those selected almost at random from the 

 patch ; and, what was worse, none of them reproduced the parent or ' ' came 

 true to seed," and all further motive for inbreeding was gone. My labor, 

 therefore, amounted to nothing more than my own edification ! My experi- 

 ence in crossing pumpkins and squashes has now extended through five years, 

 and although I have obtained about one thousand types not named or de- 

 scribed, I have not yet succeeded in fixing one ! The difficulty here is an 

 aggravated one, however. The species are so exceedingly variable that all 

 the mongrel individuals may be unlike, so that there can be no crossing be- 

 tween identical stocks, and if inbreeding is attempted, it may be found that 

 the flowers will not inbreed ! And the refusal to inbreed is all the more 

 strange because the sexes are separated in different flowers upon the same 

 plant. In other words, in my experience, it is very difficult to get good seeds 

 from squashes which are fertilized by a flower upon the same vine. The 

 squashes may grow normally to full maturity, but be entirely hollow or con- 

 tain only empty seeds. In some instances the seeds may appear to be good, 

 but may refuse to grow under the best conditions. Finally, a small number 

 of flowers may give good seeds. I have many times observed this refusal of 

 squashes (Cucurbita Pepo) to inbreed It was first brought to my attention 

 through efforts to fix certain types into varieties. The figures of one season's 

 tests will sufficiently indicate the character of the problem. In 1890, 185 

 squash-flowers were carefully pollinated from flowers upon the same vine. 

 Only 22 of these produced fruit, and of these, only 7, or less than one-third, 

 bore good seeds, and in some of these the seeds were few. Now these 22 

 fruits represented as many different varieties, so that the ability to set fruit 

 with pollen from the same vine is not a peculiarity of a particular variety. 

 The records of the seeds of the 7 fruits in 1891 are as follows : 



Fruit No. 1. — 4 vines were obtained with 4 different types of fruit, 2 of 

 them being white, 1 yellow and 1 black. 



Fruit No. 2. — 23 vines : 15 types, very unlike, 12 being white and 3 

 yellow. 



Fruit No. 3. — 2 vines : 1 type of fruit which was almost like one of the 

 original parents. 



Fruit No. 4. — 32 vines : 6 types, differing chiefly in size and shape. 



Fruit No. 5. — 20 vines : 19 types, of which 10 were white, 8 orange, 1 

 striped, and all very unlike. 



Fruit No. 6. — 13 vines : 11 types, 8 yellow, 2 black, t white. 



Fruit No. 7. — 1 vine. 



These offspring were just as variable as those from flowers not inbred, 

 and no more likely, apparently, to reproduce the parent. These tests leave 

 me without any method of fixing a pronounced cross of squashes, and lead 



