A STUDY OE RECIPROCAL CROSSES. 97 



these 200 pistils 18 well formed fruits were obtained; of the 

 18 fruits only 6 produced seeds with embryos, and from these 6 

 only 5 contained fertile seeds. Germinative tests of these seeds 

 showed on the average that 30 per cent were viable. These facts 

 emphasize the importance of a practical study of the obscure 

 problems concerning plant development as opposed to generali- 

 zations from a study of the minute structure of the organs. 



From the foregoing results the following statements seem 

 reasonable : 



I. Reciprocal crosses with crookneck and golden custard 

 squashes are not absolutely impossible. 



II. Golden custard pollen varies in its degree of potency and 

 shows all possible stages, ranging from perfect fertility to per- 

 fect sterility. These stages are shown in the decay of crookneck 

 pistils, when pollinated with golden custard; in microscopical 

 sections of such pistils ; in the partial and complete development 

 of fruit and seed ; and in the sterility and fertility of seed. 



GENERAL SUMMARY. 



From the foregoing investigations the following conclusions 

 seem justifiable : 



Failure in making reciprocal crosses may be explained by 

 some incompatibility in the structure of the flowers. Such 

 barriers may result from delicate structures, and constitute 

 mechanical difficulties. 



Failure of pollen to produce fecundation may be explained by 

 the incomplete development of pollen tubes, or, in cases of com- 

 plete development, by the non-fusion of nuclei. 



Failure in the fusion of male and female nuclei is indicative 

 of disturbance of the sexual organs. Such disturbances may be 

 chemical, but are probably constitutional in nature. 



