C. Pki.lkw and p. M. Durham 165 



been kindly allowed to examine stock.s in various nurseries. We have 

 never seen true heterostyly in Kewensis ; although the styles may vary 

 in length, the anthers are always in the thrum position. The pollens 

 examined microscopically never show the difference in size which is so 

 marked a feature in hetei'ostyled primulas. 



We have also examined a very large number of plants with regard 

 to the condition uf the leaf surface, and we find on careful inspection 

 that many of the plants are really hairy. The hairs are short and 

 sparsely scattered, sometimes visible to the naked eye, but often only 

 to be seen with the aid of a lens. 



We shall refer to this tetraploid form as Keiuensis, and the sterile 

 hybrid we shall refer to as diploid Kewensis. 



Evidence from cro.ss-fertilization, and the que.stion 

 OF parthenogenesis. 



We began our experiments in 1911 by crossing verticillata with 

 jiorihunda, and in 1912 we also made many crosses between these 

 species and Kewensis. We obtained offspring which resembled the 

 maternal parent exactly, and which generally bred true on self- 

 fertilization. These results were in accordance with those obtained 

 in recent years at Kew and at Messrs Veitch's. 



We then started a series of experiments in 1913 to find out if 

 parthenogenesis actually occurs. The plants were carefully emasculated 

 and isolated. In 1914 we repeated the experiments with additional 

 precautions, isolating the plants in glass cages, which ensures isolation 

 but has the disadvantage of bad ventilation. Nevertheless we obtained 

 some positive evidence of parthenogenesis. 



One of the difficulties we have had to contend with in this work 

 has been the difference in the behaviour of the flowers of the same 

 plant on crossing. Several flowers may be pollinated by the same male 

 parent, one only may produce viable seed, the rest either giving seed 

 which fails to germinate or none at all. On account of these irregu- 

 larities, and of the uncertain germination of the seeds, it is impossible 

 to lay much stress on the details ; we shall therefore only consider the 

 general course of the results. 



There is some evidence that individual jjlants, though alike in their 

 power of giving good seed when selfed, yet differ as regards the capacity 

 for giving good seed on cross-fertilization. For instance, a single plant 



