N. Barlow 63 



a Long $ X Short ^ cross. If this is an error, therefore, it must be 

 due to foreign pollen or seed in the soil. 



In crosses with Longs as fathers, and Mid or Short mothers, 

 there is nothing conclusive to show that the non-parental type has 

 not arisen from the mother's pollen-grains reaching the stigma. But 

 in Table VIII the Mid-plants setting seed have not given rise to a 

 Short, nor have the Short-styled plants given rise to a Mid. But the 

 numbers are few, and the evidence rather negative than positive. 



Referring to Tables I and II, we see that the year 1909 gives the 

 largest number of these possible errors. In this year I have great 

 suspicion of the earth used for sowing, as in four separate cases one 

 or two more plants came up in the pots than the number of seeds 

 sown warranted '. 



But putting aside all the results of 1909, we are still left with one 

 case in 1908 and another in 1910. Whether these rare appearances 

 must be regarded as accidental introductions, or as a vital part of the 

 problem, must still remain in doubt. 



Lythrum salicaria. 

 Only the F-^ has so far been raised from this plant. 



TABLE IX. 



Long-styled female x Mid-styled male Mid-styled female x Long-styled male 



Longs Mids Shorts Longs Mids Shorts 



6 5 93 84 



Long-styled female x Short-styled male Short-styled female x Long-styled male 

 50 46 123 103 



8= VI- T- 



jttid-styled female x Short-styled male Short-styled female x Mid-styled male 



53 63 145 71 56 162 



Longs and Mids crossed together give only the two parental forms 

 in a total of 188. 



' None of these four cases coincided with a family in which the non-parental form 

 made its appearance. 



- This cross is under suspicion. The flower spikes were covered either by paper bags 

 or by muslin. The muslin was found to be an insufficient protection, as several times bees 

 were seen crawling on the surface, whilst the flowers inside were so near that it might have 

 been possible for them to receive foreign pollen from part of the bee's body. The above 

 case was from a muslin covered spike. All plants selected for further generations were 

 from paper-covered spikes. 



