3l6 WILLIAM LAWRENCE TOWER. 



scribed by De Vries in Oenothera Lamar ckiana, may well be the 

 variability which follows complex processes of hybridization. 



L. decemlineata X L. oblongata X L. multitceniata. 

 Exp. No. C. H. 156.4 Escamela. 



In 1906 operations were begun at Orizaba, and in May the 

 same three species from the same original stocks were mated. 

 Conditions at Orizaba are decidedly different from those in the 

 Balsas Valley. The city is 2,000 feet higher in altitude and 

 the climate is very different. In the Balsas Valley during the 

 summer the days are bright and hot, with evening showers. At 

 Orizaba, in the location chosen at the foot of the Sierra Escamela, 

 it is never above 90° even on the hottest days, and the nights are 

 always cool, owing to the downward draught of cool air from the 

 mountains which flows over the valley at night. The relative 

 humidity is high at all times, and the precipitation during the 

 season was 74 inches. 



Under these conditions the crosses which were made thrived 

 as far as certain members were concerned; the L. muUitceniata 

 individuals were decidedly reduced by the conditions under 

 which they were living and the L. oblongata individuals were 

 hampered considerably, but to a lesser degree. Crossing was 

 observed, however, among the component species in all directions, 

 and progeny emerged in July, showing a combination to have 

 been formed between L. oblongata and L. decemlineata, with the 

 L. muUitceniata type and attributes practically wanting. The 

 population, when examined, showed individuals which were ap- 

 parently dominated by L. decemlineata (A) to the exclusion (as 

 far as visible) of all others; individuals which were very clearly 

 intermediate between L. decemlineata and L. oblongata (B) ; 

 and individuals which were more or less intermediate between 

 L. decemlineata and L. muUitceniata (C). Of these the inter- 

 mediate between L. decemlineata and L. oblongata existed in by 

 far the greatest numbers, as shown by the following proportion: 



131 I 397 92 



