( 18 ) 



true in normal surroundings. The generation also included a 

 single male specimen of the form tortuosa, which crossed with 

 a female decemlineata yielded hybrids with decemlineata 

 dominant. These in succeeding generations split up in 

 Mendelian fashion, the tortuosa (extracted recessives) breeding 

 true. 



Experiment 4. — The subject of this experiment was the 

 tropical species Leptinotarsa multltaeniata. Here, again, the 

 parent beetles were exposed during the first period of egg- 

 laying to artificial conditions of heat and humidity, while 

 during the second period the surroundings were normal. The 

 results obtained were analogous with the preceding. The 

 second batch produced nothing but multltaeniata for four 

 generations, when the experiment ceased. The first batch gave 

 together with typical multltaeniata a much larger number of 

 the form melanothorax. The two forms were separated, and 

 each continued to breed true under normal conditions. 



Experiment 5. — In this experiment, the subject of which 

 was again L. multltaeniata, the egg-laying period was divided 

 into three instead of two portions. The first and third were 

 passed under normal conditions, the second under the artificial 

 conditions of the last experiment. The first and third 

 batches of eggs gave nothing but typical multltaeniata ; the 

 second consisted entirely of the forms melanothorax and 

 rubicunda, each of which continued to breed true. 



Experiment 6. — Finally, a similar experiment was tried 

 with another tropical species, Leptinotarsa undecimlineata. 

 Again the result was the production by artificial means of a 

 varietal form (anguslovittata), which continued to give pure 

 descendants showing no tendency towards reversion. 



After these experiments there seems no possible room to 

 doubt that the germ-plasm is accessible to external influences, 

 and that it may by these means be transformed in such a way 

 as to give rise to a permanent race of descendants showing 

 marked differences from the parent form. 



I will not now stop to discuss Tower's interpretation of the 

 results of these and his numerous other experiments. His 

 theoretical views seem to me to differ in value, some being 

 weighty in the extreme, and a few, to say the least, question- 



