THE HISTORY OF THE INVESTIGATION 7 



seeds of Guilandina bonducella was unintentionally left over- 

 night in the pan of the balance. Out of curiosity I noticed its 

 weight and found that it was 2 per cent., heavier than when 

 originally placed in the oven. In other words, a sample of 

 100 grains, which had been reduced in weight by exposure to a clue to a 

 a temperature of 100° C. to 92 grains, on the following morning P„^^iJ.y^°'^ 

 weighed 102 grains. This result was startling, and quite a 

 new road of investigation was opened up, occupying much of 

 my time for three years and supplying materials for several 

 of the chapters of this work. Under the stimulus of this 

 discovery I made at the time a few speculative comments in 

 my notebook, which proved to be the starting point of a theory 

 of cosmic adaptation to which Chapter XX is devoted. Here 

 again the seeds of Guilandina bonducella have been a source 

 of inspiration, and I soon got to realise that I owed much to 

 these interesting seeds. 



But in the meanwhile my horizon had been greatly Thewiden- 

 extended. With the old theory gone it was evident that I hofizo^of 

 could no longer treat the seed independently of the fruit, and ■ *e inquiry, 

 that I could no longer ignore the facts that the seed had 

 coverings, that the embryo in the resting seed was in all stages 

 of development, and that the reserve of food within the seed 

 presented great variations in amount as well as in disposition. 

 The investigation promised to branch out in a multitude of 

 ways, provokingly divergent in their direction. However, I 

 continued the method of following indications and was soon 

 hard at work again with the balance and the oven. 



Since the seed-coats had played a variety of parts in the 

 experiments, one of the first of the new inquiries begun was 

 concerned with the seed-coat relation, meaning thereby its The seed- 

 relative weight as part of the entire seed. Great variety in coat relation, 

 this respect soon displayed itself. At the same time I began 

 to compare the seed-coat relation of the resting seed with that 

 of the soft, unripe seed and of the seed swollen for germina- 

 tion. Now commenced the separate treatment of the coats 

 and kernel in the oven experiments, when I was surprised to 



