THE EGGS AND YOUNG TADPOLES f 



points on which it is based is a risky procedure. If the line is curved 

 or has some comphcated shape, then extrapolation is quite unjustified, 

 for no one knows what twists and turns the line really takes beyond 

 the points that have been determined. Mathematics can often help to 

 decide the validity of experiments or observations, but is of no use in 

 deciding relevance, a matter that depends on the exercise of judgment 

 in the light of all available facts. Great caution is therefore always 

 needed before accepting that the results of experimental zoology or 

 physiology really apply to ecology, for however important they may 

 be in themselves, as additions to scientific knowledge, they may be 

 quite irrelevant to ecology, if they are merely accurate descriptions of 

 events that never happen in nature. The test of relevance is to try and 

 find something in the field that corroborates the laboratory results. 

 It need not be as extensive evidence as that provided in the laboratory 

 but the two sets of observations must fit. If there is no fit, then the 

 laboratory results are not applicable to the field. It is never the other 

 way round, £or, provided a field observation is correctly reported, and 

 is therefore vahd, it must be relevant, for no process o£ extrapolation 

 is involved — it is direct evidence. These remarks are inserted here not 

 because they apply especially to the work of the authors who have 

 studied this special subject, but simply because this is Chapter i, and 

 a suitable place to emphasize a general principle. 



To return now to the effect of lethal temperatures on distribution, 

 no one seems to have reported the death of eggs and embryos in the 

 field under circumstances that could be attributed to high temperatures. 

 Dickerson (1906) stated that in Texas, tadpoles and water frogs are 

 often killed in large numbers in the shallow pools, and considered that 

 death occurs at 40°C. This is not quite the evidence that is being sought, 

 but it shows the kind of tiling that is wanted. For R. temporaria there 

 is a httle evidence, not enough to establish the case, but at least sug- 

 gestive. The spring of 1949 was exceptionally warm in south-eastern 

 England, and in a pond on the borders of Middlesex and Hertfordshire, 

 some spawn reached a temperature of 2i'5°C. This is perilously near 

 Douglas's figure of 24°C, or Moore's of 25X, but not quite there. 

 The spawn in two other ponds in the same area on the same day did 

 not nearly reach this temperature, but was in each case at about 1 5°C. 

 When values show a range as wide as this, it often indicates that even 

 the extreme value actually observed would be exceeded if more 

 observations had been available, and it is rather likely that somewhere 



