132 ECOLOGY AND LIFE HISTORY OF THE COMMON FROG 



io°C, but even if this did turn out to be the case, it is probably nothing 

 more than a minor manifestation of the ''poikilothermal effect," and 

 we can dismiss temperature from the picture. With both rainfall and 

 humidity, a remarkable result attended a more detailed analysis. 



In rainfall, there was a highly significant association between the 

 rainfall and spawning, but the astonishing thing is that the direction of 

 the effect was the reverse of that expected. Frogs appeared to spawn 

 less in wet weather than random association could account for. This 

 was so surprising that the analysis was carried a stage further. When 

 the rainfall was divided into day rain and night rain, it was found that 

 when the night (during which the spawn was probably laid) had been 

 wet, but the following day (when the observation was made) had been 

 dry, the association was in the expected direction. I suggested that the 

 reason for this remarkable reversal of the effect was that the observers 

 were deterred by the rain from making their observations on a wet 

 day, but postponed them to the next dry day. At the time, I pointed 

 out that if this were the case, there should be an accumulation of 

 reports on the first dry day after a wet spell, but that there was no sign 

 of this. In view of some remarkable features of rainfall relation in 

 the spawning month, to be discussed later, I am now less sure that 

 errors accounted for the discrepancies. There is a good deal of evidence 

 to show that the phenological observers were accurate, and in this case, 

 for example, they observed on both week days and week-ends alike, 

 suggesting that most of them made their observations during their 

 ordinary duties, perhaps on the way to work. It seems unlikely that 

 such people would be deterred by the weather to such an extent. It is 

 possible that we have here a real effect, in which wet weather at the 

 time of spawning delays the event. More information is needed on this 

 point, collected by observers who undertake to observe whatever the 

 weather. There was a similar effect when humidity was examined. 

 No relationship was apparent when maximum humidity, which 

 generally occurs at night, was examined, but minimum humidity, 

 which occurs by day, gave a significant association suggesting that 

 frogs avoided days of high humidity. These are often wet days, so 

 that the humidity effect may not be independent of the rainfall effect. 



Joint Functional Regression Diagrams 



Of all the methods used to investigate this problem, these diagrams 

 were probably the most important. They take a long time to construct, 



