EXTERNAL FACTORS ON THE SPAWN DATE 



135 



Ml, Temperature and Rainfall. In Fig. 36 a much more compli- 

 cated state of affairs is presented. The same variables arc used, but for 

 the month before spawning, Mt. It is one of the most important. 

 Starting at the bottom of the diagram, at a temperature of — 2-5°C, 

 we fmd, as might be expected, some late dates up to 85°C. It may 

 come as a surprise to those living in these islands that a mean monthly 



ISOPHENES 

 0^90 80 70 60 50 



c 



N YEAR-DAYS 

 40 30 



100 200 300 



RAINFALL IN M ILLI METRES 



Fig. 36. Joint Functional Regression Diagram, Showing the 



Relation between Rainfall and Temperature and Spawn Date 



for the month Mi 



Note: (i) a trough of earliness at about o°C, (2) a ridge of lateness at about 



3''-4°C, (3) signs of another ridge of lateness above 7°C, (4) the latest dates 



occur at opposite ends of the diagram, when it is either very cold or very 



hot for the time of the year. 



temperature of — 2*5°C has been recorded, but here is to be found 

 the immense advantage of working with a large scheme such as tliis. 

 The data for these low temperatures come from very cold winters in 

 parts of the country that have cold climates anyway, and a single 

 observer has no chance of collecting observations such as these, as 

 well as the other extremes that are needed for a complete survey. 



Passing now to slightly warmer conditions, we reach the freezing 

 point, o°C, and here encounter what was perhaps the most dramatic 

 event in the whole investigation. (It would be quite wrong to suppose 



