296 Group Discussion 



use an arith./log scale, arith. for time and log for survival, so that 

 constant mortality gives a straight line. 



Tanner: I have been sitting here for the last half-hour with a very 

 strong feeling of dejd vu. The people who are interested in growth 

 have been fitting growth curves with decreasing enthusiasm for about 

 35 or 40 years. It seems to me that you are pursuing a vertiginous 

 and descending pathway ! 



I do not think there is anything in the general aspects of growth 

 which leads one to suppose that the allometric relationship is very 

 useful in general. There may be instances where it is necessary, not 

 from any theoretical considerations but because using logs produces 

 a straight line; I do not think that there can be any other justifica- 

 tion. 



Sacher: A transformation, such as the logarithmic, cannot increase 

 the amount of information contained in a set of data, so if the cor- 

 relation of the transformed variables is • 99 + this expresses a fact 

 about the data, i.e. that only a fraction of a percentage of the total 

 variance is error variance when the proper functional relation be- 

 tween the variables is found. No a priori justification is needed for 

 the use of the power function. In my own and Brody's data there 

 is no question about its appropriateness. The fact that some other 

 data are adequately rectified by a linear plot is interesting but it 

 cannot contravene the allometric relations as they have been estab- 

 lished in many other cases. 



Chitty: My particular problem is to find out why animal popu- 

 lations in nature do not go on increasing indefinitely, and what it is 

 that they die of. ^lost people up to the present have considered that 

 deaths in nature could be almost entirely accounted for through 

 predation or epidemic disease, heavy infestation with parasites, or 

 food shortage, but it is now clear that this is a wholly inadequate 

 explanation, particularly for the huge mortalities which occur in the 

 young stages. The problem arises — what exactly is it that they die 

 of? The suggestion was first made by P. H. Leslie and R. M. Ranson 

 in 1940 (J. Anim. EcoL, 9, 27), for the field mouse, that the life-table 

 type of explanation might be applied to field populations. In the 

 laboratory you recognize that, with age, there is an increasing prob- 

 ability of death from a variety of causes which are peculiar to the 

 particular environments — that is to say a group of mice in one 

 laboratory would not have the same final causes of death as they 

 would in another — but in each case there would be the common fact 

 that as they grew older they became increasingly liable to die of 

 whatever it was that was peculiar to those environments. The 

 field evidence strongly suggests that this may be a profitable way of 



