6 ECOLOGY AND LIFE HISTORY OF THE COMMON FROG 



ponds, as will be seen from the following account of the jelly mass in R. 

 temporaria (Savage, 1950). 



The jelly mass is not an amorphous lump, in which are embedded the 

 vitelli. On the contrary, it possesses a structure, which may be likened 

 to a bunch of grapes glued together only where they touch, and having 

 wide channels (the intercapsular channels) between each egg. Indian 

 ink dropped on a freely floating cluster falls right through in a very 

 short time, and small swimming animals can be seen under the micro- 

 scope to emerge from the depths. The diffusion of gases so essential 

 for respiration does not therefore take place over distances of, say, 

 five or ten centimetres but only over the semi-diameter of the envelope, 

 a matter of a few milhmetres. The structure cannot act like this unless 

 the clump is floating or at least supported under water on some 

 irregular surface, because the jelly is so weak that, if it has to support 

 its own weight, it collapses and obliterates the channels. It is for this 

 reason that large clumps of spawn brought into the laboratory often 

 develop irregularly. The clumps are commonly placed in only enough 

 water to cover them, in a dish with an impervious surface, and the 

 lower ends of the channels are then blocked by the collapse of the 

 mass. In the field, not only are the clumps laid in a greater depth of 

 water, but there is almost always enough wind to disturb the water 

 and produce a gentle oscillation of the clumps which produces a 

 current of water through them. 



Now, in all this discussion of the probable effect of the jelly mass and 

 temperature on the distribution of frogs, there has been one feature 

 missing — confirmation from the field. All laboratory experiments that 

 are performed, not for the intrinsic interest of the facts discovered, but 

 for the hght they may throw on ecology, must satisfy the requirement 

 that besides being valid, they must also be relevant. This point comes 

 up over and over again in ecology, so that it may be as well to make 

 it clear now what is meant by this distinction. An experiment is vahd 

 if it is competently performed and correctly analysed so that it is 

 clearly significant, whether or not a statistical test is actually apphed. 

 It is, however, only evidence of what happened in the experiment, and 

 this remains true even though it is successfully repeated many times by 

 many people. The conclusions drawn from the experiment can only 

 be transferred to other conditions by a process of inference that 

 corresponds to the graphical process of extrapolation. It is well 

 known that the extrapolation of even a straight line far beyond the 



