EXPERIMENTAL EQUIPMENT AND PROCEDURES 



circuiTiscribe potentialities. In these qualitative experiments large nunnbers of con- 

 trols are not necessary, but similarly placed transplants of other donor material, or 

 varied locations within the host area of the same anlage', would constitute adequate 

 controls. The function of the control is to prove positively that the results obtained 

 could not have been obtained with any other set of circumstances. 



THE NUMBER. AGE. SEX. ETC. OF ANIMALS USED : The results of many experi- 

 ments are the more convincing the greater the nunnbers of experimentals and controls 

 used, within certain limits. If every experimental animal shows a particular res- 

 ponse never seen among the controls, then 10-20 controls and experimentals should 

 suffice. However, if the differences between experimentals and controls are (calcu- 

 lated to be) small, then, on a purely statistical basis, larger numbers are necessary. 

 In the latter instance the results are primarily quantitative, even though they may be 

 concerned with qualitative differences. Thus, it is important to plan on a sufficient 

 number of animals all used at one time. There are, for instance, seasonal differ- 

 ences even within a species, and also diurnal differences (cycles) in certain rhyth- 

 mic functions, so that in order to obtain large numbers it is not safe simply to ac- 

 cumulate data from small numbers over an extended period. 



The purpose of any experiment is generally to study the effect of a single vari- 

 able, e.g. , temperature. It is inaportant therefore that all other physical and 

 biological variables be eliminated between the experimental and controls, or at least 

 be reduced and recognized. One variable among animals that is not always recog- 

 nized is sex , though the differences are usually obvious! There are few biological 

 experiments in which sex does not play a part, though sometimes a negative one. In 

 general the female is the more hardy, the more resistant to trauma, etc. Another 

 variable that must be recognized is age. Certainly the embryos of different ages are 

 very different, so that one cannot put together data from temperature observations on 

 the blastula, gastrula, and neurula stages of the frog tadpole, for instance. Another 

 area of variation is in physiological activity, the more difficult to recognize. Thus, 

 one must try to obtain a large number of identical animals or embryos; half of which 

 may then be subject to the one experimental variable in order to determine the valid 

 biological response. 



THE SINGLE EXPERIMENTAL VARIABLE IMPOSED : The above statements empha- 

 size the necessity of preparing the animals or embryos so that the results obtained are 

 clearly due to the single variable imposed. If one is experimenting on temperature 

 effects he must be absolutely certain that no other variable is involved. Aside from 

 the few biological variables above mentioned, among the physical variables would be 

 light, agitation, salinity, number of organisms per unit of space (concentration), etc. 

 any one of which could easily becloud the results of the intended variable of temper- 

 ature. Both the biological and the physical and chemical variables must be eliminated 

 or so controlled that there remains but a single variable between the control and the 

 experimental animals. 



METHOD OF COLLECTING DATA : This will depend entirely upon the material, and 

 the nature of the experiment. In any case it should be systematic and complete, so 

 that anyone else could step in and complete the experiment at any point of progress. 

 Dates should appear by every set of data collected, and every measurable or detect- 

 able bit of information should be recorded in a book which never leaves the laboratory. 

 The practice of collecting data on scraps of paper, later to be transposed to the rec- 

 ord book, is to be discouraged. 



Sketches, accompanied by photographs, are very useful in qualitative experiments. 

 But in every case the illustration must be fully labelled . No one can remember ac- 

 curately such minute data for any length of time. When the data are all available it 

 is well to reduce it to a table and then, if possible, to a graph. 



