sect, xi.] INTRODUCTION. 65 



these facts give no support to the view that the exactness or 

 perfection with which the proportions of the normal \m\i\ are 

 approached is a consequence of Selection. It appears rather, that 

 there are two possible conditions, the one with five joints and the 

 other with four, either being a position of Organic Stability. Int.* 

 either of these the tarsus may fall; and though it is still conceivable 

 that the final choice between these two may have been made by 

 Selection, yet it cannot be supposed that the accuracy and com- 

 pleteness with which either condition is assumed is the work of 

 Selection, for the "sport" is as definite as the normal. 



This interesting case of Meristic Variation in the tarsus of the 

 Cockroach illustrates in a striking way the principle which is 

 perhaps the chief of those to which the Study of Variation at the 

 outset introduces us. We are presented with the phenomenon of 

 an organ existing in two very different states, between which no 

 intermediate has been seen. Each of these states is definite and 

 in a sense perfect and complete ; for the oscillations of the four- 

 jointed form around its mean condition are not more erratic than 

 those of the normal form. Now when it is remembered that just 

 such a four-jointed condition of the tarsus is known as a normal 

 character of many insects and especially of some Orthoptera, it is, 

 I think, difficult to avoid the conclusion that if the four-jointed 

 groups are descended from the five-jointed, the Variation by which 

 this condition arose in them was of the same nature as that seen 

 as an individual Variation in Blatta ; that as the modern pheno- 

 menon of the individual Variation which we see, so thai pasl 

 phenomenon of the birth of a four-jointed race, was definite and 

 complete, and that the change whose history is gone, like the 

 change to be seen to-day, was no gradual process, but a Discon- 

 tinuous and total Variation 1 . 



1 Since this Section was written it has seemed possible that the account Riven 

 above may be found to need an important modification. It is well known that 

 Blatta, in common with many other Orthoptera, has the power of reproducing the 

 antennae and legs after amputation or injury, and we have made 80me observations 

 shewing that the tarsi of these regenerated legs sometimes, if not always, contain 

 four joints. The question therefore arises whether the 4-jointed tarsus la a truly 

 congenital variation, and not rather a variation introduced in the process of 

 regeneration, somewhat after the manner of a bad- variation. To determine this 

 point a considerable number of immature specimens were examined, and it was 

 found that the percentage of individuals with 4-jointed tarsi is considerably Less 

 in the young than in the adult. These facts lend support to the view that the 

 4-jointed condition is not congenital. A quantity of individuals were also hatched 

 from the egg-cocoons, and among them there has thus far been found no ease of 

 4-jointed tarsus. On the other hand the total number thus hatched is not yet 

 sufficient to create any strong probability that none are ever batched in the 

 4-jointed state. We have also seen the 4-jointed tarsus in three very young in- 

 dividuals, which, to judge from their total length, must have been newly hatched. 

 The statistics shew besides that the abnormality is distinctly commoner p females 

 than in males, and that it is commoner in the legs of the 2nd pair than in the 1st, 

 and much more common in the 3rd pair of legs than in the 2nd. These facts some- 

 what favour the view that the variation may be congenital. It seems also ex- 

 ceedingly improbable that in the specimen with all the tarsi 4-jointed. the six legs 

 could each have been lost and renewed. There seems on the whole to be a pre- 



B. 5 



