AFFINITIES AND GENEALOGY. 187 



supplanted and disappear as soon as they have given birth 

 to other and more perfect groups. The latter, though 

 victorious over their predecessors, may not have become 

 better adapted for all places in the economy of nature. 

 Some old forms appear to have survived from inhabiting 

 protected sites where they have not been exposed to very 

 severe competition; and these often aid us in constructing 

 our genealogies by giving us a fair idea of former and lost 

 populations. But we must not fall into the error of look- 

 ing at the existing members of any lowly organized group 

 as perfect representatives of their ancient predecessors. 



The most ancient progenitors in the kingdom of the 

 vertebrata, at which we are able to obtain an obscure glance, 

 apparently consisted of a group of marine animals* resem- 

 bling the larvae of existing Ascidians. These animals 

 probably gave rise to a group of fishes, as lowly organized 

 as the lancelet ; and from these the Ganoids, and other 

 fishes like the Lepidosiren, must have been developed. 

 From such fish a very small advance would carry us on to 

 the Amphibians. We have seen that birds and reptiles 



* The inhabitants of the seashore must be greatly affected by the 

 tides; animals living either about the mean high- water mark, or about 

 the mean low-water mark, pass through a complete cycle of tidal 

 changes in a fortnight. Consequently their food supply will undergo 

 marked changes week by week. The vital functions of such ani- 

 mals, living under these conditions for many generations, can hardly 

 fail to run their course in regular weekly periods. Now it is a mys- 

 terious fact that in the higher and now terrestrial vertebrata, as well 

 as in other classes, many normal and abnormal processes have one or 

 more whole weeks as their periods; this would be rendered intelligi- 

 ble if the vertebrata are descended from an animal allied to the 

 existing tidal Ascidians. Many instances of such periodic processes 

 might be given, as the gestation of mammals, the duration of fevers, 

 etc. The hatching of eggs affords also a good example, for, accord- 

 ing to Mr. Bartlett ("Land and Water," Jan. 7, 1871), the eggs of 

 the pigeon are hatched in two weeks; those of the fowl in three; 

 those of the duck in four; those of the goose in five; and those of the 

 ostrich in seven weeks. As far as we can judge, a recurrent period, 

 if approximately of the right duration for any process of function, 

 would not, when once gained, be liable to change; consequently it 

 might be thus transmitted through almost any number of genera- 

 tions. But if the function changed, the period would have to 

 change, and would be apt to change almost abruptly by a whole 

 week. This conclusion, if sound, is highly remarkable; for the 

 period of gestation in each mammal, and the hatching of each bird's 

 eggs, and many other vital processes, thus betray to us the primor- 

 dial birthplace of these animals. 



