ON THE ANATOMY OF THE TODIES. 353 



acquired exhibited by its existing descendants. As most of the 

 Anomalogonatae possess either well-developed caeca, or a tufted oil-gland, 

 whilst all lack the ambiens and accessory femoro-caudal muscles, it may 

 be presumed with some certainty that the ancestor of the group generally 

 possessed both well-developed caeca and a tuft to the oil-gland the 

 first having disappeared in the Piciforrnes, the latter in the Passeriformes, 

 and both in the highly specialized Cypseliformes : at the same time it 

 was destitute of both ambiens and accessory femoro-caudal muscles. 

 The existence of Todus therefore exactly substantiates what might have 

 justly been inferred a priori on purely theoretical grounds ; whilst its 

 insular habitat, the small number of species, and their diminutive size 

 are exactly what might have been expected of a very ancient and 

 synthetic form, which has been unable to hold its own, on the larger 

 areas, with more lately developed and highly specialized forms. On the 

 other hand, it is not to be expected, on the doctrine of descent, that any 

 living form, however synthetic, should be exactly intermediate between 

 any other two living groups, because it is nearly certain to have been 

 modified in some points pari passu with those forms to which it (or, 

 rather, its ancestors more or less remote) gave origin. There are 

 structures in other families of the Anomalogonatae as, e. g., the biceps- 

 slip of the Caprimulgidae, the glutens quintus of the Coliidse, the vomer 

 and the glutens primus of several which are not represented at all in 

 Todus. These may, of course, have been independently reacquired ; 

 inasmuch as, however, they are all structures met with in the Homalo- 

 gonatous birds from some form of which I cannot doubt that the 

 Anomalogonatae are descended it is more probable that they have been 

 inherited directly from a common ancestor which possessed these along 

 with the other structural characters of the Anomalogonatae. That one 

 or more of such structures should have disappeared in Todus, though 

 present in the hypothetical common ancestor, is in no way surprising. 

 I submit, in conclusion, therefore, 



(1) That Todus is a much isolated form, with affinities to both the 

 Passeriformes and Piciformes of Garrod. 



(2) That it cannot be substantiated that Todus is clearly allied to any 

 particular living form of these. 



(3) That this view may be most correctly expressed by making a group 

 Todiformes, equivalent to Passed-, Pici-, and Cypseliformes, for the sole 

 reception of the genus Todus. 



(4) That in all probability Todus, though in some respects much 

 modified and specialized, represents more nearly than any other existing 

 form the common stock from which all the living groups of Anomalo- 

 gonatous birds have been derived. 



