CIRCLES OF THE SHRIKES AND THRUSHES. 339 



(417-) Thus we see that our new circle has this 

 advantage over the old, that the variation of the series 

 composing it turns out to be in accordance with the 

 variation of all other ornithological groups. It can 

 consequently stand three tests, its circularity, its pa- 

 rallelism of analogy with other groups, and its coinci- 

 dence with the established mode, or progression, in 

 which nature varies her groups. The old group would 

 not bear this latter verification, notwithstanding it ap- 

 peared to be circular, and notwithstanding its analogies 

 could be traced in the family ofMerulidce, although the 

 series in the latter family, being made to correspond with 

 the erroneous disposition of that of the Laniadce, ne- 

 cessarily shared in the error, the analogies being correct, 

 but the series in which they are made to follow incorrect ; 

 the exposition of the two groups, as now re-formed, 

 being as follows : 



True Circle of True Circle of 



the Shrikes. the Thrushes. 



Lanius. The most typical of their respective families. Merula. 



Thamnophilus. Bill hooked at the tip. Myothera. 



Edolius. Feet very short. Srachypus. 



Ceblepyrus. Rump feathers more or less rigid. Oriolus. 



Tyrarinus. Frequent the vicinity of water. Crateropus. 



(418.) But it is not sufficient that each of the divi- 

 sions in these two families, as divisions, are correct ; for 

 some of them are either very numerous in species, or 

 contain many striking deviations in their form. Be- 

 fore, therefore, we can pronounce that either of these 

 families are strictly proved, in all their parts, it becomes 

 necessary to institute a further analysis, to select any 

 one of the subordinate divisions, and to submit its con- 

 tents to the very same tests as we have just applied to 

 its family, a* a whole. For instance, the MYOTHERIN^;, 

 or ant thrushes, represented by the genus Myothera, is 

 one of the divisions, or lesser groups, in the circle of the 

 thrushes. Now, is this a truly natural group ? that is, 

 is it circular ? We may fairly conjecture it is natural, 

 because its most obvious distinctions are in accordance 

 with analogies to be traced in other circles, and with the 

 principle of variation. But this, strictly speaking, is 

 z 2 



