THE GENERAL SHAPES OF ANIMALS. 173 



genera of this group, as in Cesium, Eurhamphcea t and 

 Callianira, that almost equal distribution of parts which 

 characterizes the Beroe is quite lost. 



Here seems a fit place to meet the objection which some 

 may feel to this and other such illustrations, that they amount 

 very much to physical truisms. If the parts of a Medusa 

 are disposed in radial symmetry around the axis of motion 

 through the water, there will of course be no means of 

 maintaining one part of its edge upwards more than another ; 

 and the equality of conditions may be ascribed to the radiate- 

 ness, as much as the radiateness to the equality of conditions. 

 Conversely, when the parts are not radially arranged round 

 the axis of motion, they must gravitate towards some one 

 attitude, implying a balance on the two sides of a vertical 

 plane a bilateralness ; and the two-sided conditions so 

 necessitated, may be as much ascribed to the bilateralness as 

 the bilateralness to the two-sided conditions. Doubt- 



less the form and the conditions are, in the way alleged, 

 necessary correlates ; and in so far as it asserts this, the ob- 

 jection harmonizes with the argument. To the difficulty 

 which it at the same time raises by the implied question 

 Why make the form the result of the conditions, rather than 

 the conditions the result of the form ? the reply is this : 

 The radial type, both as being the least differentiated type 

 and as being the most obviously related to lower types, must 

 be taken as antecedent to the bilateral type. The indi- 

 vidual variations which incidental circumstances produce in 

 the radial type, will not cause divergence of a species from 

 the radial type, unless such variations give advantages to the 

 individuals displaying them; which there is no reason to sup- 

 pose they will always do. Those occasional deviations from 

 the radial type, which the law of the instability of the homo- 

 geneous warrants us in expecting to take place, will, however, 

 in some cases be beneficial ; and will then be likely to estab- 

 lish themselves. Such deviations must tend to destroy the 

 original indefiniteness and variability of attitude must 



