THE MORPHOLOGICAL COMPOSITION OF PLANTS. 35 



prone like the kinds thus far described, but growing erect, 

 the stem and attached leaflets become dependent upon a 

 single root or group of roots; and being so prevented from 

 carrying on their functions separately, are made members of 

 a compound individual : there arises a definitely-established 

 aggregate of the third degree of composition. 



The facts as arranged in the above order are suggestive. 

 Minute aggregates, or cells, the grouping of which we traced 

 in 182, showed us analogous phases of indefinite union, 

 which appeared to lead the way towards definite union. We 

 see here among compound aggregates, as we saw there 

 among simple aggregates, the establishment of a specific 

 form, and a size that falls within moderate limits of varia 

 tion. This passage from less definite extension to more 

 definite extension, seems in the one case, as the other, to be 

 accompanied by the result, that growth exceeding a certain 

 rate, ends in the formation of a new aggregate, rather than 

 an enlargement of the old. And on the higher stage, as on 

 the lower, this process, irregularly carried out in the simpler 

 types, produces in them unions that are but temporary ; while 

 in the more-developed types, it proceeds in a systematic way, 

 and ends in the production of a permanent aggregate that is 

 doubly compound. 



contemplate divergent types now existing, would not arise if we had before 

 us all the early intermediate types. The Mammalia differ in sundry respects 

 from all other kinds of Vertebrata Fishes, Reptiles, Birds ; and if the 

 absence of hair, mammae, and two occipital condyles, in these other verte 

 brates were taken to imply a fundamental distinction, it might, in the absence 

 of any known fossil links, be inferred that the Mammalia belonged to a sepa 

 rate phylum. But these differences are not held to negative the assumed 

 relationship. Similarly among plants. We must not reject an hypothesis 

 respecting a certain supposed type, because the existing types it must have 

 been akin to present traits which it could not have had. We are justified in 

 assuming, within limits, a hypothetical type, unlike existing types in traits of 

 some importance. Hence results the answer to a criticism passed on the 

 above argument, that it implies relations between the undeveloped and devel 

 oped forms of the Jungermanniacece such as the facts do not show us. This 

 objection is met on remembering that the types in which the supposed transi 

 tion took place disappeared myriads of years ago. 



