76 rnociiiiuiNGs oi' the 



resL'iirch iiny disclose the existence of a number of examples. 

 But, so far as I know, the hypothesis of alternating habitats has 

 hitherto only in one case received a botanical application. 

 Dr. I). U.Scott*, raanj^ years ago — two years, in fact, before 

 Dollo on the zoological side formulated the 'Law of Irreversi- 

 bility' — put forwai'il the illuminating suggestion tliat the expla- 

 nation of tlie ajiatoniical peculiarities of the genus Gumiera might 

 lie in an a.i!cestral liistory in which an original terrestrial period, 

 followed by an jupiatic phase, has been succeeded by a second 

 terrestrial period. Scott showed that the single stt^le of the stem 

 of the acpiatic genus Mi/riuphifllum was com|)arable with one of 

 the steles of a polystelic (^raanera, and more recent work has con- 

 firmed the justice of this comparison t. He further suggested 

 that such a terrestrial plant as Ganneta might possibly have been 

 descended from an aquatic ancestor anatomically similar to Mi/rio- 

 phxjllam — that is to say, characterised by the condensed stele and 

 lack of secondary thickening which are so often associated with 

 water-life. A descendant of such a reduced aquatic plant can 

 only revert to terrestrial life if it is able to develop an increased 

 supply of vascular tissue ; but no longer having a cambial system 

 available for the purpose, it cannot achieve this end except by 

 falling back upon the multiplication of existing steles. Ex|)ressing 

 this exam|)le in tt'ruis of the ' Law of Loss,' we may say that the 

 cambial system once discarded could not be regained, while the 

 expedient of adding to tlie number of reduced steles represents a 

 device for repairing this irrevocable loss by means of such substi- 

 tutes as are to hand. 



If we accept the views of Samuel Butler so far as to admit 

 that there is at least an analof/ij of a verv intimate kind between 

 heredity and unconscious memory, each example of the 'Law of 

 liOss' may perhaps be visualised as representing a lapse or failure 

 of memory. If an organ be lost, the remembrance of it presum- 

 ably in course of time becomes more and more remote, until 

 finally, even if circumstances renew the need for it, the memory 

 has so entirely faded that the plant cainiot, as it were, recall how 

 to reconstruct it. It is thrown, s > to speak, on its own resources, 

 and is thus compelled to discover for itself some method of 

 responding upon new lines to the ancient need. 



In connexion with the ' Law of Loss ' it is important to notice 

 how often the evolutionary tendencies within an\' particular 

 group, as soon as we are able to analvse them in detail, resolve 

 themselves into phases of reduction. This is obviously true in the 

 special case of the aquatic Angiosperms, but it seems to have a 

 much wider application. The theory, for instance, is now very 

 generally held that Monocotyledons are reduced descendants of 

 a Dicotyledonous stocic. In the case, again, of floral evolution 



* S.-ott, D. II. (ISOl) ; see also Riissow, E. (187.")). 

 t Schincller, .\. K. (1904). 



