BY J. H. MAIDEN. 2S 



3. Apparenfcly close similarities of different plants 

 may arise from : — 



(a) A difference not easy to detect with the naked 

 eye, and error in stating they are the same may be 

 classed under the heading of "error of experiment,'' or 

 "error of instrument." I do not know of any such 

 case, for very close similarities (not identities) are 

 found to involve real differences as our eyes become 

 better trained. 



(b) Every systematist knows that he ha« made mis- 

 takes, partly because of the paucity of material, and 

 partly because he has over-estimated the reliability of 

 witnesses. These eri'ors are especially liable to occur 

 in a prot( an genus like Ibat of Eucalyptus, whose 

 members are spread over an area larger than that 

 of Europe. 



(c) Homoplasy. We know that really dissimilar 

 plants prove themselves as soon as the floral or fruit- 

 ing organs are exhibited, but that there are closely 

 simulating examples in the vegetative stage every 

 botanist knows. 



4. In a fossil leaf, one cannot take count of adaptive 

 characters as a very general rule ; one must take the leaf 

 as we superficially see it ; of its morphological characters. 

 Let us consider the question of timber for a moment. 



No two sticks of timber in a timber yard or in the forest 

 are precisely alike. The timber is subject to all the limita'- 

 tions of variation of the species from which it sjorang. -tind 

 if these nuances of variation are difficult to record in the 

 species itself, they are difficult to interpret in the quanti- 

 tative records of the tim ber tester. A.11 that we can say is 

 that these records vary between such limits asi have been 

 (perhaps arbitrarily) asigned to the species by the system- 

 atic botanist. To say that the quantitative results are 

 variable between certain limits is another way of saying 

 that the species is variable, that certain forms have been 

 admitted under the banner of the species by the botanist. 

 If the botanist changes his views as to the direction and 

 amount of variation in a species, the timber-tester must 

 modify his figures accordingly, or persuade the botanist to 

 alter his views. There is nothing final about timber tests, 

 and the only way to render them comparable is to render 

 available with them the fullest particulars as to habitat, 

 size of tree, season of felling, and subsequent treatment, 

 relative position in the trunk of the tree of the test piece, 

 particulars in regard to ijie meteorological conditions of 



