37U T. G. HALLE. [Nov. 1913. 



as applicable to the whole species. Palaeobotany in general 

 is, indeed, dependent on the possibility of so nsing, with all 

 possible precaution, the palaeobotanical species, and the em- 

 ])loyment of specific designations mainly serves just to mark 

 utf such units. It is clear, therefore, that it must be an im- 

 perative demand to make these units as pure and unmixed 

 as possible, so that a statement made in regard to some spe- 

 cimens of the species may with the greatest possible proba- 

 bility apply to all others designated by the same name. An 

 example may serve to illustrate the dangers w^hich, from this 

 point of view, attend a broad classification. We may imagine J 

 two sets of specimens of fern-leaves, which an adherent of h 

 narrow classification would prefer to treat as two species but 

 which, with a liberal delimitation of the species, are regarded 

 as one. Specimens of the first of these sets are later found 

 to have sporangia of, for instance, Osmundaceous type, and 

 the result will be that the whole species will be stated to 

 belong to that group. If then, after some time, specimens of; 

 the second set are found to belong to stems showing the ana-^ 

 tomieal features of the Marattiales, this character, by still 

 regarding the species as a unit, will probably be applied to 

 the species as a whole. The simple explanation may be that 

 the name is used for two different natural species which may 

 well belong to quite different groups. — It is clearly imposs- 

 ible to guard satisfactorily against this risk, even with a 

 narrow delimitation of species, and no doubt a large group 

 of mistakes in palaeobotany is of this nature. But the use 

 of specific names in a narrow sense will naturally tend to 

 lessen this risk, whereas it hardly brings any other danger 

 instead of it. In the case quoted, a mistake in this direction 

 would only mean that the process of reconstruction of the. 

 particular species is delayed. 



However one may look at the problem, it would appear that 

 the use of specific names in a restricted sense would be less apt to 

 cause confusion and mistakes. The most conspicuous objection, 



