1896 | BOTANICAL OPPORTUNITY 203 
restricting the species cultivated to the smallest number capable 
of adequately expressing the facts to be shown. Perhaps it 
is safe to say that an institution able to maintain an herbarium 
of half a million specimens, representing one-fifth as many 
species, is doing exceedingly well if it has in cultivation at 
any one time 10,000 species of the higher plants, and there 
are very few gardens which actually grow half of this number, 
while no inconsiderable percentage of the plants cultivated are 
so deformed, distorted, dwarfed, and imperfect, as a general 
thing, that they can scarcely be said to represent the species 
whose name they bear, either in appearance or technical char- 
acters. 
This leads to the conclusion that not only class gardens but 
research gardens should be kept within reasonably narrow 
bounds, so far as permanent planting is concerned, while allow- 
ing sufficient elasticity for rapid and ample temporary expan- 
sion in certain directions along which work is planned. This 
does not necessarily mean that any considerable amount of land 
not used in the permanent plantation need be reserved for 
Special expansion. Asa rule the more important gardens are 
Situated in or near large cities, and the high price of land alone 
would prevent such reservation in most instances; but the 
impure atmosphere of many of the larger cities is a further and 
even stronger reason for selecting for any large experimental 
undertakings a suitably located and oriented tract of farming 
land easily rented for one or several years at a relatively low 
gure. 
Granting the wisdom of such temporary adjuncts to a 
research garden, a step further leads to a recognition of the 
Possibility of securing the most varied climatic conditions by 
€stablishing branch gardens located where particular kinds of 
Study can best be carried on. In no other way can gardens be 
made to contribute to the fullest extent to the study of marine 
°F seaside plants, alpines, or the great class of succulents. etc., 
characteristic of the arid regions of our southwestern states and 
territories, and in no other way, except in the field, can these 
