646 



SCIENCE 



[N. S. Vol. XXXI. No. 800 



tifying of the grorinds, which must receive 

 attention whether a true botanical garden 

 is developed or not. And this brings me 

 to the second of the three conditions which 

 must be met in collegiate gardens. All col- 

 leges desire to have their grounds as beauti- 

 ful as possible, in order to create attractive 

 surroundings for undergraduates, pleasing 

 memories for graduates and favorable im- 

 pressions for parents and benefactors. 

 Now, to this end, the extensive use of trees 

 and shrubs is indispensable. It would 

 seem at first sight possible to combine a 

 good landscape use of these with a system- 

 atic arrangement to illustrate relationships, 

 but I have found, as no doubt have many 

 others before me, that this is only partially 

 possible. Thus, some families contain far 

 more plants of attractive form than others. 

 Imagine confining Coniferse strictly to one 

 section ! Again, the proportion of trees to 

 shrubs is so different in the various fam- 

 ilies that if these were confined to special 

 areas some sections would have few or no 

 trees and others no shrubs. Thus Legu- 

 minosffi have several ornamental trees, but 

 hardly any ornamental shrubs, while this 

 case is reversed in Rosacea, reaching an 

 extreme in Caprifoliaeese, which has no 

 ornamental trees at all. Hence a strictly 

 systematic arrangement can not be com- 

 bined with good landscape results, and the 

 best that can be done is to make sure that 

 representatives of a given family are pres- 

 ent in the appropriate area, even though 

 not confined thereto. But on this plan, a 

 very good collection of trees and shrubs, 

 both pleasing to the eye and useful for 

 study, can be assembled on a college cam- 

 pus. Moreover, trees and shrubs are in 

 condition for study earlier in spring and 

 later in autumn than herbaceous plants, 

 and besides can be studied to considerable 

 advantage all through the winter when 

 herbaceous plants are not visible at all. 



Hence my experience has shovra that of the 

 outdoor garden, the trees and shrubs are 

 far and away the most valuable part ; next 

 come natural gardens, and last of all the 

 systematic garden. There is one other 

 matter worth mention in this connection. 

 The absolute necessity which colleges are 

 under to keep their grounds attractive in 

 any case, makes it possible to develop them 

 as a botanical garden with comparatively 

 little additional expense, for the extra cost 

 of the other features is not relatively great. 

 This applies in part also to the greenhouses, 

 because where these are developed it is 

 possible to give profitable and congenial 

 employment to a good gardener during the 

 winter, and consequently a more competent 

 type of man can be kept, to the great ad- 

 vantage of all the interests involved. 



Another matter which I am finding im- 

 portant in connection with the outdoor 

 garden, but which applies equally to the 

 greenhouses, is this. It is far better to 

 concentrate upon good effects with a few 

 things rather than upon the collection of 

 many. In my ovra garden, we are re- 

 ducing the number of species, but are 

 giving better massing and surroimdings to 

 those we retain, which include especially 

 the kinds the observer is likely to meet with 

 again. Primarily this is in order to con- 

 form to an educational principle of which 

 the importance steadily grows upon me, 

 viz., that the scientific merits of a garden, 

 or of anything else, are not of themselves 

 sufficient to attract persons to their study, 

 but attention must be paid to the peculiari- 

 ties of human nature which demand that 

 things shall be made attractive also. I 

 therefore consider it important to so ar- 

 range plants that they will evoke attention 

 and admiration first, on which basis in- 

 struction is far more easily given. And as 

 the human capacity for attention and ab- 

 sorption is strictly limited, it is no use to 



