March 8, 1889,] 



SCIENCE. 



177 



in the cases where the plant was less than one foot in diameter at 

 the ground, and they do not become a conspicuous feature until the 

 tree is nearly adult ; i.e., until it has attained a diameter of eigh- 

 teen inches or more. At this stage of growth, if the crown be per- 

 manently wet, the knees become an extremely conspicuous feature, 

 fifteen or twenty often being found grouped about a single stem at 

 the distance of from five to twenty feet from the base of the bole. 

 It thus appears tolerably certain that the need of having a portion 

 of the roots above the water-level will be found in certain other 

 trees. Thus far my note-books supply me, however, no certain 

 indications of this fact. Indeed, it is only in the case of Taxodium 

 and the tupelo that I have found the plants under circumstances 

 which would show clearly their needs in this respect. 



There is another feature concerning the growth of water-loving 

 trees, or at least those which are tolerant of permanent moisture, 

 that deserves attention. I have reference to the form of the bole 

 or trunk as it is exhibited in the specimens of the Southern species, 

 which occupy situations diversely affected by moisture. On very 

 wet ground the trunk appears to be generally expanded at the 

 crown, in a measure, which is not the case in specimens of the 

 same species growing in dryer situations. Thus, in the cypress, we 

 not infrequently find the bole at the crown, and for some feet above, 

 having a diameter twice as great as it is at ten feet above the sur- 

 face. Where, however, the tree grows on a dryer soil, the expan- 

 sion at the base is much less considerable. The same appears to 

 be the case in the tupelo, which often has a remarkable expansion 

 of the trunk near the surface of the ground, where the plant oc- 

 cupies very wet situations. In a somewhat less degree, this fea- 

 ture appears to exist in all our trees, except the willows, which 

 occupy sites characterized by diversity in the measure of wetness. 

 I should state that this opinion rests entirely on eye-measurements. 

 I have long intended to submit the impression to the criticism of a 

 careful determination, but have not been enabled to do so. The 

 impression has, however, been so often repeated to me in different 

 regions, that I am inclined to believe there is little chance of error 

 in the statement. I trust that some one who is well placed for such 

 observations will subject the suggestion to a careful statistical in- 

 quiry. 



If I be correct in the opinion that trees in very wet situations 

 develop an enlarged bole near the surface of the ground more 

 frequently than those which occupy dryer situations, we may per- 

 haps account for the fact in the same way in which I am disposed 

 to explain the occurrence of knees in Taxodhini and of root-loops 

 in the tupelo ; viz., through a need of an aeration of the sap, which 

 is denied in roots that are under water. 



It appears to me from eye-observation, as yet uncorrected by 

 measurements, that the buttresses which the water-loving trees 

 form about the trunk are more considerable than they are in the 

 same species on higher land. If this be really the case, it may 

 perhaps be due to the same physiological need which has led to 

 the formation of knees, and to the enlargement of the bole near the 

 crown of the tree. I feel less confident as to this increase in the 

 buttress prominences than I do concerning another observation 

 which I have above set forth. I state the impression for the reason 

 that it has very frequently been borne in upon me in my studies on 

 the development of swamp-plants. At first I was disposed to at- 

 tribute the peculiarity to the fact that the roots of swamp-trees do 

 not usually extend far beneath the surface, and therefore the but- 

 tresses were enlarged in order to give greater stability to the trunk. 

 This hypothesis was disproved by the fact that trees growing in 

 such situations are very rarely uprooted by storms. I failed, in- 

 deed, to find a single case of such uprooting by the action of the 

 wind in several thousand miles of journeys through the morasses 

 in the eastern part of the United States. The only cases in which 

 such overturning met my eye appeared in the swamps near the 

 Mississippi, which, on the whole, exhibit buttress structures much 

 less conspicuously than the trees of the Atlantic coast morasses. 



There is another interesting series of facts connected with the 

 effect of excessive water on our forest-trees which are tolerant of 

 swamp conditions. These relate to the variations in the character 

 of the bark, the mode of branching, etc., of the plants in situations 

 diversely conditioned as regards the amount of moisture. In al- 

 most all our forest-trees, which range from dry to very wet stations. 



there are noticeable diversities as regards the above-mentioned fea- 

 tures, according to the station they occupy. Thus the ordinary 

 chestnut oak varies in a very noticeable manner between dry ground 

 and wet. The tree in very wet situations has a much smoother 

 bark than it exhibits on high land, and I am told by the woodmen 

 that this bark in trees which grow within the swamp is unfit for 

 the purposes of tanning. The variety of tupelo known to the wood- 

 man of the Dismal Swamp as the " pawpaw gum," appears to owe 

 its peculiarities to the fact that it normally grows in much wetter lo- 

 calities than the ordinary Nyssa. It differs from the parent species 

 in that the bole is singularly enlarged near the crown, often having 

 a diameter for some feet above the surface of the water two and 

 one half times as great as it has at the height of ten feet above the 

 ground. In this connection it may be noted that this variety of the 

 tupelo is less disposed to develop the root-loops than the more 

 common form, it appearing indeed as if the great extension of the 

 bole near the crown made the development of these processes un- 

 necessary. 



The variation in the character of our forest-trees when exposed 

 to swamp conditions affords an extremely interesting field for an 

 important class of inquiries concerning the influence of environ- 

 ment, and the effect of natural selection, on the development of 

 organic forms. In the Dismal Swamp, where the water-level 

 during the growing season is subject to relatively little variation, a 

 difference in altitude of six inches, or at most a foot, will greatly 

 affect the character of the timber-trees and other plants. With 

 each such variation in height, we perceive a noteworthy change in 

 the character of the vegetation. 



MENTAL SCIENCE. 

 Statistics of Visual Images. 



The American Society for Psychical Research has devoted con- 

 siderable time to the study of unconscious mental habitSiJ^^.'a. field 

 that abounds in suggestions applicable to the class of pljjl^gmena 

 which such a society investigates, but is still more valuable as 

 contributing to our knowledge of obscure mental traits. Thus 

 Professor C. S. Minot has shown that we are by no means as likely 

 to think of one number as of any other, when simply asked to think 

 of a number, but that there exist certain definite and very general 

 preferences for certain numbers above others. People have " num- 

 ber-habits," or unconscious tendencies f o choose a certain few num- 

 bers (perhaps on account of greater familiarity, easy manipulation, 

 peculiar association, brevity of utterance, or other causes) when an 

 unlimited choice is offered them. In No. 4 of the " Proceedings of 

 the American Society for Psychical Research," Professor Minot 

 brings together extremely interesting material for a similar study 

 with reference to the "diagram-habit." The committee on experi- 

 mental psychology of this society sent out a large number of postal- 

 cards bearing the printed request, " Please draw ten diagrams on 

 this card, without receiving any suggestion from any other person, 

 and add your name and address." Five hundred and one such 

 cards have been collected, of which 310 were drawn by men, 169 by 

 women, and 22 had no name. 



The first point of interest in such a study is to observe how 

 various the drawings of five hundred persons will be. We are not 

 told how many different designs occurred ; but the occurrences of 

 83 different designs have been tabulated, and their sum includes 

 about half of all the drawings. But the real poverty of the in- 

 tellect when it expresses itself naturally is made evident by the 

 great preponderance of a very few simple diagrams. Thus circles 

 were drawn 209 times; squares, 174 times; equilateral triangles, 

 160 times; crosses, 160 times; letters of the alphabet, 82 times; 

 diamonds, 80 times ; oblongs (horizontal), 78 times; inscribed 

 circles, 78 times; stars, 77 times; faces (profile to the left), 61 

 times; houses, 56 times; rhombi, 56 times; scrawls, 53 times; 

 other animals and heads, 48 times; flowers, 46 times; leaves, 45 

 times ; hexagons, 42 times ; cubes, 43 times ; right-angled tri- 

 angles, 42 times ; figures of men, 32 times ; and so on. The 

 above are the twenty most frequent drawings, and, it will be 

 seen, form an aggregate amounting to nearly one-third of all the 

 drawings. On the average, each occurs So times. If we group to- 



