244 



SCIENCE. 



[Vol. XII. No. 303 



in order to get rid of these irregularities, to get together a larger 

 number of observations. This is done by simply adding together 

 all the first halves and all the second halves : that is, in this case, 

 I have added columns, etc. As I understand it, exception is taken 

 to -this operation, as bringing together quantities which are not ho- 

 mogeneous. Suppose that, instead of adding up directly each half 

 of a series, the mean rainfall at a station is obtained from the whole 

 series. Now, if the proposition as above stated be correct, this 

 mean rainfall is, barring irregular fluctuations, the rainfall of the 

 middle year of the series. Let the residuals be taken. Is there 

 any impropriety in adding up the residuals, not only in each half- 

 series in one sum, but those of all the half-series, for comparison 

 of the sums of the two half-series } 



Or, to put it in mathematical form, let R equal the mean rainfall 

 of a series, which is equal to the rainfall of the middle year, r the 

 rainfall at any time, / the interval in years before or after the mid- 

 dle year (plus when after, and minus when before), x equal the rate 

 at which the rainfall is supposed to increase, which may be assumed 

 as constant over the area, as it is a qualitative rather than a quan- 

 titative result which is sought. We desire to learn whether .r has 

 any considerable value. Then 



r— R 

 }■ := R ± t.r, and .r = ; 



and, for a single series, 



+ 



4 



fl 



t- 



the mean rainfall R being eliminated : .r being the same over the en- 

 tire area, and the mean rainfall being eliminated, the above equa- 

 tion applies to all series, and they may be properly combined for 

 the purpose of obtaining the value of x, and 



r ''n 1 \ r^'\ 

 x=\-\-\-\. 



As has been stated, this method was used to test the above 

 proposition, in the prairie region. Twenty-four stations were used_ 

 and the observations of 428 years were used in evidence. The re- 

 sult showed that there was a trifling amount Diore rain in the ear- 

 lier than in the latter half of the series. In short, it showed that 

 the rainfall had not increased. 



It was applied in Ohio, which from a forested area has become 

 with settlement mainly a deforested area. Under the terms of 

 the proposition, the rainfall should have diminished, but the 

 amount of the diminution is trifling, being but .21 of an inch per 

 year. To this result twelve stations, with 294 years of observation, 

 contributed. 



Southern New England, comprising some 20,000 square miles, 

 was originally a densely forested region. With the progress of 

 settlement it was almost entirely cleared. In recent years, say 

 since i860, a reverse movement has been going on. The competi- 

 tion of Western farms and cheap transportation is driving New 

 England farmers to other vocations, or is forcmg them to move to 

 other parts of the country. Thus the farms are being abandoned, and 

 are growing up to woods. To-day Massachusetts contains 52 per 

 cent of woodland, and Rhode Island even more. Southern New 

 England, then, presents two phases of change for investigation. 

 During the earlier period, with the cutting-away of forests, the 

 rainfall should have diminished, while during the past twenty-eight 

 years it should have increased. During the first period there were 

 used in the investigation eighteen stations, with 400 years of ob- 

 servation. The examination showed that the rainfall \\2^A increased 

 while deforesting was going on. 



In the second period fourteen stations were used and 200 years 

 of observations. The examination showed no change whatever. 



This investigation has convinced me that forests exercise no in- 

 fluence whatever upon rainfall. I wish to state this plainly, as it 

 was suggested at the last meeting that I had some doubts concern- 

 ing the results obtained. I regret that any thing in my paper should 

 be capable of such a construction, as it was certainly as far as pos- 

 sible from my thoughts. 



I am aware that this conclusion is at variance with the popular 



idea, and that a popular idea is not a thing to be disregarded, as 

 there is usually some reason for its existence. We find woodland 

 and a heavy rainfall generally co-existing. In almost all places en- 

 joying a heavy rainfall, the land is covered with forests, unless they 

 have been removed by man. It may be that in this case an effect 

 has been mistaken for a cause, or rather, since it is universally 

 recognized that rainfall produces forests, the converse has been in- 

 correctly assumed to be also true. 



Although forests have no influence upon precipitation, yet they 

 do exert a certain economic influence. Without increasing rainfall, 

 they, in common with other forms of vegetation, economize that 

 which falls, retaining it somewhat as a reservoir, and preventing its 

 rapid descent into the streams. In this way, too, forests tend to 

 reduce the magnitude of floods and to regulate the flow of rivers, 

 thus preventing disaster and improving navigation. This retention 

 of the rainfall is, however, accompanied by a rapid evaporation 

 from the leaf surfaces of the forest, whereby a considerable pro- 

 portion of the rainfall returns to the atmosphere without reaching 

 the earth. On this account it is urged, and I think with reason, 

 that in our arid region, which is dependent for irrigation upon its 

 streams, it is advisable to cut away as rapidly as possible all the 

 forests, especially upon the mountains, where most of the rain falls, 

 in order that as much of the precipitation as possible may be col- 

 lected in the streams. This will cause, not a decrease m the an- 

 nual flow of the streams, as commonly supposed, but an increase, 

 coupled with a greater concentration of the flow in the spring 

 months, and result in rendering fertile a greater area of the arid 

 region. It may be added that the forests in the arid region are 

 thus disappearing with coinmendable rapidity. 



There is no question but that forests reduce the extremes of tem- 

 perature in their immediate neighborhood. They also serve me- 

 chanically as windbreaks, diminishing the force of air-currents. In 

 these and perhaps other ways they serve a useful purpose. 



But with all this in mind, is it worth while to go on planting 

 trees for their climatic effects ? It seems to me, that, apart from 

 the uselessness of it, nature is planting trees at an infinitely more 

 rapid rate than man. For every tree planted under the timber-cul- 

 ture act, or on Arbor Day, a thousand spring up of their own ac- 

 cord. Every deserted farm east of the plains grows up to forest. 

 Half of southern New England is to-day wooded, and the propor- 

 tion is increasing every year, and yet in Massachusetts they have 

 every year an Arbor Day, when the farmers turn out and solemnly 

 plant a tree apiece. 



MENTAL SCIENCE. 



The Psychology of Deception.' 



The deceptive character of the evidence of the senses has be- 

 come attributed to them because of the failure to recognize that we 

 seldom have to do with a sicnple sensation. What deceives is not 

 the information of the sense, but the wrong interpretation of this 

 information by the mind. Such interpretation need not be conscious, 

 and often is not so. The familiar experience of raising a pitcher of 

 water, usually well filled but upon the present occasion empty, and 

 finding it dart upwards in our hands, is a case in point ; for it shows 

 that we estimate the amount of force necessary to raise the pitcher, 

 but only become conscious of this inference when it happens to 

 lead us astray. The phenomena of the stereoscope abound in illus- 

 trations of such unconscious reasonings. One of the simplest 

 types of deceptions arises when such an inference, owing to an un- 

 usual disposition of external circumstances, leads to a conclusion 

 that better evidence shows to be false. A ball held between two 

 crossed fingers seems to be double, because under ordinary occa- 

 sions an impression on the right side of one finger and on the left 

 side of its neighbor (to the left) could only be brought about by the 

 simultaneous contact of two objects. Everywhere, then, we inter- 

 pret the unfamiliar by the familiar, the unknown by the known : illu- 

 sion arises when the objective conditions change their character, 

 and real deception occurs when this change is not recognized, when 

 no better evidence is present to antagonize the false inference. The 

 child who regards a spoon half immersed in water as really bent, 



' See an article with this title by Joseph Jastrow, Ph.D., in the Popular Science 

 Monthly, December, 1888. 



