April 12, 1888J 



NATURE 



557 



spondent ''A. B." draws attention, there can be no doubt, so 

 far as it is regarded as a storehouse of information ; but whether 

 the system of classification proposed therein will not tend to 

 retard rather than to further progress is a question on which I 

 am at present more than doubtful. The two points to which 

 Dr. Hatch and your correspondent draw attention as inherent 

 weaknesses — viz. the "dyke rocks," and the subdivision of the 

 "effusive rocks" into palseo-volcanic and neo-volcanic — appear 

 to me such serious defects, that to praise a system which largely 

 rests on them is like complimenting a viaduct by saying that it 

 is an excellent viaduct but two of its piers unfortunately have a 

 bad foundation. 



But dismissing this as a question too large for discussion in 

 your columns, I will confine my remarks to some defects in 

 detail, rather serious as they appear to me, which are exhibited 

 by the classification as tabulated by your correspondent. 



(i) A mere "linear" classification fails, I believe, to repre- 

 sent satisfactorily the relation of the igneous rocks, because it 

 separates too widely rocks very closely related — such, for 

 instance, as the Dacites and Rhyolites (Liparites), and their 

 coriresponding holocrystalline representatives. Hence I believe 

 that the branching system such as I indicated in my Presidential 

 Address to the Geological Society in 1885 is more logical and 

 more in accordance with the facts of Nature. 



(2) In regard to the above example, I fail to understand why 

 Dacites should be included with Andesites and Liparites sepa- 

 rated from Trachytes, or, if we speak of their holocrystalline 

 representatives, why we should separate Granite from Syenite, 

 while we include Tonalite with Diorite. It is true that Granites 

 are common and Tonalites are rare, and possibly the latter 

 always contain some hornblende ; but until it is shown that a 

 quartz-plagioclase-biotite rock does not exist, or that the sub- 

 stitution of hornblende for biotite is of primary importance, 

 there does not seem any valid reason for suppressing the group. 



(3) As the term Diabase has long had a recognized meaning, 

 I fail to see any good reason for substituting it for Dolerite, to 

 vyhich, as generally understood, it stands in much the same rela- 

 tion as do many Serpentines to Peridotites. Neither can I 

 admit the propriety of separating Gabbro from it. 



(4) The wide separation of the Leucite and Nepheline rocks 

 from the Basalts seems also to me to be of doubtful advantage. 



(5) If the term Peridotite be used in the sense in which it has 

 generally been employed {e.g. by Dr. Wadsworth in his excel- 

 lent " Lithological Studies") — namely, to denote a rock in 

 which silicates of magnesia and iron abound, with some also 

 containing lime but with little alumina — it is surely not possible 

 to regard Limburgite as its "effusive" equivalent. That rock 

 seems to me to be more properly associated with the Picrites, 

 not as defined by Dr. Wadsworth [i.e. olivine-augite rocks), ' 

 but as equivalent to the Palitopicrites of some authors — viz. 

 pyroxenic rocks, containing a fair amount of olivine, and 

 some feldspar, which last, however, has a very variable propor- 

 tion. The true position of these rocks appears to me to be as a 

 connecting-link between the Peridotites and the Dolerites. 



There are other points in the work to which I should like to 

 call attention, but I am writing away from books, and should 

 have refrained for a season had not " A. B.'s" letter seemed to 

 me to call for a word of friendly protest. No one can be more 

 deeply sensible than I am of the value in many respects of 

 Prof. Rosenbusch's work, but until his classification rests on a 

 firmer foundation it will not, I fear, be really helpful to 

 students in leading them to clearer ideas on a complicated and 

 difificult subject. T. G. Bonney. 



The Delicacy of the Sense of Taste. 



At the Philadelphia meeting of the American Association, 

 in 1884, we presented a paper upon the general subject of the 

 "Sensitiveness of the Special Senses." We have since con- 

 tinued our investigations,' and have the honour to present at this 

 time the results of some experiments upon the sense of taste. 



The object of the experiments herein described was to find 

 out what substances, or classes of substances, are most readily 

 detected by the sense of taste, and the relative delicacy of this 



' See the following papers:— "On the Relative Bitterness of different 

 Bitter Substances " (Proceedings of the Kansas Academy of Sciences, 

 1885) ; ; " On the Relative Sweetness of different Sugars " (Reports of the 

 Kansas Board of Agriculture, 1885); '"On the Sensitiveness of the Eye for 

 Colours of a Low Degree of Saturation " (American Journal of Science, 

 in. vol. XXX. p. 27); "'The Sense of Smell" (Nature, vol. xxxv. p. 74). 



sense towards these substances. For the production of familiar 

 typical effects upon the organs of sense the following substances 

 were selected : — 



I. (bitter) 

 II. (sweet) 



III. (acid) 



IV. (alkaline) 

 V. (saline) 



Quinine. 

 Cane sugar. - 

 Sulphuric acid. 

 Sodium bicarbonate. 

 Common salt. 



The attempt was made to include other substances in this list, 

 but it was difficult to find any, not embraced in the five classes 

 above mentioned, which would not betray their presence either 

 by colour or odour. Indeed it is surprising, to one who has not 

 given the subject attention, to what an extent we are accustomed 

 to depend upon the aid of the sense of smell in the classification 

 of tastes. The fact has been noticed by several authors that, if the 

 nostrils are closed, the range of our taste becomes very much 

 limited. 



Our method of testing the deligacy of the sense of taste was 

 to make solutions, of known strength, of the different substances ; 

 then, by successive dilutions, to make from these several series 

 of weaker solutions ; each one being of one-half the strength of 

 that preceding it. All the bottles containing these substances, 

 and several bottles of water, being placed side by side without 

 regard to order, the person to be tested was requested to taste of 

 each solution and place it in its proper class. In each series the 

 lowest solution was so very dilute that it was deemed impossible 

 to distinguish it from water. Unknown solutions were to 

 be classed with water. As the tasters were found to be 

 liable to mistake occasionally even the stronger solutions, an 

 opportunity was given, at the close of the test, to correct such 

 accidental errors. In carrying out these tests we found that the 

 most trustworthy results were to be obtained by instructing each 

 operator to pick out the stronger solutions, temporarily classifying 

 with water all which were not immediately recognized ; and then 

 to go over the latter solutions a second time, properly classifying 

 such as could further be detected. 



These tests were made by 128 persons, between the ages of 

 twelve and fifty— eighty-two men and forty-six women. The 

 average results are given in the following table : — 



Table of Averages. 



I. Quinine. 



Male observers detected i part in 390,000 parts of water. 

 Female ,, „ i „ 456,000 „ ,, 



II. Cane sugar. 



Male observers detected i part in 199 parts of water. 

 Female „ „ i ,, 204 „ 



III. Sulphuric acid. 



Male observers detected i part in 2080 parts of water. 

 Female „ „ i „ 3280 ,, 



IV. Bicarbonate' of soda. 



Male observers detected i part in 98 parts of water. 

 Female ,, „ i ,,126 „ 



V. Common salt. 



Male observers detected i part in 2240 parts of water. 

 Female ,, ,, i ,, 1980 ,, ,, 



This table indicates only the average delicacy of taste for each 

 substance included in our experiments. The tests brought to 

 light ma iy astonishing individual peculiarities. For instance, 

 there were persons who could detect with certainty I part of 

 quinine in 5,120,000, while others failed to notice i part in 

 160,000. How far this difference is due to education it is not 

 possible to say. Among the tasters were quite a large number 

 who had been accustomed for several years to the handling and 

 recognition of drugs and chemicals. Their record was consider- 

 ably above the general average, but they were, on the other 

 hand, surpassed by a few individuals who had had no previous 

 training. 



The results of our experiments may be briefly summed up as 

 follows : — 



1. The sense of taste is much more delicate for bitter sub- 

 stances than for the others included in our list. (The relative 

 delicacy for quinine and sugar is very nearly 2000 : i . ) 



2. Taken in the order of their effect upon the organs of taste, 



