I02 



NA TURE 



[May 31, 1894 



crown, we ought to find it in a more or less incipient stage in 

 such remote forms as the cretaceous mammals, whereas the very 

 reverse is the cise, as admitted by Osbom himself ; ' in them 

 the heel is mostly better developed than in such of later forms, 

 marsupial or placental, which morphologically most closely 

 approach them. 



So that, in conclusion, the superior as well as the inferior 

 molars, far from "just emenjin? from the primitive trituber- 

 cular stage," are more remote from it, both morphologically and 

 chronologically speaking, than the tertiary and living forms. 



According to the tritubercular theory, the so-called quadri- 

 tubercular superior molars (composed of four principal cusps), 

 were derived from the tritubercular pattern by the addition of a 

 posterointernal cusp (hypocone). In order to test the value of 

 this hypothesis, let us see what relation the cusps of the superior 

 bear to those of the inferior molars. In a " quadritubercular " 

 mammal, i.g. Erinaccus, the antero-internal cusp (protocone) 

 of a superior molar works as a pestle in its mortar, below 

 which is the posterior valley, formed chiefly by the talonid, 

 whilst the hypocone functions in a similar manner in the hinder 

 molar's anterior valley, formed by the trigonid. According to 

 the theory, the superior "talon" (hypocone) and the inferior 

 "talonid," are later additions to the crown : from this it would 

 follow that the phylogenetically older of the two superior pestles, 

 the protocone, was preformed before its mortar, and vice versii, 

 the posterior of the two mortars was in existence long before 

 its pestle, the hypocone ; so that the second pestle and the 

 second mortar would make their appearance only in a distant 

 future or never appear at all, the function of their respective 

 mortar and pestle thus remaining for ages sinecures. The fol- 

 lowing reasoning, which appears to me more logical, may, how- 

 ever, be not unworthy of consideration. Starting from the 

 assumption that the more complicated pattern, in our special case, 

 the "quadritubercular" molar, is primitive as compared to the 

 tritubercular, we had originally, viz. in the ancestral form of tri- 

 tubercular molar, two superior pestles working in the two inferior 

 mortars, as in the case of the recent Erinaceus. According to 

 this view, the tritubercular type is derived from the quadrituber- 

 cular, by the gradual suppression of the posterior pestle (hypo- 

 cone), fiiri passu with the reduction or suppression of its mortar, 

 the anterior valley, formed by the trigonid. 



The impropriety of the term lrilui<ercular, as applied to many 

 patterns which in reality are multitubercular, is apt to create not 

 a little confusion. Partially to this circumstance I ascribe some 

 obviously contradictory statements to be met with in Mr. 

 Goodrich's recent letter in this journal : On the tritubercular 

 theory,' zs, well as in a previous paper by the same author,^ to 

 which he refers us for further information as to his views on the 

 primitive mammalian molar. 



On looking at the conclusions contained in the appendix (" On 

 the primitive mammalian molar ") to Mr. Goodrich's valuable 

 paper on the mammalia from the Stonesfield slate, I felt gratified 

 to find that he considers it extremely probable that the molars "of 

 the ancestors of both Monotremes and Dilremes were of an in- 

 definite multituberculate pattern," because I had for years 

 advocated the view that the mammalian molar is derived from 

 a palybunous form, a view in favour of which I have reopened 

 the discussion on a recent occasion.' .So, I naively believed at 

 first that it was for this reason my paper was recorded in Mr. 

 GrifKlrich's List of kefcrenecs. My gratification was, however, 

 cm uicrably abated on finding that 1 am merely alluded to in a 

 foomole, where I am stated to have expressed views which are 

 the very reverse of those which every attentive reader will find 

 in my paper to which he refers. 



Id his letter on the tritubercular theory, Mr. Goodrich 

 stales : — " There is much evidence for the view that the upper 

 molars of the pro-mammalian ancestors were of the trituber- 

 cular, and the lower molars of the tubercular-sectorial types ; 

 in fact, I think, we cannot do belter than accept Prof. Cope's 

 generalisation, if not as a definitely established theory, at all 

 cvenu a.s an excellent working hypothesis . . . that the superior 

 molars of both ungulate and unguiculate mammalia have been 

 derived from a tritubercular type ; and that the inferior true 

 molars of both have been derived from a tubercular-sectorial 

 type. " 



• ■I MAmmalt of the Upper Cretaceous Ilcdi.'* 

 ^ V. i9<jj, p. 33J. 

 pp. 67. 

 . i'ti fiom the Sioneifield Slate " (Quart, /ourti. 

 '*• Pt>- i'i-'ll- 



1 H. 



» N/> 

 3 ••<) 



Micr. 



* I' 



NO. 1283, VOL. sol 



I would put the following questions: — (l) How does this 

 last statement agree with the one above quoted from Mr. 

 Goodrich's former paper, referring to "the ancestors of bith 

 Monotremes and Ditremes? " For these alone can be alluded 

 to by the term "Pro-mammalia." (2) How does the accept- 

 ance of Cope's " generalisation," quoted by Mr. Goodrich, and 

 which is the very quintessence of the tritubercular theory, agree 

 with his criticism of this same theory in the beginning of his 

 paper, where he exclaims, "Never before have its weaknesses 

 been so obvious, its errors so plain" ? 



I leave Prof. Cope to reply to Mr. Goodrich's implication, 

 that he extends the tritubercular and tubercular-sectorial types 

 to the molars of the " Pro-mammalian ancestor." 



C. I. Forsyth Major. 



Natural History Museum, 

 S.W., May 10. 



The Determination of Latitude and Longitude by 

 Photography. 



Latitude and longitude may be determined on shore with 

 considerable accuracy by means of an ordinary photographic 

 camera ; and this method will, I think, prove useful for several 

 reasons. The observation part of the lousiness, which consists 

 in taking the photographs, is separated from the measuring and 

 calculating part, and may be performed by different persons 

 at difl'erent times and places. For taking the photographs no 

 scientific apparatus is wanted besides the camera and a wa'.ch, 

 the latter to measure intervals of not more than a few hours 

 with an accuracy of a second or so. .-Anyone may easily be 

 instructed to take the photographs, as no mathematics and very 

 little astronomical knowledge are wanted, only the measuring 

 of the plates and the calculation requiring some scientific 

 training. 



I will first describe the determination of latitude. The 

 general plan is well known to astronomers ; it is here only 

 adapted to the use of an ordinary camera. 



If the lens is directed to the zenith, and the stars are allowed 

 to draw their trails over the plate, it will evidently be possible 

 to determine the latitude from the plate, provided the point of 

 the plate is known which corresponds to the zenith. Now this 

 point may be found in the following manner : — Let the camera 

 float and be turned while floating. It will turn exactly round a 

 vertical axis. The zenith will then be the only point of the sky 

 whose corresponding point on the plate will remain unaltered. 

 The way I arranged the experiment was this : — I pl.aced the 

 camera in a rectangular tin box, lens upwards, fixing it firmly 

 by wedges. The box was ballasted so as not to be capsized 

 when floating. Photographic plates will do .as b.iUast, but it is 

 better to take a plate of lead, which brings the centre of gravity 

 lower down, and thereby increases the stability. I then filled 

 a second rectangular box, somewhat larger than the first, with 

 so much water, that when the first box floated in it, the rims of 

 both boxes were in the same plane. Two opposite points on 

 the rim of the inner box were fastened to the rim of the outer 

 one, each with two strings, forming an obtuse angle. This pre- 

 vented the inner box from turning during an exposure. The 

 strings need not be tight, as the forces they have to resist are 

 small. 



Having placed the whole apparatus on a firm t.able in the 

 garden, I removed the cap from the lens. As the trails of 

 brighter stars may generally be measured more accur.ately than 

 those of faint ones, I exposed long enough for a star of the third 

 magnitude to pass sufficiently near the zenith to appear on the 

 plate. After replacing the cap, I turned the whole apparatus 

 carefully through an angle of 180°. For this purpose I h.ad pressed 

 a row of flat nails into the board along one edge of the outer 

 box, and having turned it, I let the same edge of the box touch 

 the row of nails from the other side. Before again removing the 

 cap, I waited about three minutes to allow the oscillations of 

 , the water and the inner box to subside. The second exposiirc 

 I made long enough for the image of a star of about the third 

 magnitude to be thrown on the plate during the star's culmina- 

 tion near the trail of the star of the first exposure. In order to 

 see that it is possible to determine the latitude from tliis plate, 

 imagine first the trails ol the two stnrs to touch. That would 

 be a proof that they had |)assed the meridian on difl'erent sides, 

 but at the same distance from the zenith. The mean of their 

 declinations, taken from the almanacks, would therefore be equal 

 to the distance of the zenith from the equator that is equal to 



