3Q 



NA TURE 



[May 8, 1884 



ject, let us look to the growth of modern opinion on a much 

 older question, but connected with the same axis of rotation, 

 viz. the number of solar days in a tropical year ; reckoned now 

 to be 365-2422414 ± .1, but assumed in the old Julian year = 



Pope Gregory's reformation of the calendar, by introducing 

 the former (or something like it) in place of the latter quantity, 

 caused sufficient disturbance to all the ordinary affairs of men in 

 every nation when it was first adopted ; and has some arguments 

 which may be alleged in its favour still. 



But if I read aright a recent tract by so consummate a physi- 

 cal astronomer as Prof. Simon Newcomb, he holds that the 

 Gregorian alteration has done so much more harm than good, 

 being quite a needless refinement, and is so totally unsuitable to 

 calculations in physical astronomy, compared with the Julian 

 year, that civilised nations should, and presently will, return to 

 that year and reckoning, or " Old Style,"— leaving a few 

 curious computers, whom it may concern, to prepare tables of 

 corrections where they are absolutely required for their own 

 abstruse and recondite purposes. C. Piazzi-Smyth 



15, Royal Terrace, Edinburgh, April 26 



The Ancestor of the Dipper (Cinclus) 

 In NATURE for April 3 (vol. xxix. p. 524) the Duke of Argyll 

 desires Mr. Romanes to prove "that the dipper once had an 

 ancestor which began to dive in water, &c." The Duke well 

 knows that such ancestry cannot be exhibited, but seems unaware 

 that there are other landhvcis Hint are divers besides the dipper 

 ( Cinclus). 1 have 1 iften seen the winter wren dart or dive through 

 a sheet of water, and remain in the damp and dripping space 

 behind the little cascade. The water-thrushes [Seiurus, sp.) all 

 wade in water, ami often, seeing minute mollusca on the bottom 

 of the stream, plunge both head and neck beneath the surface ; 

 so that, often for several seconds, a large part of the body is 

 submerged. Now, these birds, like the winter wren, still have 

 the plumage "pervious to water, and so are liable to be drenched 

 and sodden " ; but they have also the faculty of giving these 

 drenched feathers such a good shaking, that flight is practicable 

 a moment after leaving the water. Swallows, too, are often seen 

 flying in and through spray and thin sheets of falling waters, yet 

 with no detriment to their flight power. Certainly the water- 

 thrushes or wagtails {Seiurus Indovicianus, auricapillus, and nove- 

 boracensis) have taken many preliminary steps towards becoming 

 as aquatic as the dipper 1 Cinclus), and the winter wren, and even 

 Maryland yellow-throat, are not far behind. The Duke can 

 scarcely derive any- comfort from the dippers; Mr. Romanes 

 can. Chas. C. Abbott 



Trenton, New Jersey, U.S.A.. April 18 



Double-storied Houses and Concave Roofs 

 In your issue of January 31 I notice a review of Mr. Im Thurn's 

 book on the Indians of Guiana, in which attention is called to 

 the manner in which a pile dwelling may be converted into a 

 two-storied house, and Prof. Moseley's suggestion that the Swi s 

 chalet did so originate is quoted, the general impression intended 

 being apparently that this, in the majority of cases, is the origin 

 of a double-storied dwelling. Now in that portion of the Hima- 

 layas lying south of the snowy range, to which my personal 

 experience is confined, double-storied houses are almost uni- 

 versal, the lower story being used as a cattle-shed, the uppei as 

 a dwelling ; at the level of the floor a platform is carried out 

 from the building on one side at least, usually on three, or, if the 

 house stands clear of the hillside, on all four sides. The only 

 means of access to this platform and the upper story 1- by a 

 ladder or flight of steps — it is difficult to say which it should be 

 called -but 11 consists of the trunk of a tree split in half on the 

 flat surface of which a series of notches are cut to give foothold ; 

 this is placed in a sloping position leading to the outside edge of 

 the platform, or if, as is often the case, the platform is inclosed 

 by boarding, through a hole in the floor. It will be seen that 

 this is a principle of construction such as might easily have- 

 descended from a pile dwelling, and yet I cannot believe that 

 this is the case ; my reasons are : (1) there are no lakes in the 

 Himalayas in which the habit of building on piles could have 

 been acquired ; (2) the houses are built of dry stone, strength- 

 ened at intervals by timber frames, these frames being without 

 exception horizontal, and built into the wall in courses, such a 

 thing as a vertical post being unknown, while had the style of 



structure descended from a pile dwelling, some trace of the piles 

 would probably have remained, and the house been built on 

 vertical posts whose interstices were filled in with stone, &c. ; (3) 

 though two stories are the rule, it is by no means a 'universal 

 rule: the temples are frequently three, and occasionally four 

 and even five stories high, while those of the natives who are 

 rich enough to afford it build three-storied houses, the ground- 

 floor being used for the cattle, the intermediate one as a storeroom, 

 and the u] iper one as the dwelling. On the other hand, the Kolis 

 or Koltas, an aboriginal race who are as a rule the servants, or 

 practically slaves, of the so-called Brahmin and Rajput land- 

 owners, generally inhabit a single-storied cabin, but where these 

 Kolis themselves own land and cattle, they, too, have double- 

 storied houses. The true origin of this style of building lies, I 

 fancy, in the fact that stone or wooden slabs are practically the 

 only available roofing material, and the preparing and collecting 

 these, not to mention the timbers required, forms a very serious 

 part of the labour involved in building a house, and it is conse- 

 quently an advantage to make one roof cover both the cattle and 

 their owner rather than to undertake the labour and expense 

 involved in two separate roofs. 



In this connection there is a point to which I would wish to 

 call attention. When first entering the Himalayas I was struck 

 by the fact that, whereas the roofs of the villagers' houses were 

 made with a single straight slope from ridge to eaves, those of 

 the temples were as a rule steeper near the ridge, so as to pre- 

 sent a concave outline, and as the ends were usually ornamented 

 with deep weather boards fringed with pendent wooden oma 

 ments, while the corners often had what can best be desi ribed 

 as a wooden tassel, the appearance of the whole was decidedly 

 Chinese ; as I worked higher up into the hills, towards the 

 legion of the deodar, the origin of this construction revealed 

 itself. Where deodar is abundant the roofs of the common 

 bouse,, as well us of the temples, are made of split planks the 

 whole width of the tree, and from 6 to 8 feet long, the ridge 

 being made water-tight by a coping cut out of a single deodar- 

 trei shaped into a ridge above, while in the lower side a V-shaped 

 group is cut. If the row of planks next the ridge were set ul a 

 low angle, it might be difficult to lit this coping; but when the 

 angle of the slope is high, the fitting is easier, and besides the 

 beam by its weight grips the planks of the opposite slopes and 

 holds them together effectually without the need of nails, a con- 

 sideration [of probably far greater weight at the time when this 

 method of construction originated than at the present time. 

 Thus the origin of the high slope near the ridge is explained, but 

 to carry this high slope down to the eaves would necessitate the 

 use of an inordinate amount of material, and so the second row 

 of planks is arranged with a gentle slope; in those rare cases 

 where the roof is large enough to require three separate rows of 

 planks, the middle one is arranged generally with a slope inter- 

 mediate between those of the two marginal ones, and the roof 

 assumes a concave form. As deodar gets scarce the first roofs in 

 which it disappears are those of the villagers' houses, and it is 

 invariably the lower or gentle slope that is the first to be roofed 

 with stone ; then this spreads on to the steeper part of the roof, 

 but here the slope has to be lessened or the slates nailed on. 

 Where roofing slabs can be obtained, large and well-shaped, the 

 latter alternative is adopted, and merely the deodar coping 

 remains of the original wooden roof ; as a last stage this too 

 disappears, and the ridge is made water-tight by carrying the 

 slates of one slope over the edges of those of the other. This 

 style of roof, however, only persist, where rooting slabs are- 

 obtained in such abundance and of such size that they can be cut 

 to the desired shape ; where slates are only procurable of small 

 sizes and irregular shapes, the concave roof is soon found to be in- 

 applicable, the higher slope near the ridge disappears, anil the 

 root assumes the form of a single gentle slope, but in the temples 

 the archaic form survives. I have called attention to this con- 

 cave outline of roof because a similar concavity of outline in 

 Chinese roofs is commonly said to be a survival from the time 

 when the .Chinese dwelt in tents ; this can hardly be the true 

 explanation, for, as Fergusson has pointed out, the Tartar tents, 

 and those of all nomads with which I am acquainted, have a 

 convex and not concave outline. 1 do not know whether there 

 are in China any trees from which roofing slabs of good quality 

 are or could be made, nor have I at present means of access to 

 any books by reference to which I could settle this question, but 

 if such be the case it is more probable that the concave outline 

 originated as I have indicated above than in the manner 

 suggested by Fergusson in his book on Eastern architecture. 



