J UNE 29, 1899] 
classes, &c. Thus an analysis of the grades o. chest-girth of 
255 boys before and after a three terms’ course of compulsory 
gymnastics showed that the following improvement had been 
made. The numbers in the lower line give the percentage of 
the boys examined, who made the number of grades improve- 
ment indicated in the line above. r 
Improvement :— ‘ 
rgr.|2er.|3 er. | 4gr-|5er-|6gr.|7gr-| Sgr. | o gr. | rogr. | tr gr. 
Per cent. :— 
ode | 2. [mle th ORS |! 4 eo] CRA ieee 
This with the omitted fractions gave 73 per cent. of the boys 
who had made more or less marked improvement relative to the 
general mass of boys of their age, the improvement in some 
cases being very marked indeed. 
An analysis of the growth of 161 boys by means of their 
grades showed that the scheme of growth corresponded to the 
scheme indicated by the curves in the diagram in 31 per cent. 
ofthe casesexamined. There was a steady rise relative to this 
standard in 17 per cent. ; a steady fall in ro per cent. ; a period 
of rise followed by one of fall, or vice versd, in 18 per cent. In 
9 per cent. the variation was erratic, and the remaining 15 per 
cent. probably belonged to the first group; but not within the 
limits of variation allowed. 
In 68 per cent. the type of structure, as indicated by the re- 
lation of height to weight, was stable throughout the period 
examined ; but in about one-fourth of these cases there was a 
considerable constant difference between the grades of height 
and weight, amounting in the most extreme cases to as much as 
eight grades. 
The lesson drawn from these observations was that, in order 
to form a correct opinion relative to a boy’s physical progress by 
means of his measurements, it is very desirable to keep a regular 
record of his growth, in order that the general scheme of his 
growth may be determined, and that any irregular fluctuations 
due to external and removable causes may be noted and properly 
dealt with. 
The Giant Tortoises of the Galapagos. 
I NOTICED in your issue of June 15 a paragraph about the 
Galapagos tortoises. I do not know if this information is of any 
interest, but during my residence in Hawaii I knew of two 
living there. One of them lived ina garden near Hilo, and 
belonged to the late Captain Thomas Spencer ; I last saw it 
about 1880. The other one lived on the Waimea plains in a 
perfectly wild state, and I used frequently to come across it 
when out shooting. It used to wander about within a radius of 
three or four miles. 
It was blind of one eye, and its shell had lichen growing on it, 
and it could move about with a man sitting on its back. 
I last saw it in 1890, but it may possibly be still living ; this, 
however, could easily be ascertained. 
They were, I believe, brought to Hawaii from the Galapagos 
in whalers, and were of great age. If desired, I shall endeavour 
to find out if they are still alive. W. HERBERT PURVIS. 
10 Alexandra Place, St. Andrews, Fife. 
School Laboratory Plans. 
COLLEGE plans are not always safe precedents. Boys need 
more supervision. Can any of your readers advise as to the best 
arrangement of benches for a class of twenty-four to thirty boys, 
aged fourteen to seventeen, doing chemistry and physics with 
elementary quantitative experiments ? 
(1) Is the double back-to-back bench the best form? It may 
economise woodwork, but it makes the class face both ways, and 
attention to verbal instruction is less easy. 
(2) Is the superstructure of shelving necessary? If qualita- 
tive analysis is not done, fewer bottles are needed. The super- 
structure hinders conversation across double benches, but it 
stops supervision also. 
(3) What is the best way of arranging the benches so as to 
allow of supervision and keep wall spaces free for shelving ? 
They may be (a) all round the wall, leaving no space for shelves 
and cupboards ; or (4) single bench along two walls and double 
bench down the middle ; or (c) across the room, double benches 
alternating with windows, well lighted but difficult to supervise ; 
(d) central aisle with double bench extending to walls right and 
left ; (ec) double benches, Jengthways, free from walls ; (/) single 
benches, cross-ways, like the desks of an ordinary class-room. 
I shall be grateful for any help or advice. 
Bootham School, York, June 23. HucuH RICHARDSON, 
NO. 1548, VOL. 60] 
NATURE 
199 
Pair of Brazilian Marmosets Breeding in England. 
A PAIR of marmosets, which for the two past winters have 
had a free run of our greenhouse and garden (in Buckingham- 
shire), produced two young ones on May 24. They seem to 
thrive on freedom and exercise, and the young ones are now be- 
ginning to feed themselves. In hot weather they like to remain 
out all night, but at first they came in to their box in the green- 
house every evening, the male parent always carrying the twins 
on his back, their little round furry heads merely looking like 
small excrescences each side of his neck; and only handing 
them to the mother at feeding-times, and then carefully lifting 
them back with both hands and settling them into position, 
where they seem to cling on without being held. 
Their favourite garden house appears to be an old bird’s nest, 
rather high up ina pink thorn-tree, some distance from the green- 
house. They very rarely come down to the ground, but the 
female will answer a call and come to feed from the hand. 
Bananas, milk and water, insects and young birds are the foods 
they like best. Dora WHITMORE. 
THE DIFFRACTION PROCESS OF COLOUR- 
PHOTOGRAPHY. 
Ape production of colour by photography has been 
accomplished in two radically different ways up to 
the present time. In one, the so-called Lippmann process, 
the waves of light form directly in the photographic film 
laminze of varying thickness, depending on the wave- 
length or colour of the light. These thin laminz show 
interference colours in reflected light in the same way 
that the soap-bubble does, and these colours approximate 
closely to the tints of the original. 
The technical difficulties involved in this process are so 
great that really very few satisfactory pictures have ever 
been made by it. The other, or three-colour process, has 
been developed along several distinct lines, the most 
satisfactory results having been produced by Ives with 
his stereoscopic “ Kromskop,” in which the reproduction 
is so perfect that, in the case of still-life subjects, it would 
be almost impossible to distinguish between the picture 
and the original seen through a slightly concave lens. 
The theory of the three-colour method is so well known 
that it will be unnecessary to devote any space to it, 
except to remind the reader of the two chief ways in 
which the synthesis of the finished picture is effected 
from the three negatives. We have first the triple lantern 
and the Kromscope in which the synthesis is optical, 
there being a direct addition of light to light in the com- 
pound colours, yellow being produced, for example, by 
the addition of red and green. The second method is 
illustrated by the modern trichromic printing in pigments. 
Here we do not have an addition of light to light, and 
consequently cannot produce yellow from red and green, 
having to produce the green by a mixture of yellow and 
blue. Still a third method, that of Joly, accomplishes an 
optical synthesis on the retina of the eye, the picture 
being a linear mosaic in red, green and blue, the in- 
dividual lines being too fine to be distinguished as 
such. 
The diffraction process, which I have briefly described 
in the April number of the Philosophical Magazine, is 
really a variation of the three-colour process, though it 
possesses some advantages which the other methods do 
not have, such as the complete elimination of coloured 
screens and pigments from the finished picture, and the 
possibility of printing one picture from another. The 
idea of using a diffraction grating occurred to me while 
endeavouring to think of some way of impressing a 
surface with a structure capable of sending light of a 
certain colour to the eye, and then superposing on this a 
second structure capable of sending light of another 
colour, without in any way interfering with the light 
furnished by the first structure. This cannot, of course, 
be done with inks, since if we print green ink over red, 
the result will not be a mixture of red light and green 
