July 7, 1923] 



NA TURE 



Letters to the Editor. 



[The Editor does not hold himself 7-esponsible for 

 opinions expressed by his correspondents. Neither 

 can he undertake to returfi., nor to correspond with 

 the writers of rejected manuscripts intended for 

 this or any other part of Nature. No notice is 

 taken of anonymous communications.'] 



Positive Ray Analysis of Copper. 



The chief difficulty in analysing an element with 

 a high melting-point by means of positive rays lies 

 in the construction of a suitable furnace for evaporat- 

 ing the metal. I have recently succeeded in obtaining 

 rays of copper by using a molybdenum furnace, 

 heated with a coil of molybdenum wire enibedded 

 in alundum cement. Three isotopes were observed 

 separated by two units in atomic weight. The rela- 

 tive intensities were about 1-4:1:1, the lightest being 

 the strongest. Rays of rubidium were also obtained, 

 probably from the cement, and showed two isotopes, 

 as found by Aston with his method of analysis. The 

 relative intensities gave a mean atomic weight of 

 85 '5 1, in good agreement with the chemical atomic 

 weight 85 -45. To obtain agreement with the chemical 

 atomic weight of copper 63-57, ^^ is necessary to 

 suppose the isotopes to be 62, 64, and 66, since this 

 gives a mean atomic weight of 63-76, which is as 

 close as would be expected. A direct comparison 

 with rubidium is desirable, but further experiments 

 will be necessary before the comparison can be 

 regarded as conclusive, since the rubidium rays 

 probably start at the surface of the cement and may 

 fall through a different potential from the copper 

 rays. A few comparisons suggested the even atomic 

 weights, so that we may provisionally take the 

 isotopes of copper as of atomic weights 62, 64, and 66. 

 This seems to mark the first exception to the rule 

 observed by Dr. Aston to hold for chlorine, potassium, 

 bromine, rubidium, and antimony, that elements 

 with odd atomic numbers have isotopes with odd 

 atomic weights, and may be connected with the fact 

 that copper occupies a place in the series of elements 

 where the atomic weights begin to increase rapidly 

 with atomic number. A. J. Dempster. 



Ryerson Laboratory, Chicago, 

 June 9. 



Expansion of the Wings of Lepidoptera after 

 Emergence from the Chrysalis. 



No one who has watched a butterfly or moth 

 emerging from the chrysalis can fail to have been 

 impressed by the rapid expansion of the wings. This 

 expansion is not real growth, 

 but merely the opening out of 

 the contents of a carefully 

 packed parcel, and the general 

 character of the changes which 

 ^^^w- ^t' occur in the process is well 

 1^^^ ^ known. 



^^HP The true growth of the wings 



^Hf takes place and is completed in 



membranous sacs just within 

 the walls of the chrysalis, and 

 the form of the wings can be 

 distinguished from the outside. 

 The position of the wings during 

 their development is such that 

 the upper surface of the fore 

 wing is next to the wall of the chrysalis, and within 

 a day or two from the time of hatching the colours 

 and markings can in many cases be recognised. 



Each wing consists of two separate membranes, 



NO. 2801, VOL. 112] 



,^1^ 



Fig. I. — Pupal and extended 

 wings of V. Jo (distance 

 between lines = i inch). 

 The pupal wings were 

 removed from the chry- 

 salis just before emerg- 

 ence. 



united with the nervures, on which the scales ^are 

 mounted, the stems of the scales entering sockets in 

 the membranes placed in fairly symmetrical rows, 

 though the irregular shape of the spaces between the 

 nervures prevents the symmetry being exact. 



The point to which the present note is intended to 

 direct attention is the numerical relation between the 



. i »llf ll.lli 



Fig. 2. — Section of pupal Fig. 3. — Section of pupal 

 wings parallel to the wings at right angles to 



nervures. x 60. the nervures. X 60. 



Sections in Figs. 2 and 3 were cut from the chrysalis, and 

 show both the fore and hind wings. 



size of the pupal and expanded wings, and the reason 

 for the constancy of this relation. In all the lepi- 

 dopterous wings which I have examined the pupal 

 wing has very nearly one-third of the dimensions of 

 the wing of the perfect insect (Fig. i). 



If the fully developed wing is removed from the 

 chrysalis and sectioned, the reason for the one-to- 

 three ratio is immediately evident so far as regards 

 extension parallel to 

 the nervures,' but the 

 " accordion " folding 

 whereby the scale- 

 bearing membranes 

 expand in a direction 

 at right angles to the 

 nervures is rather 

 more complex. 



The section parallel 

 to the nervures is 

 shown in Fig. 2 and diagrammatically in Fig. 6. Here 

 the wing membrane is seen folded so that the distance 

 from fold to fold is the same as the depth of the fold, 

 and therefore the extended is three times that 

 of the folded dimension. To realise the character 

 of folding in the other principal direction, imagine a 

 series of camera bellows fully extended AjAj, etc., 

 to be placed side by side, Fig. g, so that the sides CiC,, 

 C.Co, etc., will remain in contact when the bellows are 



6 c 



Y-mti^ 



Fig. 4. — Section of 

 extended wings 

 parallel to the 

 nervures. X 60. 



1m(, 5 ^> I tion of 

 extendeil wings at 

 right angles to the 

 nervures. X 60. 



These sections are from the posterior part 

 of the fore wing not far from the margin. 



Fig. 6. — Diagrammatic section of Fig. 7. — Diagrammatic section 



pupal wing parallel to the of pupal wing at right angles 



nervures. to the nervures. 



The letters refer to those in Figs. 2 to 7. — (a) Wall of chrysalis ; (b) scales ; 

 (c) wing membrane ; (d) sockets in membrane. 



contracted. Then remove the lower sides B^Ba, etc., 

 and join the free edges of CiC^, C^Ca, etc. It is clear 

 that the surface thus formed is developable, and that 

 if, to start with, the bellows are compressed to one- 

 third of their extended length the developed surface 

 will in all directions have three times the dimension 

 which it has when folded. 



The section of the membrane cut in this direction 

 presents a much more complex appearance (see Figs. 

 3 and 7) than that parallel to nervures. 



The compression to one - third of the extended 

 dimension in the transverse direction appears to be 



