April 1, 1891.] 



KNOWLEDGE 



71 



of the week under the required month. The days of the 

 week are now opposite their proper days of the month. 



III. — If the year is given in thick figures it is a Leap 

 Year, and tlie -tanuary and February in thick letters must 

 be used ; in other cases the plain .January and February 

 must be used. 



IV. — The table may be used for any other centuries 

 (N.S.), the calendar being repeated every fourth centixry. 

 Thus 1600-1699 is identical with 2000 ; 2100 with 1700 ; 

 2200 with 1900, &c. 



V. — For Old Style dates, in the first motion of the disc, 

 the day of the week given in the following table must be 

 set opposite to the given year ; the second motion being 

 the same as in New Stj'le dates : — 



N.B.— That the years 1600, 1700, 1800, &c., are all 

 leap years, O.S. ; but only every fourth of them, e.y. 

 1600, 2000, 2400, &c., are leap years according to N.S. 



THE MAGIC SQUARE OP FOUR. 

 To the Editor of Knowledge. 

 Dear Sib, — For the sake of accuracy, I should like to 

 point out a mistake in my estimate of the number of 

 varieties of type D. The types A, 1> and D are mutually 

 convertible by a few simple transpositions, and therefore 

 must have the same number of varieties. Now it is 

 mathematically demonstrable that there can be only 48 

 varieties of A and B ; hence there must be just 48 of D. 

 How it happened I wrote 96 I do not know. 



On the other hand, Mr. Cram tells me he can make 56 

 each of G, I, .J, and L. Consequently, we are still 32 

 short of Freniele's total of 880. 



T. S. Barrett. 



To the Editor of Knowledge. 



Dear Sir, — In reading Mr. -J. Pentland Smith's very 

 interesting article on " Contrivances for the Cross-Fertili- 

 zation of Plants," , in Knowledge, Feb. 2,1891,1 have 

 been completely puzzled by the following problem : 



In speaking of the case of Aspidistra elatior, Mr. Smith, 

 either in his own words or quoting Dr. Wilson, says in 

 effect : — 



(I.) There is no access to the pollen until the stigma is 

 fertilized. 



(II.) Fertilization cannot be effected until pollen has 

 been deposited on the stigma. 



>Vluit I want to know is : How is the first flower fer- 

 tilized, so as to commence the process ? 



Yours faithfully, 



F. .]. Pkovis. 



Coleford, Gloucester, 16th March 1891. 



[Until the decay of the stigma the slugs cannot get to 

 the pollen. One can easily imagine that sooner or later 

 the stigma of the first flower of the season will decay, 

 whether pollenated or not, and that then access can be 

 obtained to the pollen below. Pollenation, that is the 

 resting, and ultimate germination of the pollen on the 

 stigma, precedes fertilization in all Angio-sperms or plants 

 wliicli have their seeds enclosed by the carpels. — J. P. 

 Smith.] 



STELLAR SPECTRA. 



By E. W. Maunder, F.E.A.S. {Assistant superintendiiu/ the 

 Spectroscopic Department of (ireeninch Observatory). 



WHEN the publication of Dr. Elkin's determina- 

 tion of the parallax of Arcturus rendered it 

 probable that we must class this star as one 

 of the most distant of its magnitude, it 

 became clear to me, as I tried to show in the 

 February number of Knowledge, that it must be of most 

 gigantic size, and must move with most amazing swift- 

 ness. But a fiu'ther point for inquiry also suggested itself. 

 If Arcturus be, as would appear to be the case, the star 

 which actually gives the most light of any we know of at 

 present, then its spectrum should be typical of the largest 

 and hottest stars, whereas it has been customary to regard 

 it as of a markedly lower class. 



The subject of the classification of stellar spectra has 

 been occupying special attention of late ; it may, there- 

 fore, be worth while to see if any further light, however 

 feeble, can be thrown on the matter from the point of view 

 suggested by the magnitude of Arcturus. 



The earliest classification of stars was effected by Sir 

 W. Herschel, who grouped them according to their 

 colours ; the only arrangement possible at a time when the 

 spectroscope was still undreamed of, but a real step, never- 

 theless, towards the more delicate discrimination which 

 that instrument renders possible. Fraumhofer, in his en- 

 deavours to solve the secret of the spectrum, noted the 

 strongly marked diflerences between the spectra of diflerent 

 stars, and drew the important conclusion that the dark 

 lines which crossed them were due to something in the 

 stars themselves, and not to any effect of our own atmo- 

 sphere, or any absorption of light in space. But the first 

 spectroscopic classification of stars was due to Rutherford 

 in America, and Secchi in Em-ope. The latter, as was 

 inevitable in beginning so new a research, made several 

 alterations from liis first scheme ; but his final classification 

 divided the stars into " types " of spectrum as follows : — 



Type I. — The white and bluish white stars, like Sirius 

 and Vega. The spectra in this type show the four lines of 

 hydrogen intensely dark, broad, and with diffused and 

 shaded edges. Metallic lines are faint and narrow, and 

 not easily seen. The principal stars of Orion form a 

 variety of this type, in which the hydrogen lines are much 

 less marked, and much narrower than in other stars of the 

 first type. 



Type II. embraces the yellowish stars, such as our Sun, 

 Arcturus, and Aldebaran. The hydrogen lines are well 

 seen, but narrow and fairly sharp ; the entire spectrum is 

 full of well-marked metallic lines, some of which are more 

 pronounced than the lines of hydrogen. 



The stars of Type III. arc mostly orange in hue ; 

 o Ononis and a Herculis are the best examples. The hydro- 

 gen lines are faint, or no longer seen ; but a succession 

 of dark bands, dark and sharp towards the violet, and 

 shading away into nothingness towards the red, makes this 

 type of spectrum the most strongly marked and the most 

 beautiful of any. 



Type IV., the stars of which are mostly red, also shows 

 a banded spectrum, but the bands fade off in the opposite 

 direction to those of Tj-pe III. 



Secchi further called attention to the existence of a 

 couple of stars showing briijlit lines in their spectra, the 

 forerunners of a fifth type ; and, a little later, the dis- 

 covery by MM. Wolf and Rayet of a curious group of 

 stars in Cygiius added a sixth type. 



Secchi's classification was purely an observational one, 

 and it was independent of theoretical considerations as to 



