September 12. 19.18 



NATURE 



well in fronl o( the lattei I in explanation was, 

 however, simple I he beam was nol an aurora beam 

 at all, but part of the 22° halo with the horizontal 

 mock moon. The illusion was, however, ver) strik- 

 as there were vertical beams ol auro 

 id abovi the halo beami" ( Kf 

 I have given above the actual words -. 

 the time, as it is dangerous to describe an observa- 



several yeai s lab 1 , 1 spei ially if a mat b 

 opinion is concerned I feel compelled, however, 

 few further remarks. 

 With regard to (a) ii should be realised that as 

 . outlines ol the mountain were 

 indistinct, and the bands of light due to the 

 moon shining on the shoulders won- ver) faint. 

 Normally, he] could scarcely bi seen, but when an 

 beam above the mountain coincided with a 

 band, the eye was guided down from the beam to 

 the band of light, which then attracted the atti 



.1 continuation of the aurora beam. 

 \ cloud-cloak similar to that referred to in (b) 

 observed to cover the top of Erebus, especially 

 a blizzard. It fn near to the 



mountain, anil at night its outline would not be dis- 

 tinguishable from that of the mountain. It was onlj 

 when the gap between thi mountain anil the cloud, 

 through which the aurora app d that 



the pn Hi itself was suspei ted. This 



in excellent illusion, and nine casual observers 

 . \n of ten would have been convinced thai the aurora 

 irly visible in front of the mountain. 

 The third illusion was p most impres- 



parenth obsi th< Scott Antarctic Expedi- 



tion of 1911-iB are ingenious. But in two cases it 

 is noi explicitl) stated whether the original observe! 

 accepted the proffered explanation, and in the remain- 

 ing case it appears that, with Dr. Simpson's explana- 

 tion before them, the majoril) the observers 

 lined of their original opinion. sir Douglas 

 Mawson's list included a greater numbei of apparent 

 cases of aurora at low altitudes, bul « hether the 

 observers possessed the exact shad cepticism 

 desirable in observers in such a case I am 1 ible to 

 say. 



M\ own view is that ... ' are not as yet in 1 . oti 

 to deny the possibility of the occurrence of aun 

 low levels near the magnetic poles. It is • ■ 

 that the observers of the next Arctic or An 



expedition should lie familiar with what has 

 written on the subject, and that they should 

 specially careful in dealing with any apparently lov 

 level aurora. Also observation of auroral heights bi 

 photography after Prof. Simmer's method should be 

 a fundamental part ol the programme of any such 

 expedition; and while .1 long base should be usi 

 seme of the observations, others should employ .1 base 

 sufficiently short to deal satisfactorily with hei^l 

 only a few kilometres C. Chres. 



Vugusl 17. 



sive. The mock moon, with part of the vertical 

 cird was formed in 1 mist, it-. If quit; invisible, 

 which lay over the foothills of Erebus. The moon 

 was shining in a clear sky without any sign of a halo. 

 none bul a trained observer would have connected 

 this detached beam of light with the moon. It ap- 

 part ol the aurora display 

 which was taking place apparent!) in its immediate 

 neighbourhood, and the conclusion reached b\ Meares 

 thai it was a part of the aurora was very natural. 



In mi paper on "Atmospl ElectrichA in High 



Latitudes" (Phil. Trans., \. vol civ., p. 92, 1905), I 

 notha 1 isi ol an illusion in Lapland, w hi< h 

 made the ai rni ath low -King clouds. 



Thus both in the north and in the south I have seen 

 - in which the mrora appeal d to be in the lower 

 atmosphere, and in all these cases ver) careful observ- 

 ing \, ; illusion. Tli< mere 

 nent, there/on . thi i has been seen 

 isi rvi is and a m ai object cannot be 



p ool of tin peni ol the aurora 



into the lower atmosphi G. C. Simpson. 



Nfeteorological Office, Simla, India. June 1;. 



Electricai phenomena resembling aurora have 

 hitherto been observed in the laboratory only a 

 low atmosp ! ' mina- 



tions of the altitude of aurora b) Prof. Stormei 



i. w heights so low as 50 km. Thus 



rations of aurora bi low the summi 



mountains are naturalK viewed with suspicion b) 



physicists. The explanations given by Dr. Simpson 



of'the three cases he describes the onh ones ap- 



2550, VOL. 102] 



Hybrid Sunflowers. 



Ix crossing the different species and varietii 

 llilianthus some peculiar results have been obtained. 

 The crosses referred to have all been made by my 

 wife at Boulder, Colorado, and the results may be 

 classified as follows : — 



(1) The varieties of Helianthus annum (including 

 //. lenticulariSi regarded by some botanists as a dis- 

 tinct species) when crossed together produce p 

 which are as fertile as the parents. In some of the 

 mongrel varieties there is, how T ever, a marked 



v of pollen. 



(2) The annual species of sunflowers (typical Hi 

 thus) crossed together are quite fertile, but the h 

 are themselves nearly sterile. H. annuus has 

 crossed with three species, H. argophyllus, II. petio- 

 laris, and //. cucamerifolius. 



(3) The annual species can rarely be crossed with 

 the perennial, and when this occurs the offspring 

 closely resemble one or the other parent species. One 

 such hybrid was recorded in the " Standard Cyclopedia 

 of Horticulture" (vol. vi., 1917, p. 3281) as between 

 //. pumilus and H. annuus. Renewed study of the 

 living plants this year convinces me that this is an 

 error; the perennial parent was, in fact, H. we - 

 rhomb oideus. Both species occur here, and Mrs 

 Cockerell, at the time of making the cross, did not 

 distinguish between them. Morphologically they are 



iall) distingui x I b) the fact that H. sub- 

 rhomboideus has underground migratorv branch- 

 on ails of which it spreads, while II. pumilus is 



stricti) reproducing only by seed. The 



isel) resembles H. subrhomboideus 

 (though this was the pollen parent), but is much 

 with larger broad leaves, has small or • 

 : bram ties, but, nevertheless, is sta 

 . the migrators are present, bin thi 

 spread b) them in all direction- as do the 



migratory forms. Comparing the detail 1 struc- 



1 Found that the ray-fton is of the h i were 



eh. mi pistils, win reas these wen ' loped 



(though not functional) in the II. subrhomboideus. 



However, further investigation showed, to my sur- 



. that some head- of the wild H. subrhomboideus 



he ray-florets whollv without pistils. The involu- 



