August 25, 188 1] 



NATURE 



393 



Third stage . . . 



Fourth stage ... 

 Fifth stage 



Continuous absorption at the blue 



I end not reaching to the less refran- 



I gible end. (This absorption may 



break up into channelled spaces.) 



Continuous absorption at the red end 



I not reaching to the more refrangible 



t end. (This absorption may break 



up into channelled spaces.) 



Unique continuous absorption. 



So that the story is one of absolute conformity, absolute con- 

 tinuity from one end of the series to the other ; but on this subiect 



I need not say more, because my friend Capt. Abney will have a 

 great deal to say about the red molecules and blue molecules 

 when he comes to deal \vith the red end of the spectrum, and I 

 may safely leave this part of the subject in the hands of one who 

 has so brilliantly distinguished himself by his investigations 

 upon it. 



Replies to Objections 

 Now I think it is time that I should reply, or attempt to 

 reply, to some objections that have been made to these views. 

 So far as I can gather, the serious objections which have been 

 made are not many, but some of them are objections to which 

 considerable value should be attached. The chief one now 



Inte r- \ 

 mediate 



Inter- \ 

 med ial;; 



tciiiperature. 



Fig. 44. — Diagraai showing the action of three different temperatures on a hypothetical substance assuming three stages of complete dissociation ; and also 

 of intermediate temperatures at which the vapours are only partially dissociatett 



urged is that one is "misled in the conclusions that one draws 

 from these observations of spots and storms by the fact that the 

 solar lines corresponding with the lines which we consider to 

 be common to two substances are really double, and that 

 the lines common to twj substances appear common simply 

 because we have not sufficient dispersion to separate them. Now 

 that is a very important objection indeed, but let us examine it. 

 It has been pointed out that of the 62 iron lines which re- 

 mained as the result of the purification of the first part of the 

 map (between wave-lengths 39 and 40) only iS were left ; all 

 the rest being found common, not only to two substances, but in a 



great many cases to four or six substances, and we found also 

 that our rough observation-book, as we went on, suggested that 

 the solar line was double ; but if we had gone on in that way we 

 should not have been able to produce a map at all, because there 

 would have been few lines which were not complex, so that it 

 would have been a piece of cowardice to remain there and not 

 attempt to find out what it meant. Now let us suppose a great 

 many of the solar lines are double. It is fair to assume that 

 these double lines would be irregularly distributed throughout 

 the spectrum. We cannot imagine some spiteful freak of nature 

 choosing out to be double a particular set of solar lines which 



Fig. 45. — The ph.itagra] 



some one should eventually find common to two substances ; we 

 must imagine an impartial distribution of double lines if we are to 

 consider them as double. Now the argument we can bring 

 against that is foun led on this, that two things always hang 

 together, the alleged complex nature of tlie line and the fact that 

 this particular line is picked out for special prominence in spots 

 and storms. For instance, take the line 4923-2 referred to 

 in Fig. 41. If that hue is double, and is one of two lines 

 seen in flames, the probability that that line should be double, 

 supposing that the solar line is double, would be as 2 to i, but 

 that line is picked out as 50 to I in the case of prominences. 

 The betting in favour of the coincidence is not so great in the 

 case of the spots, but when we come to the prominences, when 

 wc are dealing with 100 lines with the probability therefore of 



iipounds. 



50 double lines if they are equally distributed, and that every 

 other Ime is double — when we come from 100 lines to I, it is JO 

 to I against this particular line being double ; and yet this is the 

 particular line which we always find to be common to two sub- 

 stances when we really discuss the observations of the flames. I 

 think then that the idea that these basic lines are simple crea- 

 tures of the imagination, simply chance coincidences, will really 

 not stand at all. 



Prof. Young has lately brought forward this objection, 

 although in 1872 he was the first to point out the very extraor- 

 dinary fact as it appeared to him, and as it still appears to every 

 one else, that an enormous number of coincident lines which he 

 got from the tables extant at that time seemed to cluster round 

 the bright lines seen in his observations. The credit of that is 



