September 5. 1907] 



NA TURE 



477 



Discovery of Seventy-one New Variable Stars. — The 

 wholesale discovery of new variable stars from photo- 

 graphic plates is proceeding at Harvard, and in Harvard 

 College Observatory Circular, No. 130, Prof. Pickering 

 announces the discovery of a further batch of seventy-one 

 new variables. These were found by Miss Leavitt on the 

 Harvard maps Nos. 9, 12, 21, 48, and 51. Prof. Picker- 

 ing gives a table showing the proportion of newly dis- 

 covered variables to the total number now known to exist 

 in each region examined, and arrives at the general de- 

 duction that about one-third of all the variables in the 

 three northern regions examined, and about one-half of 

 those in the two southern regions, yet remain to be 

 found. The designations, positions, and magnitude ranges 

 of the newly discovered variables are given, and the list 

 includes thirteen probable Algol and seven long-period 

 variables, the proportion of the former being remarkable, 

 as in the case of Region 50 discussed in Circular No. 122. 



The Electrical Action of the Sun. — In these columns 

 for March 14 we referred to a discussion, by Dr. Albert 

 Nodon, of the nature and effects of the sun's electrical 

 charge. The whole discussion is now published as an 

 extract from the Revue des Qtieslions scientifiques for 

 April and July, and will be found to be of great interest 

 by all workers on solar physics and the allied terrestrial 

 phenomena. In the first part of the paper Dr. Nodon 

 discusses the observations, the apparatus used in making 

 them, and the theories deduced from them. The second 

 part contains a discussion of the application of the results 

 to the explanation of cometary, planetary, and terrestrial 

 -phenomena, whilst in the third part of the paper the 

 author discusses the deductions relative to terrestrial 

 physics. The paper is pu"blished by J. Polleunis, 45 rue 

 Sans .Souci, Brussels. 



Micrometer Measures of Double Stars. — In No. 4193 

 of the Astronomische Nachrichten (p. 277, July 26) Dr. 

 H. E. Lau publishes a further list of Struve double stars 

 measured by himself, and discusses the mean probable 

 errors of his measurements. In addition to the eighty 

 measures made by Dr. Lau, the list also contains twenty- 

 eight measures made by Herr Luplau-Jannsen. 



THE MAY OR GORSEDD YEAR IN ENGLISH 



AND WELSH FAIRS. 

 CIR NORMAN LOCKYER has taught us to call the 

 '~^ year indicated by alignments of stone monuments in 

 Britain the May year. The quarter days of that year are 

 astronomical, being the half-way stations of the sun 

 between the solstices and equinoxes. In fixing these dates, 

 of course, the solar quarter days were marked as well, 

 which year is conveniently called the solstitial year. It 

 will clear the way for the discussion of some figures bear- 

 ing on the subject if the two series of quarter davs are 

 presented here side by side, as given in " Stonehenge 

 Astronomically Considered," p. 23 : — 

 May Vear ... Feb. 4 May 6 Aug. 8 Nov. S 



Solstitial Year ... March 21 June at Sept. 23 Dec. 23 



The quarter in both series is of the same length, ninety- 

 one days, and the distance from a solstitial quarter day 

 to a May-year one is roughly forty-five days. 



Though the name May year is a very happy one, as 

 the May festival was certainly the most popular, it is 

 really the Gorsedd year, the very raison d'etre for that 

 institution which, in form, purpose, and ritual, is the 

 temple-observatory brought up to date. We know now 

 for 'vhat purpose the megalithic monuments were raised, 

 and that knowledge has been acquired bv working from the 

 known to the unknown. By assuming that the Welsh 

 Gorsedd is a much truer representation of ancient 

 Druidism than the manifestly inaccurate, second-hand 

 observations of Caesar and other classical writers, we are 

 able to see at the Welsh National Eisteddfod in this 

 twentieth century the actual use to which the temple- 

 observatory was put. If such a broad assertion causes 

 surprise, that surprise is considerably lessened by what 

 seems to me to be an incontrovertible fact, that, instead 

 of having one Gorsedd, and that in Wales, a true survival 

 from late Neolithic times (to fix an indubitable downward 

 limit), we have in Britain more than one thousand Gorsedds 



N >. 1975, VOL. 76] 



the pedigrees of which are as unimpeachable as that of the 

 Welsh institution. I refer to fairs still held on the 

 quarter days of the May year. To a student of the Welsh 

 Gorsedd this fact at once dispels any a priori doubt as to 

 the antiquity of that institution. It is only one among a 

 thousand, though, I would maintain, it is the only one 

 that shows what all the others were at first. 



The Gorsedd and the popular fair are one and the same, 

 constituting a true monument as ancient as a temple- 

 observatory in stone. A better way of putting it is, the 

 temple-observatory has survived in (i) stone, in (2) tradi- 

 tion, and in (3) festival. The Welsh Gorsedd presents this 

 triple evidence. 



There is, I think, no need for a formal proof of the 

 prevalence of the May year in Ireland, Scotland, and 

 Wales, or the "Celtic fringes." It reigns supreme over 

 still purely Celtic ground. It is when one comes to 

 England proper that even one accustomed to mark time 

 in May-year terms must confess to a feeling of surprise. 

 The evidence from the Celtic fringes is, of course, indis- 

 pensable to understand and explain the English May-year 

 fairs, but a brief presentation of the English case may be 

 helpful by way of enlisting the cooperation of English 

 archaeologists to make that case as strong as possible. 



I take Owen's " New Book of Fairs " for 1S24 as 

 source. The book was published by Royal licence, but 

 as regards Wales it is incomplete, and I would infer as 

 much as regards England. The following figures, except 

 those given by counties, include the Welsh fairs as given 

 in the list. That inclusion cannot affect seriously the 

 English case, as will be seen. 



The relative popularity of the May and of the solstitial 

 years may first be ascertained by comparing the number 

 of all fairs in May with those in June. May fairs, 510; 

 June fairs, 250; 2 to i for May. 



There are two lists of fairs in Owen's book, one by 

 counties and the other by dates. I take the latter first. 

 The figures in every case are my own. .As the book is 

 incomplete, and all lists of fairs I have consulted are so, 

 I have thought it sufficient for the present purpose to 

 make only one rapid reckoning of the fairs. The chief 

 fair days can be easily noted by large groups of fairs. 

 The fairs corresponding to the May-year festivals are to 

 be looked for under several dates. The astronomical day 

 is in many cases observed eleven and twelve days later. 

 Generally, that day has given way to the first of the 

 calendar month. In both cases new and old style dates 

 must be noted. Then there are fairs depending on such 

 dates. .^11 fairs held during the first twelve days of the 

 month should be numbered. In the case of November, the 

 inclusion of Martinmas fairs needs no comment, as 

 .N'ovember 11 is a Scottish quarter day, and the Scottish 

 quarter days, with the fact that in Gaelic-speaking Scot- 

 land the months, as well as the seasons, are still reckoned 

 in the true May-year order, is sufficient formal proof of 

 the predominance of that year on Celtic ground. 



241 



D. = Dependent fairs. 

 Thus we have 873 plain May-year fairs. I claim now 

 the Church-year fairs, which are plainly the old May-year 

 festivals. For February 4 I claim 28 fairs between 

 Candlemas and the beginning of Lent ; for May 6, 358 

 Whitsun and .Ascension fairs ; and for November 8, 53 

 fairs at Michaelmas, 71 on October 10 (Old Michaelmas), 

 and 32 on December 11 (Old St. Andrew's Day). Though 

 -Michaelmas and St. Andrew's Day are both a month 

 away from November i, they constantly occur as half-year 



