478 



NA TURE 



[September 5, 1907 



^lays corresponding to May Day. It is very likely, how- 

 «ver, that some of the August fairs have been absorbed 

 into Michaelmas. As that day occurs so near to the 

 autumnal equinox, some concession must be made also to 

 the solstitial year. There is no need, however, to decide 

 these points at present. 



We can now add 542 to our list of May-year fairs, 

 .altogether 1415 fairs which may reasonably be claimed as 

 ■so many Gorsedds or prehistoric monuments, 96 in 

 February, 717 in -May, 205 in August, and 397 in 

 November. It is curious to note that the number of plain 

 Mav fairs and of Church-year May fairs is the same. 

 The ratio for November seems to be too high, and the 

 number for that month has grown evidently at the expense 

 ■of August. Dividing the total for August and May, we 

 get 301 for each of those months to match the figure for 

 May, which is always at least twice as high as the corre- 

 sponding figures. For obvious reasons February is a poor 

 time for fairs, and the intrusion of Lent has very generally 

 broken up that end of the May year. A more thorough 

 scrutiny will be the means of recovering many February 4 

 fairs. 



If the above estimate is considered too generous, my 

 estimate of the solstitial-year fairs must err more in that 

 pleasing direction, for I include, against strong reasons, 

 all the Easter fairs in that estimate. 



Sept. 19 

 ,. 21 

 Oct. 2 .. 



Fairs 

 31 



104 



December 21 

 2S 



'5 



There are, then, 276 true solstitial fairs and 347 Church- 

 year fairs to bring the total up to 623 ; but the figure for 

 Easter shows evident borrowing from February, the 

 vernal equino.x, and May. A fair ratio would be obtained 

 by counting sixty Easter fairs for the vernal equinox 

 and the remainder for February. 



In claiming the Easter fairs for the May vear, I have 

 a larger number of fairs to add to the solstitial estimate. 

 Beside the May year, with the portions of the Church year 

 which are evidently based on it, and the solstitial year as 

 such, there are two other series of dates to consider. The 

 one I would call the Roman year, being important dates 

 in the old Roman year, which were early associated with 

 the names of Christ, St. Mary, and at least six of the 

 Apostles. I refer to groups of fairs on the 25th of the 

 month. There are 25 fairs on March 25 and 43 on 

 September 25, 68 fairs which I would add to the solstitial 

 estimate. 



The other series of dates I would call the Petrine year, 

 with groups of fairs on the 2qth of the month. When 

 the old Celtic saints of Llandaff Cathedral, Teilo and 

 Dyvrig, May-year saints, were superseded in Anglo- 

 Norman times by SS. Peter and Paul, June 29 was fixed 

 as the beginning of the year in that cathedral. The 

 canons there still mark their time of residence as from 

 that date. The Petrine year is fairly general, though it 

 is altogether subsequent to the middle of the twelfth 

 century a.d., at any rate in South Wales. There are iS 

 fairs on March 29, 35 on June 29, 45 on July 10 (O.S.), 

 S3 on September 29, and 71 on October 10 (O.S.), in all 

 251 fairs which I would add to the solstitial estimate, 

 which now stands at 948 fairs. 



The May year is still 467 ahead, and the May-year 

 figures are certainly much more satisfactory than the 

 large figures I have juggled for the solstitial estimate. If 

 the latter is fairly correct, I must now add it in a lump 



NO 1975, VOL. 76] 



to that of the May year, and say that we have still in 

 England and Wales 2363 fairs, relics of festivals held at 

 the same spots or thereabouts when the dates were 

 obtained by direct solar observations by means of aligned 

 monuments. Several capable archjEologists have expressed 

 the opinion, anent the astronomical theory, that they admit 

 the solstitial alignments, while doubting the very existence 

 of the May year in connection with the monuments. 

 Such admission is substantially complete. The solstices 

 and equinoxes were of little direct practical use to the 

 ancient farmer as dates to commence farming operations. 

 The Welsh farmer of to-day is finelv indifferent to the 

 almanac statements that spring begins on March 21 and 

 winter on December 23. He knows better. The solstitial 

 quarter days were observed as points from which the 

 infinitely more practically important May quarter days 

 could be correctly marked. 



So far I have made only a rough preliminary recon- 

 naissance of the list of fairs. I now take up Owen's list 

 of fairs by counties, not to learn more about the relative 

 prevalence of the May and the solstitial year so much as 

 to learn more concerning the May year itself. E.xcept 

 May 6 and November 8, I have counted all fairs in 

 February, May, August, and November. From the 

 Church year I have borrowed only some pre-Lenten fairs 

 for February and the Whitsun fairs for May. The total 

 is just the same, but the analysis is much more to the 

 point. The Welsh fairs are included, though the list is 

 very defective. The letters F.M..'\.N., variously grouped, 

 represent May-year quarters : — 



