452 



kno\\li:i)gp:. 



December, 1912. 



the sections caused li\' tlie partinl ilistriution of tlic 

 coast-line heartlis l>y tlie sea) it is found to consist 

 of a layer, not more tiian three or four feet thici< 

 and frequently much less, of very angular stones 

 packed in dark soil mixed with much charcoal and 

 charcoal-dust. The stones, rarely larger than pieces 

 broken for road metal, are burnt black, yellow or 

 bright brick-red in colour, 

 and obviously are frag- 

 ments of larger blocks of 

 sandstone, quart/ite or 

 other hard rocks which 

 have split up after fre- 

 quent heating and sudden 

 cooling. Many of tin- 

 mounds have been almost 

 destroyed under the 

 plough, others are mere 

 lines of "pot-boilers." 

 while others still contain 

 several tons of burnt 

 stones. 



Of the six hearths founil 

 by me near the shore. 

 the three at Swanlakc 

 and one described below , 

 are all good examples and 

 quite typical, but the last 

 is of more than ordinary 

 interest. 



The hearth at Marros 

 (see Figures 474. 476, 

 and 477) lies twentv- 

 tive yards from an 

 excellent S[)ring which 



issues about twenty or twL-nty-five feet above the 

 shore. .\ streamlet trickles past the hearth and 

 now oozes away below the storm-beach which, since 

 the occupation of the cooking-place, has been banked 

 fullv twenty feet above the shore by winter storms 

 and has not only blocked the course of the little 

 stream but has 

 almost completely 

 buried the mound. 

 On the seaward 

 side, however, 

 where a recent 

 storm has torn 

 away a large seg- 

 ment of the hearth, 

 a clear section is 

 exposed down to 

 the underlying 

 yellow drift clay, 

 layer three feet 



Ficrui- 475. A K'roup of five Iie;irtlis (indicated by tin 

 eiiclcs) on Warren Corse, near C'astleniartin, PeiiibroU' 



where there is a row. six feet long, of large blocks of 

 sandstone: this ma\- be a remnant of a low retaining 

 wall. 



Of relics, which might throw light on the period 

 when this cooking place was used, it may be said, in 

 the words of Dr. Johnson, that " the negative cata- 

 logue is very copious," for no traces of jjottery or other 

 rooking vessels, bones, 

 shells, stone or other 

 implements, could be 

 found. The remark is 

 equally true (with two 

 possible exceptions) of all 

 the hearths yet examined 

 but very few have \et 

 been carefully searched. 



The exceptions are at 

 Swanlake, where from 

 each of two of the hearths 

 I obtained a flint Hake 

 amongst the burnt stones 

 and charcoal. No other 

 clues to the age of the 

 hearths have as \"et been 

 found actually in any of 

 them, but Messrs. Cantrill 

 and Jones picked up 

 several flakes on a mound 

 broken up by the plough, 

 and they record the 

 occurrence of flakes and 

 a core near a spring 

 adjoining a hearth at 

 Westhook. During the 

 year 1911 I found flint 

 flakes and cores in the vicinity of other hearths : 



(1) Near the spring at Marros. 



(2) In rainwash near a hearth at Manorbicr. 

 (j) In rainwash near a hearth at C"\\m Mawr. 



small 

 shire. 



Streara 



In rainwash near 

 Newgale. 

 (4) In rainwash. throe feet decj 



Recent lilowivsaTicl and .sl\iT\glt 



FlGfKK 476. 



ill) = Yellow stony clay 



The true hearth consists of a 

 thick of rather small angular 

 pieces of sandstone, grit, and quartzite, (all from 

 the local rocks), burnt and discoloured by fire and 

 closely packed in dusty charcoal. It is noteworthy 

 that none of the burnt stones seem to be fragments 

 of beach pebbles, although the whole mound is now- 

 surrounded and covered by such pebbles. .\ quite 

 exceptional feature appears on the seaward side 



). a few \ards from 

 the hearths 

 at Swan- 

 lake. 

 (5) On a hearth 

 on Warren 

 Corse. 

 It is not probable 

 th.it in all these 

 itistances the asso- 

 ciation betw een the 

 hearths and the 

 ll i n t flakes i s 

 accidental ; least of all at .Swanlake. w here flakes 

 occur not only (1) in the hearths and (2) in rainwash 

 very close to them, but also (i) on a chijiping-floor 

 on the cliff-top above the hearths. It seems fair to 

 infer here a connection between the cooking-places 

 and the presumably Neolithic chipping-floor. 



Usually the mounds occur singly: but a few groups 

 of two or three have been noted. The largest 

 number yet observed in close association is a group 



niai;raiuni.atic section of a prehistoric cooking place at 

 Marros, Carmarthenshire. 



^.>.,) = Storm beach. 



