THE TRAPPEAN ROCKS. 33 



ant un total de 688, ou, en nombre rond, de 700 dykes de 

 trap dans la totalite" de 1'ile de Arran. Doublant meme 

 encore ce nombre si Ton vouloit, pour y comprendre tous les 

 dykes caches par les bruyeres vastes et etendues dans 1'in- 

 terieur, par les greves de sable sur les rivages, ou places dans 

 des recoins inaccessibles des ruontagnes, on n'arriveroit pas 

 encore au nombre de 1500; et pourtant en parlant de telle 

 ou telle cote, de telle localite d' Arran, il est souvent echappe 

 a ceux des geologues qui ont decrit Arran, a moi-meme peut- 

 etre tout le premier, de dire qu' on y voyoit des innombrables 

 dykes de trap. Or, je crois avoir maintenant montre que 

 loin de ne pouvoir etre compte"s, on peut a present concevoir 

 1'esperance de voir chacun des dykes de cette lie individuelle- 

 ment etudie, nume"rote, decrit et enregistre dans un catalogue 

 descriptif et raisonne, analogue a celui que j'ai aujourd' hui 

 1'honneur de mettre sous les yeux de la Societe Royale." * 



Mr. James Napier, of Glasgow, has published a short 

 paper on the dikes between the bays of Broclick and Lamlash 

 in the Edinburgh New Philosophical Journal, New Series, 

 vol. ii. No. 1, July, 1855, accompanied by a map, on which 

 the dikes are laid down. He reckons altogether fifty-four 

 dikes as visible along the shore, but considers that not a few 

 may have escaped his notice. " Struck," says Mr. Napier, 

 "by the large number of trap dikes cutting through the 

 sandstone, in a direction at right angles to the sea-line, it 

 occurred to me that if such dikes continued round the coast 

 to Lamlash, and still at right angles to the sea-line, they 

 must in all probability have proceeded from a common 

 centre, lying somewhere between the two bays." To test 

 this idea by observation, he measured and marked down the 

 position of every dike, and the result confirmed his " antici- 

 pation, that they proceeded from one, or possibly from two 



* Documents sur les Dykes de Trap d'une Partie de File d' Arran. 

 Transactions of the Royal Society, Edinburgh, vol. xiv. Part 2, 1840, 

 p. 684. The paper is a model of patient and generally accurate re- 

 search. 



D 



