^{page 42 note 1} is to be taken in its ordinary sense of “bringing close together”. It is thus practically synonymous with .

^{page 42 note 2} My conjecture of is confirmed by Professor E. H. Parker, who says that it is quite the ordinary “grass”.

^{page 43 note 1} My father is of opinion that the Chinese cannot yield the meaning which I adopted first, namely, “all are freely accessible from the outside.” On the other hand, I have ascertained from M. Pelliot that there is no internal communication between the grottos themselves. His letter, however, which reached me just too late for insertion in the original article, suggests the true solution of the difficulty: “Pour la question que vous me posez, il va sans dire que j'ai sur l'aménagement des grottes de Touen-houang des souvenirs visuels et documentaires fort précis. Il y a plusieurs centaines de grottes, et il n'y a pas de passage intérieur de l'une à l'autre. Mais pour les grottes qui n'étaient pas au niveau même du sol, beaucoup étaient réunies par des galeries, des balcons parfois couverts et dont certains subsistent; vous en avez probablement des spécimens dans certaines des photographies de Stein. Presque tous les balcons subsistants sont très anciens; il en est du x^e siècle.”

^{page 43 note 2} The two characters “ten” should, I think, be deleted, which would reduce the hill to the more reasonable proportions of 8 × 4 li.

^{page 43 note 3} I cannot quite accept Mr. Hu's assertion (iv, 2) that no stroke has been omitted in . That at least one stroke is wanting seems to me as plain as a pikestaff. But I am inclined now to believe that the character was so written simply as a semi-cursive form, and not because it was taboo.

^{page 43 note 4} There are two reasons, according to my father, why must be plural here: (1) the natural meaning of is “among” or “in between”, as seen in , p. 13, col. 4; (2) the words , just below, can only mean “all clamber up some high peak” (not “to the summit”), implying that there is more than one. The latter argument seems pretty conclusive; but as regards my former rendering of .(on the hill), I can point to a similar use of the word in the Liao Chai, Tan Ming-lun's edition, chüan 1, f. 10 v^o, col. 6: “the old priest took some scissors and cut out a circular piece of paper like a mirror, which he proceeded to stick on the wall”.

^{page 44 note 1} is evidently a stock phrase for tapering mountain peaks. I have just come across it again in the preface to .

^{page 44 note 2} The word , as my father points out, can hardly be a well here, though the Sha chou chih has the gloss . It is simply a mysterious hole, such as our mediaeval writers have termed a, cunnus diaboli.

^{page 44 note 3} I have adopted MrHu, 's correction (i, 1)Google Scholar, but though he is doubtless right in saying that is a verb, and that closes the sentence, I do not feel quite so certain about . It is a fact that the name “spiritual sand” was applied to the hill. See Ta, Ch'ing I T'ung Chih, ch. 170, fol. 4 r^o, col. 1: . Professor Parker has also pointed out my mistake with regard to , but he goes on to say: “I don't think it will be possible to find anywhere, at any date, an example of yen being followed by anything but a , and (as I showed) it seems always to = the French en or y.” In reply, I must confront him with his own words in the China, Review, vol. xxiv, p. 260Google Scholar: “Finally, yen occurs in a medial position between two parts of one idea.… For instance, ‘and [they said] there was no one to prevent him’: neminem quidem preventurum. Yen here has the force of quidem.”

^{page 45 note 1} , the reading proposed by MrHu, (iii, 1)Google Scholar, appears to me a doubtful and unnecessary conjecture. In any case, I have to deal with the text as it stands, and there can be no doubt that the character written by the copyist is . It is used again in the same figurative sense on p. 13, col. 5.

^{page 45 note 2} I have now come to the conclusion that the character which I first took to be is really .

^{page 46 note 1} It is an almost diabolical coincidence, from the translator's point of view, that this spirit-dragon (shên lung) should have been slain in precisely the shên-lung period.

^{page 46 note 2} Professor Parker also suggests instead of , and takes exception to my statement that is a vulgar form of ; but my authority is K'ang Hsi's Dictionary, which further states that this form was originally a variant arbitrarily introduced into the “clerkly style” of handwriting by Yen Chên-ch'ing (A.D. 709–85).

^{page 46 note 3} My father points out that is in this context not chien ⁴ but hsien ⁴.

^{page 46 note 4} More literally, “laid [an arrow] on the string.”

^{page 46 note 5} MrHu, (i, 5)Google Scholar is very severe on my punctuation here, although the sense of the passage remains unaffected. Indeed, in the English it is better to put the stop after “trees”, which in Mr. Hu's version seem to spring from nowhere. Much more important is the fact, noted by my father, that is here “a mound” and not “a fort”.

^{page 47 note 1} Mr. Hu has certainly solved the difficulty here (iii, 2). I had already received the same correction from Mr. Edmund Backhouse, of Peking, who has had considerable experience of Chinese MSS. “How often,” he says, “have I been rebuked by scholars for scratching a character out instead of keeping the page tidy by re-writing it and adding the to show that the wrongly written one was to be passed over.”

^{page 47 note 2} Omitting the characters and placing a stop after .