The lower half of the Plate explains itself.  The last figure of all, Fig. 65, is of interest, because it was drawn for an intelligent little girl of only 11 years old, after she had been closely questioned by the father, and it was accompanied by elaborate coloured illustrations of months and days of the week.  I thought this would be a good test case, so I let the matter drop for two years, and then begged the father to question the child casually, and to send me a fresh account.  I asked at the same time if any notes had been kept of the previous letter.  Nothing could have come out more satisfactorily.  No notes had been kept; the subject had passed out of mind, but the imagery remained the same, with some trifling and very interesting metamorphoses of details.

[Illustration:  PLATE III. Examples of an Hereditary Tendency to see Number-Forms, 4 Instances where the Number Forms in same family are alike 3 Instances where the Number-Forms in same family are unlike]

DESCRIPTION OF PLATE IV.

I can find room in Plate IV. for only two instances of coloured Number-Forms, though others are described in Plate III.  Fig. 64 is by Miss Rose G. Kingsley, daughter of the late eminent writer the Rev. Charles Kingsley, and herself an authoress.  She says:—­

“Up to 30 I see the numbers in clear white; to 40 in gray; 40-50 in flaming orange; 50-60 in green; 60-70 in dark blue; 70 I am not sure about; 80 is reddish, I think; and 90 is yellow; but these latter divisions are very indistinct in my mind’s eye.”

She subsequently writes:—­

“I now enclose my diagram; it is very roughly done, I am afraid, not nearly as well as I should have liked to have done it.  My great fear, has been that in thinking it over I might be led to write down something more than what I actually see, but I hope I have avoided this.”

Fig. 65 is an attempt at reproducing the form sent by Mr. George F. Smythe of Ohio, an American correspondent who has contributed much of interest.  He says:—­

“To me the numbers from 1 to 20 lie on a level plane, but from 20 they slope up to 100 at an angle of about 25 deg.  Beyond 100 they are generally all on a level, but if for any reason I have to think of the numbers from 100 to 200, or from 200 to 300, etc., then the numbers, between these two hundreds, are arranged just as those from 1 to 100 are.  I do not, when thinking of a number, picture to myself the figures which represent it, but I do think instantly of the place which it occupies along the line.  Moreover, in the case of numbers from 1 to 20 (and, indistinctly, from 20 up to 28 or 30), I always picture the number—­not the figures—­as occupying a right-angled parallelogram about twice as long as it is broad.  These numbers all lie down flat and extend in a straight line from 1 to 12 over an unpleasant, arid, sandy plain.  At 12 the