Sent H.H. Risley to K. Pearson, 11th July 1899.
11 Jul 1899
Description:

‘My dear Sir,

I enclose a card of introduction from Dr Garson & venture to ask your advice on one or two points connected with the next Census of India. I am going out in October as Census Commissioner for India and I am anxious to acquaint myself with the most modern methods of dealing with statistics, as far as that is possible for a person with no mathematical knowledge. I would ask you therefore

(1) what books in English French German or Italian are the best authorities on the subject of statistical methods including graphic presentations of facts?

(2) Do you advise that the opportunity of the Census should be taken to compile a very large series of cephalic measurements which have been rendered useless for police purposes by the introduction of finger-prints? At present the data collected by me some years ago hold the field & are better than the police statistics in that they deal with the very important nasal & nao-molar[?] indices which seem to show a closer correspondence with racial characters than the cephalic index/ Still in view of the fact that the police measurements extend to about 200,000 subjects & cover pretty well the whole of India (including Bombay which my figures did not touch) it appears to me that it would be a pity not to compile them. I should mention however that they have one serious defect – the cephalic length has been taken, not from the glabella as Topinard & Collignon & all other modern authorities recommend, but from the root of the nose. I am under the impression that Virchow at one time recommended this measurement but I have no books here to refer to. The point however leads up to my third question –

(3) Assuming that the defect I have mentioned does not condemn the figures offhand would it be possible to convert them, by adding some constant figure, into terms of statistics taken from the glabella? I gathered from Dr Garson that your researches into the question of correlation had been so elaborate that you might be in a position to say authoritatively what the correction ought to be.

Pray excuse me for troubling you with what may appear to you rather rudimentary questions. I am very anxious to make the most use possible of the great map of statistical material which now exists in India & I want to learn what are the best methods to follow.

Yours truly

H.H. Risley.’

Sent H.H. Risley to K. Pearson, 6th Sept. 1899.
6 Sep 1899
Description:

‘Dear Professor Pearson,

I venture to approach you again with a request for advice on a Census problem.

The most costly & at the same time the least satisfactory part of the India census statistics is the record of age. No native outside the educated classes had any idea how old he is or any of his family are. He will speak of himself as being bis[?]-chalis[?] “20 to 40”: an infant will be described as do-ek “2 to 1” a child slightly older as “3-4” and one older still “5-7.” The Census returns show extravagant lumping on the multiples of 10 and 5, also on the ages of 12, 15, 18 and 22 and 32. Every one admits the whole thing is worthless so far as the figures returned by the enumerators are concerned.

Attempt has been made by the English actuary Mr Hardy to correct these results with reference to the mortality statistics of the city of Madras (which are supposed to be approximately accurate) and the returns of the Rajputs of the North West Province who practice infanticide & are looked after under a special law.

But no use appears to have been made of the age-record of the large number of natives in Govt. service, whose ages from 22 or so are accurately noted.

Not what occurred to me is this:-

(1) We might disregard & not compile the bulk of the recorded ages of the 300 million we have to deal with.

(2) In each Province we might select, say 5 or 6 millions of people or some definite proportion of the total population, record it year by year (not by 5 year periods) and correct it in accordance with the results worked out from the sections of the population for which we have a more or less accurate record. For native officials we might I imagine get a fairly trustworthy life table by going back 10 years or so.

Do you think this is sound? I want to save the toil & money now expended on compiling masses of worthless figures.

Yrs truly,

H.H. Risley.’

Sent H.H. Risley to K. Pearson, 8th Aug. 1899.
8 Aug 1899
Description:

‘My dear Sir,

Very many thanks for your long and interesting letter of 19th July which encourages me to hope that I may consult you hereafter on statistical questions which may arise in the course of the Census.

Your advice as to the anthropometric data will be of great value to me. It will enable me to sarisfy the Govt. of India, who had already asked for my opinion on the subject, on two main points.

1st that the measurements should be compiled.

2nd that they should be done apart from the Census.

And I may be able to induce them to have a few hundred measurements taken in order to get the data for a formula which would enable us to convert the nasal measurements into glabella measurements. But even if we cannot do this it may be worth while to compile the police measurements as they stand. For Garson told me when I saw him in London that Virchow had at one time gone in for the length to the root of the nose. I am writing to Virchow on this subject.

I am afraid no Census that it is possible to take in India would give the data for racial fertility which you want. No native knows his own age within 10 or 20 years & the age statistics which we collect are in my opinion perfectly worthless. However I will bear your requirements in mind & will see whether the records of the infanticide area (where we watch the proportion of girls) would be of use to you. Unfortunately I have a suspicion that the Rajputs have of late got to be smart enough to kill a few boys in order to adjust the balance of the sexes! Sat at least an old Rajput hinted to a friend of mine.

I fear I shall not be in London before I return to India.

Yrs very truly,

H.H. Risley.’