3 minutes discrepancy with 19th century longitudes

[Guest guest]

Dear colleagues,

 

while working on the location of Tevaram Sites

in Tamil nadu, I have noticed that there seems

to be a systematic difference between the longitudes

of sites as can be read on modern maps (or determined

through GPS) and the longitude as it can be found

inside the Glossary of the Madras Presidency

compiled by C.D. MacLean (=CDML)

 

For instance the town of cIkAzhi (alias sirkazhi, alias Shiyally)

is to-day located by 11°14' North and 79°44' East,

but CD-Maclean gives it as being at (11°14' N; 79°47' E)

 

Numerous example could be quoted, but I will only

mention a few for the sake of brevity:

CDML to-day Place name

79°54' 79°51' akattiyAn2 paLLI

79°37' 79°34' tiruvatikai

79°31' 79°28' AmAttUr

79°41' 79°38' tiruvArUr

79°34' 79°31' AvaTutuRai

........................

 

I am not absolutely sure this question is appropriate

for Indology, but since C.D.MacLean's Glossary

is kind of a reference tool, I figure the question

might be of interest to other colleagues.

 

Thanks for any pointers explaining the discrepancy

 

-- Jean-Luc CHEVILLARD? Paris

[Guest guest]

Dear colleagues,

 

after asking my question to you, I had second thoughts

and tried to find a discussion list more appropriate for this.

 

As a result, I got an answer from Dr Andrew S Cook

(andrew.cook, Map Archivist, India Office Records, The British

Library),

which he has allowed me to reproduce on Indology (see below)

That answer was given on the MAPHIST list, hosted by a listserv at Harvard

University.

As an appendix, I also include some further explainations he gave.

 

Greeting to everybody on the occasion of the second birth of Indology

and congratulations to Dominik Wujastyk for the new-born baby

(and thanks for maintaining its elder sister/brother [what is the gender of

a list?]

active to everybody's satisfaction for such a long time span)

 

-- Jean-Luc CHEVILLARD (Paris)

 

************** 1st answer by Andrew S. Cook ************

 

"Cook, Andrew" <Andrew.Cook

"'maphist'" <maphist,

"'jlc'" <jlc

RE: 3 minutes discrepancy between to-day's and 19th century longitude

Wed, 18 Apr 2001 16:04:04 +0100

 

The reason for the systematic discrepancy between nineteenth- and

twentieth-century maps and gazetteers for longitude figures for places in

India is that India was mapped in the nineteenth century on a sheet net

based on an incorrect calculation of the longitude from Greenwich of Madras

Observatory. Post-1905 maps give the modern longitudes currently in use,

but earlier maps require a correction to be applied. The standard note on

Survey of India maps published on the pre-1905 scheme but after the error

was noticed reads as follows:

'The Longitudes are referrible to the Greenwich Meridian, taking that of

Madras Observatory as 80[degrees]17[minutes]21[seconds] East. They require

a correction of -2[minutes]30[seconds] to make them accord with the most

recent value of that Observatory, viz., 80[degrees]14[minutes]51[seconds]

East.'

Rounding to the nearest minute of longitude tends to give a standard

three-minute correction, which is what you have noticed in MacLean. Map

users (myself included) quickly get used to looking 2 minutes 30 seconds

East (or West) of the nominated point when using a combination of old and

new maps and gazetteers. When you notice two longitude values in such

cases, the lower is usually correct.

I hope this helps.

Andrew Cook

--------------------------

Dr Andrew S Cook andrew.cook

Map Archivist, India Office Records

The British Library

96 Euston Road

London NW1 2DB Telephone/Voicemail 020 7412 7828

United Kingdom Fax 020 7412 7641

--------------------------

***************** My question to Maphist

 

>Dear list members,

>while working on the location of Tevaram Sites in Tamil nadu

>(Tevaram is a collection of Tamil hymns to god Siva), I have noticed

>that there seems to be a systematic difference between the longitudes

>of the sites as they can be read on modern maps (or determined using GPS)

>and the longitudes as they can be found inside the Glossary of the Madras

>Presidency

>compiled by C.D. MacLean (=CDML)

>[i am having the second AES Reprint, dated 1990, of the 1893 edition]

>For instance the town of cIkAzhi (alias sirkazhi, alias Shiyally [CDML])

>is to-day located by 11°14' North and 79°44' East,

>but CD-Maclean gives it as being at (11°14' N; 79°47' E)

>Numerous example could be quoted, but I will only

>mention a few for the sake of brevity:

>CDML to-day Place name CDML spelling

>79°54' 79°51' akattiyAn2 paLLI (Agastyapully)

>79°37' 79°34' tiruvatikai (Trivetty)

>79°31' 79°28' AmAttUr (Trivamauttore)

>79°41' 79°38' tiruvArUr (Tirvarore)

>79°34' 79°31' AvaTutuRai (Tirvaudootoray)

>........................

>What is the explanation for the discrepancy?

>Thanks for any explaination or any pointers explaining the discrepancy

>-- Jean-Luc CHEVILLARD (Paris)

><http://www.linguist.jussieu.fr/~chevilla/>

>

>

 

************** 2nd answer by Andrew S. Cook ************

I hadn't intended burdening the list with the detritus of scholarship, but

as there seems to be an identifiable demand for it, here it is.

 

Lambton's 1805 value for the longitude of Madras Observatory was

80[degrees]18[minutes]30[seconds]. This was used for the Atlas of India

quarter-inch sheets. Warren's 1815 observations of the emersion of

Jupiter's satellites gave 80-17-21, which was used for the Great

Trigonometrical Survey and consequently for the 1:63,360 standard-sheet

provincial map series. There were further determinations between 1826 and

1858 by Goldingham (Jupiter's satellites), Taylor (lunar culminations),

Riddle, Everest and Jacob (all working off Taylor), ranging from 80-17-15

to 80-14-19.20. The 1874 Transit of Venus expedition fixed the longitude of

Suez, which was carried to Madras in 1878 by the electro-telegraphic

observations of Campbell and Heaviside to give 80-14-51.24. Apart from a

rogue determination by Walker in 1883 of 80-14-50.03, and despite

unsatisfactorily large closing errors in the subsidiary network of

longitude triangles, later found to result from imperfections in the object

glasses of the collimators, the observations of 1890 and 1893 by Strahan

varied by no more than a quarter of a second from the 1878 determination.

The correction note of -2[minutes]30[seconds] (80-17-21 less 80-14-51) was

applied to maps after 1878, and in turn gave rise at the 1891 Berne

International Geographical Congress to criticism of India for not

correcting the maps themselves. 1896 observations of the longitude

connection between Karachi and Greenwich, via Tehran and Potsdam, which

produced a Madras reading of 80-14-47.06, caused further problems. In

defining a longitude for the new 1900 Map of India and Adjacent Countries

(the basis of subsequent large-scale mapping of India) as the new

Karachi-Tehran-Potsdam-derived Madras longitude, a decision was required

whether to back-calculate the longitude of Kalianpur (the historical origin

of the Survey) by triangulation from the new Madras 80-14-47.06 figure, or

to take the astronomical value for Kalianpur based on Madras and to

calculate forward by triangulation a new, geodetically-consistent value for

Madras. The latter course was chosen, giving a new 1900 geodetic value for

Madras Observatory of 80-14-54.20, and post-1900 old-scheme maps modify the

correction note to -2[minutes]27[seconds]. Maps on the post-1905 India-wide

numbering scheme require no correction.

This is derived chiefly from S G Burrard, 'On the Values of Longitude

employed in Maps of the Survey of India', Survey of India Professional

Papers 7 (Calcutta, 1903), pp.1-7.

Since posting my previous message I have been asked whether my contribution

to this exchange is 17 days late. I can only seek to assure readers that it

is all genuine, for once.

Andrew Cook

 

************** 3rd answer by Andrew S. Cook ************

Jean-Luc

 

Yes, you are welcome to forward my postings to the Indology group.

 

The reciprocal distances and directions of all places in India were

established by traingulation by the Great Trigonometircal Survey, in an

exercise commenced by Lambton in 1802, and completed in the 1870s. From a

measured base near Madras the direction and distances of all points were

established by very precise observations to form a chains of triangles. At

each observation point ('trig. station') a theodolite of great accuracy was

hoisted into an tower to perform the observations. These points were

consequently fixed with great accuracy, and they became reference points for

all intermediate mapping. A book which deals with this, and which is easily

available, is John Keay, The Great Arc, published 2000. Also Matthew Edney,

Mapping an Empire, published about 5 years ago.

I hope this helps.

Andrew Cook

 

*************** My 2nd question to which this 3rd reply is an answer

 

>I also have an additional question:

>

>some of the places mentionned in CDML

>are referred to as being "Trig. station" and their location is given with

>greater

>precision, like for instance Aulangoody (CDML, Glossary, p.205, Combaconam

>entry)

>that is described as being at (10° 49' 55.23''; 79° 26' 43.04'').

>

>Did these places play any special role?

>Why is there a difference in the precision?

>

>Thanks for your answer

>

>-- Jean-Luc Chevillard (jlc)