South African Philately Club ForumMember Forum: Stamps, Covers, Postmarks, Postcards and EphemeraZAR Truncated Double Circle Dates …

Please or Register to create posts and topics.

ZAR Truncated Double Circle Datestamps

#1 · October 3, 2020, 11:44 pm

Quote from Bas PAYNE on October 3, 2020, 11:44 pm
I attach a draft catalogue, hoping that other members may have things to add or to correct, and offer two general comments.

"Time control letters" (a.k.a. "time index letters"): use appears to be far from consistent.

Johannesburg normally uses one letter above the date line (A,C,E,F,H or blank) and one below (A,B,C,D,E or blank) below.

Krugersdorp, Marabastad and Pretoria use only one letter, sometimes above with a 6-pointed star below, sometimes below with the star is above. A,B,C,D,E,F and H have been seen from Pretoria, A,B,C,D and E from Klerksdorp,and A and D from Marabastad.

Boksburg, Klerksdorp and Lydenburg only use the letter A; again sometimes above with a 6-pointed star below, sometimes below with the star above.

Barberton and Potchefstroom have the same 6-pointed star and don't use any letters.

If they are some form of code for the time of day, you might expect some letters to be much commoner than others (e.g. the code or codes for the late afternoon), and for this pattern not to change much from year to year. The tabulation below, for Pretoria (chosen because the sample is reasonably large without the complication of Johannesburg's two letter combinations doesn't show this expected patterning: instead, A-C dominate in 1888 (16 out of 19), but D-H are just as common as A-C in 1889 (13 out of 26). This makes little sense as time codes. A possibility that fits the observed data better is that the codes identify particular clerks, in which case the pattern shown would be consistent with increasing staff numbers at Pretoria in 1888-9. Johannesburg, with more clerks would need more codes than Pretoria, and the smaller offices would not need this kind of system. But more data is needed, and other explanations should also be considered.

Table: Pretoria "time control letters" by three-month period don't show the expected patterning.

A

B

C

D

E

F

H

6.88

4

7-9.88

3

2

10-12.88

3

1

3

3

1-3.89

2

3

1

4

4-6.89

2

1

1

1

7-9.89

1

2

3

1

1

10.12.89

1

2

2. Interestingly, the two-letter abbrevations used for the month are JA, FE, MR, AP, MY, JU, JY, AU, SP, OC, NO and DE. Two of these - MY and JY - suggest an origin from English rather than from Dutch or Afrikaans.

Any comments would be welcomed.

I attach a draft catalogue, hoping that other members may have things to add or to correct, and offer two general comments.

"Time control letters" (a.k.a. "time index letters"): use appears to be far from consistent.

Johannesburg normally uses one letter above the date line (A,C,E,F,H or blank) and one below (A,B,C,D,E or blank) below.

Krugersdorp, Marabastad and Pretoria use only one letter, sometimes above with a 6-pointed star below, sometimes below with the star is above. A,B,C,D,E,F and H have been seen from Pretoria, A,B,C,D and E from Klerksdorp,and A and D from Marabastad.

Boksburg, Klerksdorp and Lydenburg only use the letter A; again sometimes above with a 6-pointed star below, sometimes below with the star above.

Barberton and Potchefstroom have the same 6-pointed star and don't use any letters.

If they are some form of code for the time of day, you might expect some letters to be much commoner than others (e.g. the code or codes for the late afternoon), and for this pattern not to change much from year to year. The tabulation below, for Pretoria (chosen because the sample is reasonably large without the complication of Johannesburg's two letter combinations doesn't show this expected patterning: instead, A-C dominate in 1888 (16 out of 19), but D-H are just as common as A-C in 1889 (13 out of 26). This makes little sense as time codes. A possibility that fits the observed data better is that the codes identify particular clerks, in which case the pattern shown would be consistent with increasing staff numbers at Pretoria in 1888-9. Johannesburg, with more clerks would need more codes than Pretoria, and the smaller offices would not need this kind of system. But more data is needed, and other explanations should also be considered.

Table: Pretoria "time control letters" by three-month period don't show the expected patterning.

	A	B	C	D	E	F	H
6.88	4
7-9.88	3	2
10-12.88	3	1	3	3
1-3.89	2	3	1	4
4-6.89	2	1		1			1
7-9.89	1		2	3	1	1
10.12.89			1			2

2. Interestingly, the two-letter abbrevations used for the month are JA, FE, MR, AP, MY, JU, JY, AU, SP, OC, NO and DE. Two of these - MY and JY - suggest an origin from English rather than from Dutch or Afrikaans.

Any comments would be welcomed.

Uploaded files:

TDC-draft.pdf

#2 · October 4, 2020, 2:50 pm

Your suggestion that the letters might refer to clerks and not time is brilliantly leftfield. I hope you are proven right.

This is very impressive research on an interesting topic. As a result, I trawled through a file of ZAR / TVL postmarks hoping I would find some worthy examples of the truncated circle datestamp to share with Bas. I was disappointed to find just four stamps, a fact which says as much about the remaindered nature of my ZAR / TVL accumulation as it does about my ability to add anything erudite here. Be that as it may, I have had welcome fun away from this website going through that material. Here are the results, four stamps, all showing various differences described by Bas above. They are from left JOHANNESBURG with the date reversed; a typical KRUGERSDORP; a 1/2d with no town name which has an H top and a B bottom which is possibly from Johannesburg ( Bas shows something similar. This is good reference material. Thanks!); and finally a POTCHEFSTROOM with A top. The Bottom Line is that I will be looking more closely at ZAR / TVL stamps in future. I have enjoyed this excursion into the "ou Transvaal!" even if I return with no fruit.

Uploaded files:

#3 · October 4, 2020, 5:02 pm

Steve:

Many thanks. Two interesting ones!

The reversed date is the first one I've recorded for a Johannesburg TDC (in over a hundred records).

The Potchefstroom with A above is even more interesting which is new to me: the few examples I've recorded all have no tcl. This is the earliest Potchefstroom TDC I've seen (I think I read 14 JU 89 - is that right?) - van den Hurk reports an EKD of 24 NO 88, but I haven't seen it. Perhaps they started out using a tcl and then stopped - the examples I've previously recorded date between JA 90 and SP 93.

You'e almost certainly right that the the date-stamp on the 1/2d is from Johannesburg - I read "G." on the left, and it is in the right position relative to the heavy lines (as also is the "Z.A."), which excludes Boksburg.

#4 · October 4, 2020, 6:34 pm

I am so pleased that my hour's scratch-around has born fruit. I think you are right that the 'Potchiesdorp' reads 14 JU 89. The final numeral is a '9' but I cannot see an '8'. I will leave you to stick your neck out on that one. I will post the two interesting ones to you tomorrow and you can decide. Please accept them FOC compliments of the SAPC.

#5 · October 4, 2020, 9:02 pm

Steve:

That's very kind, and much appreciated.

Bas

#6 · October 4, 2020, 9:25 pm

Quote from yannisl on October 4, 2020, 9:25 pm
I am not a collector of ZAR so my comments are from similar situations in Britain and the Cape of Good Hope.

Time controls started with one letter codes both in Britain as well as the Cape. The absence of a letter was not always an error as is normally stated by philatelic dealers, but a different time slot as is shown in the image below.

For smaller offices the norm was to use fewer letters, that is longer time slots. At least for the Cape these codes were periodically published and changed. (See page 200 of Goldblatt). After 1894 two letter codes can be found the first letter denoting time and the second minutes, but again there was no uniformity.

In the British Posts, sometimes letters would be used on top or bottom of the postmark. The top being a.m. and the bottom being p.m. This can also (my guess) be achieved by the absence or presence of a star.

The norm to associate a person with a particular time was to keep counter logs. When a person took duty a note would be made in the logbook and hence one could trace from the postmark via the logbook who stamped the letter. If there was a complaint it could be investigated and dealt with.

This worked well, when the mail was small and most postings took place through the counter. Once pillar postings came into being, letters were also handstamped at the back-office. The norm for most of the more modern periods was a double-circle at the front and a single circle at the back, but also again with many exceptions.

Now what happens when you have more than one identical instrument? Again examples from Europe and Britain show that the postmarking device would then have a distinguishing mark, say a number or a letter over and above the time control. When machine cancellers started to be used they would have marks for "Machine A", "Machine B" etc.

I am not too sure if your suggestion that the letters were allocated to particular personnel can be concluded from what is a very small sample statistically. Not impossible but unlikely and the stronger argument against it is that time was too important for postmasters to have made such changes.

As an aside, time was not always uniform in South Africa, so a 12h00 o'clock time in Pretoria was different from 12h00 o'clock in Capetown. It took a conference in Bloemfontein in 1892 to agree a common standard to sort the problems with telegraph time recordings and railway time-tables.

I am not a collector of ZAR so my comments are from similar situations in Britain and the Cape of Good Hope.

Time controls started with one letter codes both in Britain as well as the Cape. The absence of a letter was not always an error as is normally stated by philatelic dealers, but a different time slot as is shown in the image below.

For smaller offices the norm was to use fewer letters, that is longer time slots. At least for the Cape these codes were periodically published and changed. (See page 200 of Goldblatt). After 1894 two letter codes can be found the first letter denoting time and the second minutes, but again there was no uniformity.

In the British Posts, sometimes letters would be used on top or bottom of the postmark. The top being a.m. and the bottom being p.m. This can also (my guess) be achieved by the absence or presence of a star.

The norm to associate a person with a particular time was to keep counter logs. When a person took duty a note would be made in the logbook and hence one could trace from the postmark via the logbook who stamped the letter. If there was a complaint it could be investigated and dealt with.

This worked well, when the mail was small and most postings took place through the counter. Once pillar postings came into being, letters were also handstamped at the back-office. The norm for most of the more modern periods was a double-circle at the front and a single circle at the back, but also again with many exceptions.

Now what happens when you have more than one identical instrument? Again examples from Europe and Britain show that the postmarking device would then have a distinguishing mark, say a number or a letter over and above the time control. When machine cancellers started to be used they would have marks for "Machine A", "Machine B" etc.

I am not too sure if your suggestion that the letters were allocated to particular personnel can be concluded from what is a very small sample statistically. Not impossible but unlikely and the stronger argument against it is that time was too important for postmasters to have made such changes.

As an aside, time was not always uniform in South Africa, so a 12h00 o'clock time in Pretoria was different from 12h00 o'clock in Capetown. It took a conference in Bloemfontein in 1892 to agree a common standard to sort the problems with telegraph time recordings and railway time-tables.

Uploaded files:

#7 · October 4, 2020, 10:42 pm

Yannis:

You are right to urge caution. However, it should also be right to be open to considering alternative explanations. The ZAR wasn't part of the British Empire, and the importance of time-keeping to the hour wasn't as large when most post still travelled by cart.

Incidentally, the sample number is large enough that the difference between 1888 and 1889 is statistically significant at the 5% level (ch-squared test with Yates' correction). But please send scans of more examples to increase sample size ....

Best,

Bas

#8 · October 5, 2020, 7:54 am

Gentlemen, this topic is yelding wonderful information and insights into the workings of the Cape and ZAR post Offices. Thank you.

There is a lesson in this topic for SAPC readers. I was disappointed find only four ZAR Truncated Circle Datestamps among my accumulation. Bas, however, found two of them very interesting, one of which was the first example he had seen in over 100 recordings. It emphasises how no stamp should be over-looked even if it is the only example you have. It may most likely be ordinary and common but could be an extraordinary new find for a specialist collector and researcher. It is incredible that this has happened to me in this forum. I looked at my four and thought "no" but posted them anyway. Bas looked at two and thought "YES!" This is a good result but we need more. Readers are encouraged to examine their postmarked stamps, compare them against this emerging new philatelic knowledge - and submit them!

Yannis, your wonderful COGH Time Control page will be useful to many Cape collectors. At some point, we need to get that under a Cape banner.

#9 · October 5, 2020, 11:01 am

Quote from Bas PAYNE on October 5, 2020, 11:01 am
Steve -

I agree completely. The take-home messages are that research is worthwhile and fun because there's still a lot that we don't know, and that sometimes what we find makes us take a fresh look at what we think we know.

In that spirit, I've taken a look at the use of "time control letters" in other Pretoria date-stamps between 1883 and 1893. If, as the conventional view has it, they reflect a local system of time of day codes, we might expect some regularity over a decade (or an occasional sharp change if the system changed).

Here's a summary (excluding date-stamps without time control letters or possible space for them):

A

B

C

D

E

F

G

H

O

blank

1883-1.88

Put 6

10+

5

6.88-12.88

Put 20 (TDC)

10

3

3

3

1889

Put 20 (TDC)

5

4

4

8

1

3

1

1890

Put 8

3

2

3

1

10

1891

Put 8/8a

1

2

1

1

10

11.91-7.92

Put 10a

10

3

11

6

9.92-3.93

Put 12

2

8

2

3

7

4

7.93 – 1.94

Put 13

3

3

2

14

2

1

1

One note of caution is needed - this is based partly on my collection, and partly on a folder of scans taken from internet sources; there is some risk of occasional double-counting. However, what does seem fairly clear is that there's no very clear consistent pattern of the kind one might expect if the letters are time codes. Why, for instance, would there be a very different pattern in early 1892 (80% A-C) as compared with later 1893 and 1893 (65% D-H) - a difference that is statistically significant at the p<.01 level? Is really likely that people posted their letters much earlier the day in early 1892 than in late 1892 and 1893?

I'm not suggesting that this proves the "clerk-code" suggestion; but it does seem to cast doubt on the traditional "time-code" belief.

think we know turns out not to be

Steve -

I agree completely. The take-home messages are that research is worthwhile and fun because there's still a lot that we don't know, and that sometimes what we find makes us take a fresh look at what we think we know.

In that spirit, I've taken a look at the use of "time control letters" in other Pretoria date-stamps between 1883 and 1893. If, as the conventional view has it, they reflect a local system of time of day codes, we might expect some regularity over a decade (or an occasional sharp change if the system changed).

Here's a summary (excluding date-stamps without time control letters or possible space for them):

		A	B	C	D	E	F	G	H	O	blank
1883-1.88	Put 6									10+	5
6.88-12.88	Put 20 (TDC)	10	3	3	3
1889	Put 20 (TDC)	5	4	4	8	1	3		1
1890	Put 8	3	2	3		1					10
1891	Put 8/8a	1	2		1				1		10
11.91-7.92	Put 10a	10	3	11	6
9.92-3.93	Put 12		2	8	2	3	7	4
7.93 – 1.94	Put 13	3	3	2	14	2	1	1

One note of caution is needed - this is based partly on my collection, and partly on a folder of scans taken from internet sources; there is some risk of occasional double-counting. However, what does seem fairly clear is that there's no very clear consistent pattern of the kind one might expect if the letters are time codes. Why, for instance, would there be a very different pattern in early 1892 (80% A-C) as compared with later 1893 and 1893 (65% D-H) - a difference that is statistically significant at the p<.01 level? Is really likely that people posted their letters much earlier the day in early 1892 than in late 1892 and 1893?

I'm not suggesting that this proves the "clerk-code" suggestion; but it does seem to cast doubt on the traditional "time-code" belief.

think we know turns out not to be

#10 · October 5, 2020, 2:41 pm

Bas

Please persevere and I am sure we are all open minded to accept alternatives. There is still much to learn. Sadly my meager collection of ZAR is not with me here in Doha. For the last few years I have been a "nomad" philatelist!

I like the way you approached the problem, especially with the additional table you have provided. The difference in time of usage is significant. You ask "is it really likely that people posted their letters much earlier the day in early 1892 than in late 1892 and 1893?",

There are many possibilities:

Clerk-code
Possibility there is an association with the time change Conference in Feb 1892, when uniform time was accepted ?
Is the assumption correct about all stamps reflect time of posting? How about collection from Pillar boxes? Was this uniform over time? Did the additional personnel clear backlogs earlier?
Additional railway mail with different cut-off times.
Changes to slot allocations of times.

If you have any "posted late" letters for the period they might provide more information. Also any telegrams with stamps.