A golden post horn on a black-red-gold tricolor. The width of the red stripe was 20% larger than the other stripes. This was adopted 11 April 1921 and abolished 31 March 1933. Its use on buildings was discontinued 9 November 1927. Source: Dreyer 1999 [dry99]. Norman Martin, 1998 and 10 December 1999
This flag...was succeeded by the 1933-1935 postal flag on 31st March 1933; however, after 30th January 1933 any black-red-gold flag was certainly rarely hoisted anymore. Marcus Schmöger, 29 March 2001
The final postal flag of the Weimar era was introduced with the regulation of 11 April 1921, finally in force 1 January 1922. This was a black-red-gold triband with a widened red stripe (4:7:4) to accomodate the post horn. The post horn was golden, and of a new design - different from the one used 1893-1919 and also different from the one used 1950-1994. Proportion 2:3. The flag was used on land as service flag of the postal authorities, but also at sea: ships carrying mail hoisted the postal flag at the main top and as a jack. Sources:
The question is far from clear-cut. Emil Dreyer (Dreyer 1999 [dry99]), Norman Martin and myself were basically only interpreting the legal prescription, but none of us described how he came to the interpretation. The legal text is not really self-explaining, so it needs interpretation:
As the legal text for the postal flag 1921-1933 was basically copied for the 1950-1994 flag, we must assume that the postal flag 1922-1933 and the postal flag 1950-1994 had the same stripe widths.
The legal text reads as follows, "Die Reichspostflagge hat die Querstreifen wie die Nationalflagge, in der Mitte des um ein Fünftel der Randstreifen breiteren roten Querstreifens ein goldgelbes Posthorn...". My translation is, "The imperial postal flag has the horizontal stripes as the national flag, in the center of the red horizontal stripe, that is wider by a fifth of the edge stripes, a golden yellow post horn...". Of course this translation already implies an interpretation.
My interpretation was: First of all we have three equal stripes 1/3 1/3 1/3; in a second step my interpretation is: we take a fifth of each of the edge stripes, i.e. of the upper and lower stripe, so we add 1/15 of the total flag width at the top of the central stripe, and 1/15 at the bottom of the central stripe: 1/3-1/15 1/3+1/15+1/15 1/3-1/15. So we end up with: 4/15 7/15 4/15, which is my ratio 4:7:4 (the central stripe has 46,66% of the total flag width).
Emil Dreyer's interpretation was (literally quoted from Dreyer 1999 [dry99]), "The red stripe was to be one fifth broader than each contiguous stripe." So that would mean a ratio of: 1 1,2 1 or 5:6:5 (the central stripe has
37,5% of the total flag width).
The obvious question is, are there no (official) drawings attached to this legal text? There are three groups of publications, where I measured the actual width of the central stripe:
official publications like Reichsministerium des Innern 1926 [rmi26], Reichsministerium des Innern 1928 [rmi28], Reichsministerium des Innern 1930 [rmi30] and Bundesministerium des Innern 1966 [wfl66];
general vexillological publications like Valentin and Neubecker 1928 [vne28] and Dreyhaupt 2000 [drh00];
vexillological publications dealing only with postal flags like 'Meyer 17' 1960 [mzs60], Karaschewski 1996 [kar96] and Dreyer 1999 [dry99].
The official publications all show the central stripe with 42% to 44 % of the total flag width; the general vexillological publications show it with about 45 %; the postal flag publications - [dry99] actually reprinting [mzs60] - show it with 46,8% to 47,8%.
So we can assume, that Emil Dreyer's interpretation is wrong; from the drawings we cannot judge the exact ratio, as especially the official publications show something like 2:3:2 ratio, whereas the publications on postal flags seem to support my interpretation of a 4:7:4 ratio. I do not trust the drawings in official publications too much, as I already found too many errors and ambiguities there.
What do actual flags show? I do not have good photos of the postal flags, so I cannot say.
All in all, I think that my interpretation of the legal prescription together with the drawings in 'Meyer 17' 1960 [mzs60], Karaschewski 1996 [kar96] and Dreyer 1999 [dry99] clearly favour the thesis that the ratio of the stripes in the postal flags (1921-1933 and 1950-1994) was 4:7:4.