The Ultra-Rare 1926 Proof Set of the Union of South Africa

Pierre H. Nortje (May 2026)

Foreword

For the first few years after the first coinage for the Union of South Africa was struck, there were quite a few changes made to the designs of most of the denominations. This could be the reason why, with the exception of the first year (1923), proof coins were only struck in very limited numbers to possibly be presented, at least initially, as “trial set” gifts to dignitaries and high-ranking officials. In 1926, there were design changes to the farthing, penny, shilling and half-crown.

Source: NGC

Introduction

This year marks the centenary of the 1926 proof set of the Union of South Africa. One would have thought that after a span of a hundred years, there wouldn't be many unanswered questions regarding a legendary set like this, but it seems that even after such a long period, there are still many questions to be answered.

The 1926 set, however, is not unique in its rarity, as, with the exception of the first year of issue set of 1923, all the proof sets from 1926 to 1946 were produced in low numbers, with some years having no proof sets issued at all by the South African Mint. What makes the 1926 set unique is that it was the only one of these low-mintage sets that was not complete from the 1/4d up to the half-crown, as it consists of only 6 of the 8 coins.

Even by 1950, when the first South African coin catalogue was published by Dr. Alec Kaplan, it seems that there was already some confusion about the number of proof coins struck by the SA Mint in the early years: Kaplan was, for example, of the opinion that for the year 1926, only 16 proof farthings were struck and no other denominations, while we know today that only the ½ penny and 2-shillings (and the two gold issues) were not issued for that year in proof condition.

Four years earlier, in December 1946, Dr. Kaplan also noted in the Journal of the Institution of Certificated Engineers (South Africa) that 16 farthings were struck in 1926.

Some early comments on Dr Kaplan’s Catalogue

The author has a first edition of Dr, Kaplan’s catalogue stamped “FOR REVIEW” which was probably sent out to fellow collectors for comments. Regarding the specimens (the word used for proofs in those days) someone crossed out the 1/4d for the year 1926 and wrote “Set 2/6 – 1/4d” as well as “16 sets” Who this person was we do not know but it seems he himself was unaware that half pennies and 2-shillings were not issued.

In the South African Numismatic Society Newsletter (Vol. 2, No. 3, February 1950), the editor, Dr. Frank Mitchell, wrote a critical review of Dr. Kaplan's catalogue, touching, among others, on the number of proofs that Kaplan provides in his catalogue. A response by Dr. Kaplan was also published in the same newsletter. He says, "The figures given in September 1948 by the Mint regarding the 1926 specimen set apply only to the 1/4d (farthing). No other values were made”.

It seems that Dr. Mitchell was also unsure about the existence of the 1926 proof sets: In De Nummis, the Journal of the Transvaal Numismatic Society (Number 2 1956/7) he wrote “Proof sets do exist for the years from 1930 to 1939, but the numbers struck were very small indeed during most of those years. In respect of the “twenties”, the position is even more serious, for apart from the original proof sets of 1923, which were fairly numerous, and the debatable sixteen sets of 1926, no proof sets at all were struck before 1930”.

In 1971, an extract from the paper on “Numismatic Rarities” by S.G. Abrahams was published in De Nummis (Number 4). Abrahams concluded that the 1926 set is the rarest South African proof set and rated it rarer than both the sets of 1930 (14 issued) and 1932 (12 issued). One must deduce from this that he did not agree with the number of 16 sets that were punted by other numismatists of the era.

South African Numismatic Society

In their newsletter of January 1956 (Vol. 4 No.12), SANS published a list of all the specimen (proof) sets issued by the SA Mint up to that year and says that “The (following) figures are accepted as correct by the Director of the South African Mint and by our Society…” For the year 1926, sixteen sets are recorded as …

Short Set, 2/6- to ¼d (8 pieces)

This is clearly incorrect as we know today. What is truly amazing is that the Director of the South African Mint at that stage, as well as the members of SANS did not know that no proof half pennies and 2-shillings were struck for the year 1926. The same mistake was made when the De Nummis, the Journal of the Transvaal Numismatic Society (Number 2 1956/7) published an article entitled “Union Coins struck at the Mint in Pretoria” by HM Stoker.

In the monthly Bulletin (No. 5) of the SA Numismatic Society, dated May 1967, Dr. Frank Mitchell noted that the 1926 farthing was something of an enigma. He stated that farthings for circulation purposes were not struck, but the odd one or two that may be found in circulation are almost certainly "run-on" proofs - that is, proofs struck for 1926 and surplus to requirements. It seems that Dr. Mitchell was implying that there were more 1926 farthings struck than the other denominations in proof - we will come back to this issue later in this paper.

At the SANS meeting of 22 May 1972, which was held at the SA Museum in Cape Town, Mr. Allen Jaffe gave a talk on the major and minor rarities and scarcities of South African silver and bronze coins from 1923 to date. Mr. Jaffe concluded his talk by stating that, in his mind, the rarest of all coins under the heading of his address is the 1926 ¼d, of which a total mintage of 16 coins went into proof sets. "...and yet, no one has ever seen a 1926 SA proof set, so the question must be asked as to whether these 16 coins were ever issued? This mystery still requires solving."

This photo of Allen Jaffe was originally published in the Afrikaans newspaper Rapport in the early 1970s, with a short article on the coin he is holding in his hand: a South African Farthing of 1926, then valued at R10 000.

More comments

Cliff van Rensburg (SA Coin & Banknote catalogue: 2002:48) wrote that three partial 1926 proof sets (no 2/- or 1/2d) were issued over a period of years at the insistence of a keen numismatist. These were unrecorded until the 1980s when three sets surfaced, as well as a set that did not have a shilling. Subsequently, a shilling in proof did surface, which was matched with the fourth set. There are only four complete sets known, making this the rarest proof set in the SA series.

In June 2011, bizcommunity.com reported that the new Absa Money Museum had reopened at its new home in Absa Towers West in the Johannesburg CBD. The report stated that the collection includes highly specialised collectors' pieces such as a 1926 proof set of six South African coins - one of only 16 sets originally minted and believed to be one of only four sets in existence today. This set was displayed at one of the coins fairs in Durban a while ago and it was not boxed: The individual coins were wrapped in cellophane and stored in small envelopes.

In September 2014, the coin collection of Robert Bakewell was sold by Dix Noonan Webb in London. Included in the collection was a 1926 set described as ...

"Proof Set 1926 PF 63-65. Only four sets are known, of which two are in museums. The location of the fourth is unknown."

According to Professor Francois Malan of the University of Pretoria, the South African Mint also has a 1926 proof set in its collection.

Among the future events at which items from the Gatsby Collection were sold at the American Numismatic Fair (ANA), held in August 2025, was a 1926 proof farthing graded PR64 by NGC ... "part of a Six-Piece Proof Set and the finest of only 16 Proofs recorded." This is not correct, as the NGC and PCGS census reports show that although 16 coins have indeed been graded, the highest grade is a Cameo Proof Shilling in PFCA 65 - see below.

NGC & PCGS Census Reports

*The author also knows of at least one ungraded specimen in a private collection. See picture below and note the die crack running through the forefront of the King’s crown.

Source: Anthony Govender

Our comments and some questions

We could not locate the original SA Mint source that states that 16 of each of the six 1926 denominations were struck in proof. In terms of what Cliff van Rensburg wrote, it would seem that three further (?) sets were issued over a period of years at the insistence of a keen numismatist, and that a fourth set later surfaced that was completed when the shilling was added.

What we understand from this is that these sets (or at least three of the four) were probably struck AFTER 1926 and were not part of the original proof issues. If this is indeed the case, then there must be at least 20 of each of the denominations, but if one looks at the census reports, it would seem that only two full sets were ever graded, of which one belonged to Robert Bakewell and was sold in 2014.

Robert Bakewell (left) standing with his friend Anthony Govender in London where the sale of Bakewell’s collection took place. Anthony wrote the foreword for the catalogue. The sale was know as the “Finest Known”.

We speculate that the other graded set was formerly in the collection of Billy Hills, a Gauteng numismatist. His collection was sold in 2017 by Johan Gouws of Pretoria. At that stage, the coins were graded as follows: - (The Penny is erroneously recorded as a Half Penny)

We suspect that these coins were later all sent for NGC conservation, and all the coins except the Penny received higher grades if one looks at the following NGC population report …

The coins in red are, in our view, the same coins. Currently there are no PF60 3d, PF60 6d, PF 62 1/- and PF61 2/6- in the census.

We know that there are also two full ungraded sets, those in the ABSA and SA Mint collections. The whereabouts of the second graded set is currently unknown but we know that it was offered for sale by Randburg Coin in July 2023 as part of a virtually complete proof set collection of the Union of South Africa in gold, silver and bronze. Because we could trace the NGC serial numbers on the slabs, we know that this is the second graded set in the NGC census – here is a picture of one of the coins (the shilling) in the original Randcoin advertisement and the actual coin in the NGC census. We believe that this is the regraded set that belonged to Billy Hills.

But what about the original 16 sets (36 coins) that were struck? Were they issued in mint boxes or dispersed individually? We will remember that Allen Jaffe said in 1972, "No one has ever seen a 1926 SA proof set." With the exception of the farthing, it would seem that never at any stage were there more than four coins of the other denominations known, recorded, or offered for sale by collectors.

The farthing, as Dr. Mitchell puts it, is indeed an enigma, as although it is extremely rare, it is the only denomination of which more than four coins are known. Dr. Mitchell believed that more were struck than were actually required. One also wonders why Dr. Kaplan, in 1950, said that "The figures given in September 1948 by the Mint regarding the 1926 specimen set apply only to the 1/4d (farthing). No other values were made."

Is it possible that he was correct and that the other denominations were only later struck at the request of a numismatist, as stated by Cliff van Rensburg? If at least 16 of the original denominations were struck, what are the chances that in each case (except the farthing) EXACTLY the same number (being 4) seems to have survived? The author speculates that this is too much of a coincidence and that there is a good chance that the four sets that survived were the only sets ever produced, and not necessarily in the year depicted on the coins.