Cannon identification


As extracted from a proposal by Rudi Roth in the International Journal of Nautical and Underwater Exploration 18.3 in 1989, and adapted to South African requirements.
With thanks to the Nautical Archaeological Society (UK)


This standard has been compiled by the researchers responsible for the Durr Estates Record “The muzzle loading cannons of South Africa” as a service to the public. The intention is to contribute to the preservation of our national heritage and to assist in the development of military tourism in South Africa.


The complex matter of identifying an ancient muzzle-loading cannon ranks very highly among the subjects which are most likely to lead to confused communications. Worldwide experts are today no more standardised than were the 17th century gun founders who, in total isolation, worked with different length and weight measurements, different proportions, patterns and methods and had different standards of marking their pieces.

The situation was exacerbated when gun founders began copying the successful patterns of others, often from other countries. This lead to what appeared to be an English pattern gun being cast in Sweden for the Dutch to measurements in localised Amsterdam units. Many other combination were also possible.

Further complications arose from the fact that all countries would make random use of captured guns, or guns from another country which were being transported by ships as cargo. The accurate identification of a wreck may thus be a very poor indication of the nationality of the guns recovered from that wreck.

The most reliable method of identifying a gun is by the correct measurement of the key parameters of the gun. The clear presentation of this data will assist experts in this field of research, as they have access to a vast amount of documentation from gun founders, proof firing authorities, arms dealers and shipyards. They also maintain large databases containing the known dimensions, patterns and proportions of many hundreds of guns which have been measured in detail.

Given the material from which the piece is made, the length, the bore, the markings, the weight and the design (or photographs), it should be possible to accurately identify any gun. It is very seldom that all this information is available, but more information that is available, the easier and more accurately the piece can be identified.


The intention of this paper is to provide, without impressive volumes of detail, a standard for the measurement and reporting of historic cannon. This will enable someone who is inexpert in the subject to communicate meaningfully with those who can assist in the correct identification of a piece.


1. Never has to much information been provided on a gun requiring identification.
2. Always provide the dimensions in millimetres.
3. Wherever possible provide several photographs of the gun, including some close-ups, but try to ignore the carriage.
4. Provide detailed descriptions of any markings on the piece, eg hyphens or colons between numbers, where on the piece the markings are, and whether in relief or engraved.
5. Rough sketches will help. No detail is insignificant.
6. Measurements should be taken with care and repeated if necessary in order to obtain an average of several readings, possibly by different people. Accuracy to within two or three millimetres is desirable.
7. Wherever there is doubt about how or where the measurement should be taken, provide a simple sketch and comment on how the reading was taken.


It is important to state from what material the piece was made. The basic options are iron or bronze (or brass). A useful sub-section of this would be if the iron could be identified as cast iron, wrought iron or a combination of these materials, but this is not essential.

Archaeologists tend to measure artefacts from extremity to extremity, this does not help in the identification of a cannon. Some experts differ in their proffered measurement, but the most common useful length of a piece is from behind the base ring to the muzzle face. This is the manner in which designers and users of the pieces recognised specific guns and is the format available in old documentation. In the case of mortars, the entire length from muzzle to the rear of the trunnions is measured.


It can be very difficult to obtain accurate measurement of the diameters of a cannon. The ideal solution is to use a large calliper or vernier in order to ensure accuracy. Such large measuring instruments are rare, with the result that an alternative method of measuring is required. Any attempt to hold a tape measure or ruler across the top of the piece and then sighting down the sides to measure the diameter is far to inaccurate.

The solution is to measure the circumference (all the way around the piece) with a good quality flexible tape measure as used by tailors. This figure is then divided by Pi (3.142) in order to obtain the accurate diameter at that point on the piece.


Provide, as accurate as possible, a measurement of the inner diameter of the bore, measured both horizontally and vertically on a clean part of the bore. Avoid a judgemental dimension such as “18 pounder” as such terms were merely a designated calibre which could vary greatly from country to country and from one year to another, even at the same foundry.

On carronades one must ensure the bore proper is measured, and not the enlarged funnel at the front of the bore. It should be noted whether the bore is smooth or rifled, and if rifled, how many grooves. If possible, indicate the shape of the rear of the bore and indicate if the piece has a chamber (narrowed section at the rear of the bore) or not.


All service pieces and many civil guns had their weight, as measured per gun after completion, engraved somewhere on the piece. Most often on the top of the base ring, but on English guns it can often be found on the lower rear of the cascable. The format of the weight markings could be:
a) 1234A – A Dutch weight of 1234 Amsterdam pounds.
b) 32 – 3 – 12 An English method of indicating a weight of 32 hundredweight, 3 quarters of a hundredweight and 12 pounds.

c) XVI : XVIII : III This was the Swedish marking for 16 Skepspund, 18 Lispund and 3 pund.
There were several other formats. The digits may be followed by letters which indicate in which measurement unit the weight is inscribed. It is important to provide the format and punctuation in detail. Possible suffixes are: – A, K, tt, S, P, q, or L.


Here the enquirer cannot reasonably be expected to recognise the design of the piece as being an “Armstrong”, a “Blomefield”, a “Finbanker”, or anything else. It is here where the photographs play an important part, as the expert will probably recognise the design immediately, unless it is a particular obscure piece.


In order to assess the validity of the dimensions given, the condition of the piece has to be known. this should be expressed not only as good, bad or worse, but some indication should be given as to the visibility or otherwise of fine detail such as crests, squareness of trunnion ends, reinforce rings and fillets.


One of Cape Town’s Noon Guns has been measured thus:-

Material: Cast iron
Length: 2740mm
Bore: 136mm
Weight: 41-3-7
Serial no: 249
Markings: Bearing the broad arrow and the crest of KG III on the second reinforce and WCo on the left trunnion end.

The official records of that time reflect that the original designation was:-
Iron Blomefield 18 pdr of 9 foot and nominal weight of 42 cwt.

From records obtained from Mr Rudi Roth, it was possible to establish that this gun was cast by Walker & Company in England and from the serial number, established that it was proof fired at Woolwich on 10 June 1794. The preceding example could be identified because all the basic information was available. It is very seldom that this is always the case, yet one should not despair at the absence of such convenient detail, much can still be done.



It is assumed that the material from which the piece is made can be established and the correct length can be measured to within 3mm of accuracy.

In the event of the bore being obscured or completely blocked, there are other dimensions which will give reasonably clear indicators of the calibre.

These are, in order of priority, as follows :-
a) The diameter of the piece at or near the vent hole.
b) The accurate diameter of the basering.
c) The smallest diameter of the piece, usually just behind the muzzle astragal.
d) The diameter of the button and neck.
e) The diameters of the trunnions. From this, or from part of this information, an expert should be able to calculate the approximate diameter of the bore.

2. THE WEIGHT (not vital, but useful)
In the event of the weight not being legible, it is advisable to provide as many measurements of both length and diameter as possible, with a sketch and indication as to where exactly the measurements were taken. This should include button, neck, trunnions and any other knob, lump or protrusion. An indication of the condition of the piece in terms of material lost to corrosion or erosion is vital for this activity. Armed with this information, a scientific calculator and a clear grasp of spherical trigonometry, an expert should be able to calculate the weight to within a useful tolerance.

If clear photographs cannot be obtained, then comment on the characteristics of the piece should be given in as much detail as possible. Here one should describe any features which may assist in identification. This includes:-

a) Lifting rings or dolphins.
b) The presence or absence of a breeching ring on the cascable, and if present, it’s shape.
c) The number and shape of all reinforce rings and fillets.
d) The number and shape of muzzle fillets.
e) The presence or absence of a vent patch, it’s size and the position of the vent and pan.
f ) Whether the vent patch, if present, has holes bored horizontally through the patch at right angles to the bore, and the spacing of these holes.
g) The presence and position of any holes, fittings, sights or patches.
h) Whether the trunnions are slightly tapered or cylindrical.
i) Whether the trunnions have shoulders against the barrel or not.

Any photographs taken of the cannon should include at least one general view of the side of the piece and then close-ups of the breech, the trunnions, the muzzle and one of any markings which can be seen. The appropriate convention is to photograph the right side of a cannon. This will place the base ring on the left and the muzzle on the right. In the case of obstructions or serious damage to the right side of the piece, this convention is dispensed with.

It is understood that a diver, working among silt, barnacles, wreckage and strong currents, using a torch in murky water on a half buried gun, cannot be expected to provide ideal measurements for the identification of what has been found. In those cases where a diver wishes to know if what has been found is worth salvaging, the following is recommended:-

a) If the means to preserve what is being salvaged is not available, then it is best left where it was found, particularly if it is made of iron. Many iron guns have survived for centuries under the water, only to decay into rusty gravel within three years of being salvaged.

b) If the piece is thought to be a rare iron piece and prior preservation arrangements can be made, then as much accurate information as possible must be provided and a positive identification obtained prior to salvage.

c) If the identification of an iron gun will assist in the classification of a wreck site, then the ideal is to bring a gun to the surface, measure and photograph it and return it to the wreck. Don’t look so shocked! This has been done before.

Please be aware, however, that in terms of the National Heritage Resources Act (Act No 25 of 1999), any wreck older than 60 years of age is protected and may not be disturbed in any way, or anything removed from such a site, except under the terms of a permit issued by theSouth African Heritage Resources Agency (SAHRA).

If a diver does wish to raise a cannon from the seabed for recording purposes, a permit must first be obtained from SAHRA, who in considering such an application would need to be satisfied that:

The proposal is in the best interest of the wreck in question;
The process of removal and replacement is handled with the necessary care; and
Neither the site or the cannon concerned are damaged in any way.

Application forms or further information can be obtained from the maritime archaeologist at SAHRA. Please also note that when measuring cannon located underwater, it is not advisable to chip away or remove any concretions or corrosion product from a gun. These concretions form a protective layer over the cannon and tend to inhibit corrosion. Their removal invariably starts a rapid corrosion, and could lead to serious deterioration or destruction of the cannon.

From the book “The Muzzle Loading Cannon of South Africa” by Gerry de Vries and Jonathan Hall.
Reproduced with permission.


In these modern times of travel and instant communication it is often difficult to visualise what conditions were like back in the 17th and 18th centuries when people were largely area bound and localised customs and standards were the rule.

The weight markings on cannon reflect some aspects of that distant way of life. Firstly there were no international standards for time and for the units of length and weight. Most countries, and indeed some districts within a country, had their own set of weights and measures and their own methods of recording them.

The official military requirement for the weight marking on a cannon was possibly twofold, firstly payment to the foundry was often based on the weight of the casting and secondly the military had various guns of the same calibre but different lengths, diameters and weights. A gun could only be accurately identified by its pattern, metal, calibre and weight.

Cannon cast in England, Wales and Scotland had their weight chiselled or engraved in the format 41 – 2 – 11 being 41 Hundredweight (CWT) , two quarters of CWT and 11 pounds. With one CWT being 112 lbs and one quarter CWT being 28 lbs a gun marked in this manner would weigh 4,659 lbs. One English lb = 454 grams.

Sweden and Norway used Roman numerals with XVI : XVII : IV being Skeppspund, Lispund and Pund with one Pund weighing 340 grams. Twenty Pund made one Lispund and 20 Lispund made one Skeppspund. The above Roman numerals would indicate a weight of ( 136Kg X 16 ) + ( 6.8Kg X 17 ) + ( 0.34Kg X 4 ) for a total of approximately 2,305Kg. When a Roman numeral such as XXVII Is preceded by an “N” this would be the gun number.

The Dutch used the Amsterdam pound which was equal to 494 grams and the marking appeared mainly on the top of the breech ring in the form 2345 A. The cross bar of the “A” was most often in the form of a shallow “V”. CAUTION – Amsterdam weights between 1650A and 1780A inscribed on the top of the base ring are often misinterpreted as the date of manufacture.

Prior to 1794 the French used their own version of the pound (poid de marc) of 490 grams in the format P 1234. After 1794 they marked their cannon in Kg of the metric system in the manner K 2134, most often on a trunnion end.

The Portuguese version of the pound was the arratel which was equal to 459 grams. Their guns were marked in the format 8q 2a 14a which represented 8 quintals 2 arroba and 14 arratel. 32 arratel = 1 arroba, and 4 arroba = 1 quintal of 58,75 Kg.

The Spanish pound or libra had a nominal weight of 460 grams. The Spanish also used the arroba and the quintal terms but they appear to have differed slightly from those of the Portuguese. 25 libra = 1 arroba, 25 arroba = 1 quintal of 46,13 Kg.

NOTE – The weight inscribed on a cannon is not absolute proof of origin, it is merely an indicator of a probability. VOC guns have been found with Swedish weights inscribed, an Indonesian gun with French weigh markings and several other combinations.

Source – [ By generous permission] – Linear Measurement and Weights (copy No. 3) – an unpublished manuscript by Rudi Roth – November 1999