Falling Horsemen - a 'Mint' set
Falling Horsemen

a 'Mint' Set

Issued for a dozen years in the middle of the 4th century AD the Falling Horseman is one of the most common ancient Roman coins. On my page on common coins I ranked these as the most common. Additionally, I have previously posted a page on Falling Horsemen and other types using the legend FEL TEMP REPARATIO and point out that the very first page posted on this website when it began in 1997 was on a special Falling Horseman. After not particularly trying over a period of ten years, I discovered that I was reasonably close to owning a coin of the Falling Horseman type from each of the fifteen mints that produced them. That seemed like good enough reason to try to find the missing ones and report the progress on this page. Some are harder to find than others. My title shows 'Mint' as it does because the term "Mint Set" is used by modern collectors to refer to a group of brand new perfect uncirculated coins. A glance below will show that these coins are not in mint condition. Some of them are just plain ugly. Neither is this a 'complete' set. Some of these mints issued several hundred minor variations on the theme. I have selected some representative coins to show but a collector much better at it than I points out that there are over 2,200 variations on this type. I'll have to trust her on that one - I did not count the ones I have seen and I have not seen even a small percentage of what exists. The vast majority of these coins are poorly made, poorly preserved, poorly cleaned or some combination of all of these faults. Finding a presentable coin from each of these 15 mints will be quite a challenge. They are interesting coins including some of the most common and most rare items in the 2,200 (or so) varieties. Below, the mints are listed in geographical order as found in RIC (Roman Imperial Coinage) volume VIII. Each mint city has special characteristics that distinguish their coins.

Late Roman Bronze Coinage(LRBC) by Carson, Hill and Kent (page 108) recognized four design groups of Falling Horsemen based on the pose of the barbarian. These are:

FH1 Kneeling Horseman kneeling on ground before horse
FH2 Sitting Horseman sitting on ground before horse
FH3 Reaching Horseman on horse turning and reaching back toward the soldier
FH4 Clutching Horseman on horse slumping forward clutching horse's neck

There is some doubt in my mind whether the differences between FH1 and FH4 are distinct enough to justify two separate numbers but that will be a discussion for another day. There are also some variations on FH2 where the horseman seems to be rising to his feet or kneeling facing the soldier rather than actually sitting on the ground. The FH numbers are more a matter of convenience than accuracy of description. I suggest becoming familiar with the above before proceeding with this page. In general, we will see a similar pattern for each mint. Most Falling Horseman coins were issued in the name of Constantius II. His long reign saw both the beginning and the end of the series. Earliest coins (c. 348 AD) will be AE2 (21-25mm) size weighing about 5g. Often these will have an A in the field of one or both sides. Small changes in weight and diameter came slowly with changes in the field letters. Before long, the average weight was down a gram but most coins still qualify as AE2. A serious revaluation followed and the coins shrank to AE3 (17-21mm) with weight dropping into the mid 2 gram range. This occurred while Constantius Gallus was Caesar (351-354 AD). His first coins are large and his last are small. Again small reductions would follow until the last coins of the type barely qualify for AE3 and weigh in the low 2 gram range. These last coins tend to be crude suggesting that they were produced in a great rush. Various mints had different markings that accompanied some of these changes and the exact dates for each step is beyond the scope of this page. Near the end of the reign of Constantius II (361 AD), the falling horseman type was replaced with even smaller coins showing the emperor standing. At the beginning, Constantius II was co-emperor with his brother Constans who died (350 AD) before the coins shrank. We will see a few large coins of Constans but most Falling Horseman coins belong to Constantius II. Constans showed a preference for another type showing him standing in a galley steered by Victory. After the death of Constantius Gallus, Constantius II was joined by his newly appointed Caesar Julian II who did not arrive on the scene until after the major reduction. All of his coins are small. That leaves only the very rare FH coins of the usurper Magnentius which are of large module. I do not have one of his to show. We will not be reviewing all issues of the other four rulers but just take a quick look at samples from each mint. This page has omitted attributions of the horsemen but it seems obvious that there were several tribes of enemies represented on the coins. Each had their own distinctive hair styles and attire that certainly would have been more generally recognized in that day than it is now. It is a matter for more study. Reverse legends are FEL TEMP REPARATIO (Felicium Temporum Reparatio) which is best translated 'Happy Days are Here Again' at least for as long as anyone lives that will get that joke. 'To the restoration of happy times' seems sterile but is a bit more literal.

Some effort has been made to present the images below in proper proportion to each other. However, individual specimens of the same coin series will be found to vary in diameter. Some mints tended to smaller diameter and thicker coins of the same weight while others were thinner and wider. When mentioned here, weights and sizes are approximations.


Amiens (Ambianum) mint was founded by Magnentius and was closed soon after Constantius II regained control of the city. All FH coins of this mint are AE2 size. While RIC rates our coin (RIC 48, page 124) as common, this will not be the easiest one to find should you try to duplicate the set. The mintmark used by its one and only workshop is AMB which is easy to confuse with the second officina of Antioch which used ANB. Some, but not all, Amiens coins have an A behind the bust making these easy to separate from Antioch shop B. The A indicates the high weight standard and is not a city mintmark. The plain ones will require you to pay attention to styles which, fortunately, are not terribly similar. Amiens portraits often have a heavy jowl look and three distinct bands at the back of the neck. Our example shows the common horseman reaching pose 'FH3' with no clear hair detail on the barbarian.


Trier was the last mint I found and the one that triggered the creation of this page but my celebration had a false start. The coin on the left is listed in RIC as #350 page 167 and should weigh 4.25g. This one is 4.12g which is within the range of acceptability. It arrived in the mail on the day this is being written. A few months ago I got the one on the right which matches the legends and types but weighs only 2.0g. This is light even for a reduced coin of the next issue (RIC 358) which should not have the A behind the bust. This smaller coin is a barbarous copy so my hopes of filling the 'Mint' set went unfilled that day. I would have preferred an earlier, full weight FH2 to fill this slot but for now this middle period but official FH3 (ranked by RIC as common) will fill this slot giving me a coin from all 15 mints to show on this page.

UPDATE: Two years after posting this page, I acquired this Trier mint (TRP) coin of Constantius Gallus. While that alone made the coin interesting, this specimen is made more appealing by its huge 25mm, 5.85g flan. While weights of late Roman coins can be quite variable, this one makes me ask questions that can not be answered. Is there an explanation for the coin being so heavy, broad and thin? Certainly nothing can be proven but the flan is what I would expect if it were prepared by hammering flat another coin of slightly greater weight standard. The previous issue from this mint was for Magnentius and many were of about this weight. Was this flan recycled from politically incorrect stock at the mint? It is a "fun" theory but nothing can be proven. If nothing else, the coin is an unusual example with full border of dots on both sides.


Compared to our previous two Western mints, Lyons (Lugdunum) was a large operation. However in the early period, when Constans controlled the mint, relatively few coins of Constantius' Falling Horseman type were made. On the left is a relatively rare example of the FH1 design (RIC 79 page 182) where the horseman is on the ground kneeling. Note that the horse's head is extended to the left rather than being tucked under as is common on most other varieties. Our second coin (right - RIC 102 page 183) is marked common but may not be easy to find. We see the horseman seated on the ground raising his hands toward the soldier about to kill him. This seated version is the classic FH2 pose. Style and workmanship at this mint are good and high grade coins can have good eye appeal. I particularly like the horse's tail which sticks straight out more like a tree branch than a tail and the hair details on the emperor's neck. The mintmark SLG* indicates the second workshop in the two workshop ordinal system (Primus, Secundus). There are later and smaller coins of this mint but I do not own one to show here. Similarly, I lack any coin of the other rulers. Remember that the number I quoted of 2200 variations will make many gaps as this small overview page progresses. Both of our samples show the horseman wearing a tall, floppy cap.

Arles - Constantia:

Continuing on our tour of Western mints we see a coin of Arles. PARL indicates the first workshop of three (adding the option of T for tertius). Again our example is an FH2 'sitting' but this horseman seems to be struggling to regain his footing. Even more distinctive to my eye is the horse's rear decorated with what I'll call Christmas balls. As with the other coins we have seen so far this is a large 'AE2' size denomination from the earlier part of the FEL TEMP REPARATIO period. Again RIC lists this (#121 page 211) as a common coin and again I'll suggest that you will not have an easy time finding one compared to most of the horseman versions. Here we find a unique situation in our list of mints. In 353, Constantius restored the city name Constantia. It had been changed to Arlate in 340 (before the FEL TEMP REPARATIO period). Our next coin (RIC 215 page 219) was struck in the same city but under the new name so we see the first workshop mintmark PCON (not to be confused with Constantinople which we will visit later). Finally we see a coin rated as common that really is common. It also is about half the weight of the first four coins we showed. Horseman coins started declining in weight and diameter until the last issues were nothing like the large, well made coins that started the series back in 348. Our example is FH3 'reaching' and wears a brimmed hat.


It may seem odd that the coins of the capital city are not the most common of the series. Perhaps more of the resources at that mint were devoted to the production of gold and medallions. The base metal coinage was centered in the provinces. Still, there are enough reasonably common Rome mint Falling Horseman that it should not be too hard to find an example. During this period, Rome used a mixed system of six officinae with the shops one, three and four (Primus, Tertius and Quartus) indicated by ordinal initials. The second, fifth and sixth shops (B, E and S) switched to Greek numerals to avoid confusion caused by the fact that the first letter of Quintus had already been used in the series as four and B is clearly two while S might be confused between Secundus and Sextus. Our early FH2 coin (left) has the mintmark RF.E indicating shop 5 (I do not understand the use of the F). R is used for the city initial but there is often some device or letter between the initial and the workshop number. The second and fourth coins have a Q (4th shop) while the Constantius Gallus (3rd from left) is shop S. Perhaps indicating some sense of humor at the mint the S was assigned to shop six. S could be the ordinal Sextus but our coin shows a squiggle shaped letter more likely intended to be the Greek numeral for 6 which modern type faces lack as part of the normal set. We need to remember that these markings were intended for mint accountability and not to make life easy for 21st century coin collectors.


Aquileia used a three shop system indicated by ordinals (P, S and T). These letters were placed after the AQ city abbreviation. We have seen that not all Falling Horsemen belong to Constantius II. Catalogs list a rare Aquileia mint Falling Horseman for Constans but I have not seen one. At the lower left is a coin of Constantius Gallus bearing the field mark LXXII (72) indicating that 72 of these were to be made from one Roman pound of metal. The lower right coin is of Julian II as Caesar (RIC 225, page 336). Shortly before Constantius II died and Julian became Augustus the Falling Horseman type was eliminated and replaced by even smaller coins showing the Emperor standing. On his accession as Augustus, Julian instituted a currency reform with completely new types. Coins of the Caesars can also be spotted by the lack of a diadem on the portrait. Julian and Gallus here are shown bare headed. Behind Julian's portrait is the letter M which indicated one of the last reductions in size and weight for this series. Some have proposed that this indicated that one thousand (M) of these coins were worth one gold solidus but this is not a certain fact. There are coins issued after the M series with even lighter weights and certainly even less value in exchange for gold. More frequently we see these late issues struck in a sloppy way with ragged edges and small flans that lose parts of the legends. Most mints placed the M on the reverse. Coins of this mint are common.


Another extremely common mint is Siscia. Our examples show one coin each for the four rulers and a spare Constantius II. Workshops were indicated by a Greek number before the SIS city abbreviation. The top row shows early, full size FH2 issues for Constans and Constantius II. At the lower left a coin of Constantius Gallus shows the unusual reverse field mark LXXII indicating that there are 72 of these coins in one Roman pound of metal. Perhaps you may recall LXXII was also present above in my coin of Constantius Gallus from Aquileia. These are also found from both mints for Constantius II but I do not own one of either. The bottom row (later) coins also show distinctive open top letters A (looking like H but not pronounced as an Eta) and the local Siscian alphabet symbol (a reversed Z) following the ASIS. The Julian shows M in reverse field indicating yet another reduction in size from the weight of the Constantius.


The Sirmium mint (closed since 326) was reopened in 351 when Siscia was lost to Magnentius and remained active even after the control of Siscia was regained. Styles of these two mints and the shop letter system are similar since Sirmium was started using personnel moved from Siscia. Compare these coins with the Siscia examples above. The Constantius Gallus (upper left) is earlier and AE2 while the Constantius II (right) is a mid-later period AE3 size. The greatest reduction in weight took place during the time of Gallus so his coins, like those of Constantius II, are found in both sizes. Julian II Caesar coins (below left) are all of the smaller module. Note the distinctive neck curls on both of these well executed later period portraits. This coin bearing the M in the reverse field is slightly smaller than the Constantius II and half the size of the Gallus.


Of the fifteen mints, Thessalonica offers the greatest variety of Falling Horseman types. This mint, at one time or another, produced all four of the FH types. Distinctive among the early issues are the FH1 coins (top row for Constans and Constantius II) that use a right facing obverse portrait showing the emperor holding a globe. This is similar to the common left facing portraits used on the middle denomination at other mints. My coins at the left of the two bottom rows show the FH2 'sitting' style with different field letters separating them by weight standards or series. My small coins are FH3 'reaching'. Unfortunately, I do not have one of the later FH4 coins for my image. Early period coins of Thessalonica used a mintmark with Greek numeral (A through E) following TS to indicate the city. Stars and dots were used to distinguish a sequence various series. Distinctive to this mint alone is the reduced size series shown in my image by the pile of one obverse and five reverse images. This series marked the five workshops (officinae) with a Greek numeral in the reverse field. Other mints, and this mint at other times, used this position for letters indicating the weight standard of the coins. This gets a bit confusing since these same letters (most commonly gamma) were common field letters for larger series. Notice that these coins use the mintmark SMTS (Sacra Moneta Thessalonica) in exergue without a following workshop letter. Compare these to the even later coin of Julian II (bottom right) which uses SMTSE (E being the shop letter) and M to mark the series in the reverse field. Also note that this mint used quite a variety of horsemen hair styles and caps. It was unusual for a mint of that day to change so many details so frequently. A complete set of Thessalonica horsemen considering every possible variation will be a lot of coins.


Heraclea follows the pattern of the other mints perhaps a little more strictly by issuing Falling Horsemen only for Constantius II in the early period and Galley types for Constans. Many mints favored these assignments but occasionally crossed back and forth with varying degrees of frequency. A point to be made here is that it is necessary to be careful to distinguish coins of Heraclea from coins of Nicomedia since the angled middle of the N and straight cross of the H can be less than obvious on some coins. I am less than confident in making this call on many coins. Heraclea issued a few FH4 coins (upper right) in the early period but no FH1 or FH2. Most coins from this mint will be FH3 'reaching'. Mintmarks used SM (Sacra Moneta) followed by the H initial and a Greek numeral (A-E) for the five shop system. One has to wonder is 'Sacred' money carried a greater penalty for counterfeiting than 'Regular'?


Considering that Constantinople was the great city of the empire founded by Constantine as the new Rome, I am somewhat unimpressed by the Falling Horseman coins issued there. There are many varieties with many field letters and weight changes but style of engraving is usually not very good. In particular the horseman often reminds me of a frog and it is not always obvious when a leg belongs to the horse or the man. The portrait work is better than the rider. Our examples include (top right) the mid weight series with .S. in the reverse field as issued by several mints and the unusual smaller coin (bottom right) showing an E in the field but the letter is made with a forked center bar not seen elsewhere. I am not clear on the exact meanings of these standards marks. Constantinople was a large mint using eleven workshops designated with Greek numerals (A through IA) following the CONS mintmark. This is convenient in separating the coins from Constantia (Arles) which preceded the CON with an ordinal initial (P=Primus, S=Secundus). Constantinople coins are very common.


Nicomedia was another prolific mint of no particular merit to my eye when it comes to cutting horseman dies. None of my coins strike me as having been produced by two well done dies. As mentioned above, care need to be taken to separate these 'N' coins from the 'H' issues of Heraclea. Care was taken in the early period to avoid issuing Falling Horsemen for Constans. For reasons unknown to me, the largest coins from the earliest Falling Horseman period are rare from Nicomedia. I do not have one. The coins become common with the intermediate issue bearing gamma in the reverse field. Of my examples I will point to the upper right coin of Constantius II with what I would call a clumsy portrait but well struck reverse compared to the average for the M series. Nicomedia is another mint that has failed to attract my praise and attention. Perhaps I ask too much from the big city mints that undoubtedly worked at a rapid pace to keep up with the demand for coins.


Cyzicus is another mint with rare coins before the gamma series. It is yet another mint suffering from a lack of artists who could draw a proper horseman. Our examples, however, include two extraordinary coins. The Constantius Gallus at the lower left still bears much of its original silver plating. All of these coins were issued with this silver wash but few retain even a trace. This example must have been fortunate in terms of where it spent the last 1700 years to be able to have so much silver. At the bottom right is another exceptionally well struck coin from the usually dumpy M series. Note that it also has the open top A's mentioned above in our paragraph on Siscia. The reverse of this coin might be considered a gem were it not for the body of the horseman consisting of three bent lines. Cyzicus also used the SM plus city initial (K) plus Greek numeral (six shops) mintmark.


To atone for all my complaining about Falling Horseman numismatic workmanship, I must now praise the mint workers at Antioch. With regularity, the quality of the reverse dies for the early, larger coins is well above average. The best examples are beautiful. In addition, many Antioch coins are reasonably well struck although my examples are not particularly good in that respect. Well struck coins of the best style do sell for a premium to persons who want only one example of this coinage. Antioch was a very large mint using a system of fifteen officinae numbered in Greek numerals (A through EI) following the city abbreviation AN. An odd feature of this mint is that all of the shops produced both FH3 and FH4 designs at the same time. Many mints switched between the FH styles but most did not make both at once. Our top two examples show my collection's best FH3 (left) and FH4 (right). Portraits of Antioch are also well done on the average. While coins of the other three rulers exist, I do not own even one of them. RIC rarity ratings suggest that Antioch made many more coins for Constantius II than it did for his co-ruler throughout the Falling Horseman period. Later, smaller coins (bottom row) are not as well executed but still well above the average of the other mints. Antioch mint coins also provided the question of stirrups being shown on these coins. That provided the subject for my very first page posted on this website when it began in 1997.


Falling Horseman coins of the Alexandria mint were produced in a style distinctive to the city. Portraits are tall and stately while the reverse scene and lettering can both be recognized as Alexandrian even without reference to the mintmarks. Mintmarks consist of the city abbreviation ALE followed by a Greek numeral. There were four officinae but delta (4) seems to have been smaller or used less consistently than the other three. The gamma series Constantius II shown at the top of my image has long been one of my favorite Falling Horseman coins. It shows a distinct style foretelling the direction of art as it moved from the Roman into the Byzantine period. While the portrait is not the most realistic, it is, in my opinion, beautiful in its formality. The Julian II FH4 at the bottom shows just how low the workmanship had become by the final years of Falling Horseman production. It is hard to find an example well struck but this one is really miserable. Its only merit is that the part of the coin showing the mintmark and field M happened to get struck well so the coin can be identified. The rare earliest coins of this mint used the FH4 design as did the last issue only for Julian II as shown here. In between Alexandria horsemen were FH3. I would be interested in knowing why one workshop (delta) alone made this change at the end just before the type was retired. My example weighs only 1.7g which is more in line with the Spes coins that replaced the horsemen. I suspect this coin carries the honor of being my newest Falling Horseman.

That brings us to the end of our survey of fifteen mints that made coins of this the most common Roman coin type. Were they rare, the action scene reverse would make them popular and expensive but as it is they get little respect from most dealers and collectors. Finding all 15 mints has been a chore since many dealers ignore such common coins.

.....but we are not quite finished.


In addition to coins of the fifteen official mints we find coins of the Falling Horseman design made outside normal channels. Some, like one shown above in our discussion of the Trier mint, are reasonable copies of official coins. Others, like the two shown here, are quite obviously barbarous. Coins of both classes (and everything in between) might be the work of counterfeiters intending to profit from their enterprise but others are most likely local products in regions that were short on coin supply but where official coins had made the people used to a coin economy creating a demand for something, anything, to serve in commerce. These are more like tokens than fakes and were important in the daily economy of their regions. My observation has been that there are more unofficial coins bearing Trier markings than any others suggesting that far off Germany was not well supplied with the real thing. The two coins shown here are the wild end of the unofficial range. Most legends degrade to marks but the style of portrait and the reverse scene are only a bit worse than the low end official coins. Barbarous coins do come much worse. I suspect they were made to circulate as coins somewhere beyond the limits of the Roman Empire where they were the only available form of coined money. Barbarous coins are not at all scarce but appeal only to a relatively small group of specialists. There is no way to tell if they date to the same years as 'real' horsemen or if they provided coinage long after the Romans had left the area. There is much room for study in this field of numismatics.

This page is dedicated to all dealers, collectors and friends that have helped in my journey of fifty years as a student of ancient coins.

Return to main page

(c) 2012 Doug Smith