From my own ball making experiences I know it is very difficult to predict how big a finished ball will be and what it will weigh because of the nature of the materials used. All pieces of leather are unique, and variability is present across a single piece of hide in terms of the thickness and the amount it will stretch when subject to load. Therefore, even if a ballmaker takes the utmost care to cut pieces of leather to exactly the same dimensions as a previous ball they are likely to end up with balls of different sizes. This is compounded by the nature of the feather filling and the skill and perseverance of the maker. I know that one must keep stuffing in more and more feathers even when it seems impossible. When I first started making balls I did not appreciate the degree to which a ball must be stuffed and it is hard to convey it in words. Simply put, for a ball to be effective, it must be stuffed until no more feathers can be inserted using humanly-applied force and the weight can easily vary by a couple of grammes according to the determination of the ballmaker. 

Ballmakers would’ve cut the leather to a set of dimensions they knew would produce a credibly round sphere when sewn together. They may have used a knife, scissors or a sharpened “cookie-cooker” metal tool to cut out the leather pieces.  Through experimentation I have found that identically sized leather pieces from the same hide can result in balls that differ in weight by as much as two pennyweights. Whilst two pennyweights (3.1 grammes) is significant it is clearly not the reason why extant balls vary between 24 and 34 pennyweights. Drawing on my own experiences I am certain that ballmakers did not stuff balls until they reached a target diameter, measured by calliper, and then stopped as concluded in 1896 by the author of the article in Golf. I know this leads to a ball that greatly underperforms even though it might feel quite full of feathers. Callipers do feature in some old photographs of ball-making tools but I believe the wrong type has been included in error. I know that callipers with round ends are useless, but those with pointed ends are extremely useful for marking out the round end caps and for running down the edge of the leather hide to mark out central strips of consistent width prior to cutting out when using a knife.    

Putting aside relatively small variations caused by the variability in leather and the manufacturing process, if a particular size and weight of ball was optimum for all conditions, then surely all makers would have aimed to make such an item. Modern technology has allowed us to study and understand the physics of golf balls to a high degree yet much still remains to be learnt. Hundreds, if not thousands, of academic papers have been written concerning golf balls, and I’ve read several dozen of them. Millions of pounds are spent annually across the sector as companies strive to gain a competitive edge and much of what is learnt isn’t published but is kept as closely guarded trade secrets. It is tempting to wonder whether the feather ball-makers had secrets of their own that they did not share with their competitors. After all, if a particular ballmaker found out how to make a ball that travelled ten or twenty yards further than those of his competitors he would’ve corned the market (as it would today!). For now, let’s consider balls of different sizes produced by particular makers. Within the sample studied there are balls made by Allan Roberston ranging in weight from 27 to 29 pennyweights. Balls of 24 pennyweights made by him exist, and a ball of 30 pennyweights was sold at auction in December 2025, therefore, he did make balls of considerably different size and weight. If we consider balls by William and John Gourlay, either working together, or by John alone after William’s death in 1844, they range from 24 ½ pennyweights to 34 pennyweights – they too clearly produced balls which spanned the range of weights used. From these observations we can deduce that a range of sizes were deliberately produced, and given Kincaid’s diary entry, this practice was likely done for many centuries.

According to McGimpsey & Neech³ “players chose balls of differing sizes and weights to suit the course being played or the weather conditions”. Kelly⁴ states, “golfers selected balls based on their skill level, the lay of the course, and the day’s playing conditions.” Picking up on Kelly’s point about skill level I think it likely any novice golfer would select a larger ball over a smaller ball when taking their first tentative strikes if given the choice, but I do not believe that once some proficiency is gained there would be a trend for less skilful golfers to use larger balls, and vice versa. I think that the ground conditions and weather are much more important. Inland soils are richer in nutrients and retain water more than coastal sandy soils and this directly influences the types of grass that can be supported. As a general rule grass on inland courses, such as at Brunstfield,  would’ve been broader leafed and longer in length than that on a coastal course such as St.Andrews. I have test hit 1.62 and 1.68 diameter balls on an inland parkland course in May and in my opinion the larger ball is far easier to play with when the grass is lush. All other things being equal a smaller ball will travel further due to less air resistance (drag) and this seems to have been understood by Kincaid in 1687.  So what is too small? The 10 ½ pennyweight ball, listed within Figure 1, was described as a “souvenir, and not for actual play”. A variety of miniature antique items exist and these are sometimes labelled “apprentice pieces” or “salesman samples”. If a youth was training to produce very large pieces of furniture, or a travelling salesman was visiting potential customers, then the production of a small-scale version would save material cost and facilitate much easier transportation, but with an item such as a golf club or ball neither would be necessary. Children’s clubs from the longnose era do exist, including one early example made by Cossar of Leith. In my opinion this ultra-small ball was either made for a small child to use, or simply as a not-for-play whimsical keepsake as it is too small and too light to be used effectively by an adult. 

James Dundas of Dundas Castle (b.1793) clearly didn’t just play with balls made by one maker, or of a consistent size, as within the batch of balls stamped with his name and sold in December 2024 there are two balls by John Ramsay, one being a 27, another a 29, together with a 30 pennyweight ball by Allan Robertson. I do not know the golfing habits of Dundas but it may be that he played golf in St.Andrews where Robertson lived, as well as at Musselburgh where Ramsay lived. He could’ve purchased these different sized balls to suit the ground conditions of where he was playing golf, or simply because he needed some new ones and they happened to be available. Some modern golfers today are very picky when it comes to the make and model of ball they use whilst others are not – we do not know how picky Dundas was! Allan Robertson made balls of all sizes, however, a study of auction sales over the last forty years indicates that the majority of extant balls by him are of size 26 or 27 – very few much larger or much smaller balls by him remain. This might be because these mid-size balls were suitable to the ground conditions at St.Andrews. Allan was renowned for playing with lightweight “toy” clubs⁵, and having personally inspected one his clubs I can confirm it is indeed very light. It is logical to assume that these would suit lighter-weight balls but without testing this is speculation. However, it would be wrong to assume that just because Allan preferred lightweight clubs and balls that all golfers at St.Andrews did.  

The survival of very light, or very heavy balls, could be because golfers didn’t like playing with them that much and they were left forgotten in desk drawers, etc, before being superseded by the new gutta percha models.

I have searched for historical texts which give clues as to why balls of different sizes were used but have only managed to locate two and these were written over a century apart. In 1790 Hoyle⁶ states, “Light balls are used when playing with the wind, and heavy ones against it.” In 1903, Alexander H. Doleman, wrote in Golf Illustrated⁷ “players in the olden days played with balls of great variety in size. Some played with light balls, some medium, and some with very heavy balls”. He continues “in the early days the powerful driver would often use a 29 or even a 30, while the weaker player would use a 26 or 27. The heavy ball was thought to give more advantage on the putting green, which was considered to outweigh a few yards loss in the drive”.

Hoyle’s statement is very puzzling. Modern scientific studies show that a golf ball is more affected by a headwind than by a tailwind^8. In simple terms you lose more yards of distance when hitting into a wind than you will gain by hitting with it, and this difference increases as the cross-sectional area of the ball increases. Logically, it would make sense to use as small a ball as possible when hitting against the wind providing it has sufficient mass to give it adequate momentum yet this is opposite to what Hoyle is advising. As with other late 18^th century texts concerning golf, it might be that this entry incorrectly reports an aspect of golf and we have no way of knowing how this “fact” was gleaned by Hoyle. If this was indeed common practice amongst golfers of the time they must have done so because of empirical evidence which is even more puzzling. It may be that a heavier ball was played against the wind because it is more likely to fly straighter than a lighter one.     

Doleman was born in 1836 and alongside his brothers he became an expert at the game. It is very likely he learnt to play as a youngster with featheries and would’ve been about sixteen years old when the gutta finally ousted the feathery for good. Unfortunately, he doesn’t make it clear in the article whether he is specifically talking about feather or gutta balls! The numbering system used on featheries was initially directly carried over onto the guttas in that it measured the weight. We can therefore apply his statements to the practices of those playing golf in the last years of the feather ball and the early years of the gutta ball. (In later years the numbers on molded guttas referred to the diameter, i.e size, which can be very confusing!). He does not refer to the wind direction at all as Hoyle did. He infers that weaker players had to use light balls to gain any credible distance - this might be true. He states that “powerful” players, presumably those with faster swing speeds, used heavier balls. As stated earlier a heavier ball will possess more momentum yet due to it being larger will present more area to the air molecules resulting in more drag. Doleman appears aware of this when he states that whilst heavier balls putted better they did not travel as far when driven. The statements made by Hoyle and Doleman are useful but not comprehensive and so we should consider other factors.

The faces of feathery era wooden clubs were typically between 15/16 and one inch deep. This is not too dissimilar to many shallow fairway woods produced today and facilitates a strike which is low down in relation to the centre of mass of the ball, resulting in more spin and a higher flight. The same effect is possible if a slightly larger diameter ball is used. Given the range of ball sizes on offer it’s likely that golfers of the feathery era were aware that different sized balls would fly higher or lower.

From experience I know that a smaller feather ball is trickier to make than a larger ball; it requires more nimble fingers. I think it likely that the first balls made by young apprentices, barely in their teens,  would’ve been on the large side, but I see no evidence that ball size was linked to skill level for experienced makers. The quality of stitching, and roundness, is consistent on balls of all sizes be they large, medium or small – they clearly possessed the skill to make balls of different sizes as required by their discerning customers.    

One such customer was Thomas Kincaid and an entry in his diary suggests balls of different densities were produced in the late 17^th century. As no extant balls can be reliably dated to that time we are unlikely to know whether this is true or whether it is purely Kincaid’s perception of balls on offer.  Having measured and weighed forty-five balls made during the first half of the 19^th century I was able to produce the following graph to illustrate the density of each ball.    

 

Figure 2. All balls inspected - average diameter versus weight.

All of the balls considerably above the red trendline have reference numbers beginning MLC denoting Museum Loan Collection. So are these balls high-compression long-hitting models? The answer is no. In Part 2 of these articles I showed that four of these MLC balls, complete with name stamps and weight numbers, had been exposed to linseed oil and their weights had consequently increased beyond that expected. The provenance of these MLC balls is not documented. This graph includes all the MLC balls without makers names, and it is very clear that all of the MLC balls are over-weight. Most are 4 or 5 grammes over which is consistent with two coats of oil, though some are considerably more. No MLC balls are near or below the red trendline and a clear gap exists between them and the rest of the balls in the sample. I believe this is further proof that all of the MLC balls have been exposed to linseed oil for purely display preservation purposes because if “oiling after painting” had been done for performance purposes then other balls in the sample would weigh more than suggested by the numbers written on them – yet none do. Checking my notebook I had recorded that many of the MLC balls were dark in colour, or felt sticky, or had gummed-up seams, but I did not realise at the time that there was a distinct pattern – I only realised this after I had started to write these articles. I am convinced these balls are definitely not in the same condition as when they left the hands of their makers. The graph below excludes the MLC balls and the reference numbers are replaced with the makers names where known.  As expected the red trend line moves to reflect the altered sample size. 

Figure 3. MLC balls excluded – average diameter v’s weight. Known & Unknown Makers.

The updated trendline confirms that weight increases as diameter increases. The volume of a sphere changes in a non-linear relationship to the radius as per the formula  and a slight upturning of the trendline can be seen which I believe is reflective of this, but it is not pronounced. 

We can see that no particular maker is revealed as making the highest density balls. Balls by the Gourlays, Tom Morris and Alan Robertson are about the same distance above the red line, and two of Allan’s are below it. This graph indicates that all these makers were masters of their craft and were very capable of producing very good balls. 

The data point lying furthest under the trendline is for ball 4. It was found in the 1950s in a wall cavity in a house in Market Street, St.Andrews and retains about 5% of its original paint. Oil paints of the period were made by mixing ground up white lead with linseed oil and through experimentation I have determined that a single coat of paint can increase a ball’s weight by up to 3g. This ball is of average diameter so I do not believe the leather has expanded by any significant amount through being used and that its position on the graph is simply due to paint loss. The proximity to the trend line of all the used balls suggest they are within a fraction of a millimetre of their original sizes. Through my ball making experiments I have come to realise that very strong leather that resists stretching is required and I have often wondered at the validity of statements which say that balls became soggy or flaccid very quickly in wet conditions to the point that new balls were used at every hole. I have carried out some rudimentary experiments to test how quickly a ball absorbs water when resting on waterlogged grass. No flabbiness was noted for balls that had sat for up to twenty minutes, which suggests these accounts are not accurate and that feather balls could be used successfully in wet conditions. Whether they could be used immediately after being totally submerged in water was not tested and this will be investigated in future experiments.   

My earlier article (part 1) confirmed that used balls are as round as new balls and the graph above indicates that the density was also maintained. This indicates that feather balls could withstand being hit repeatedly; I successfully hit a replica ball I had made several dozen times with no discernible damage. The ball was tee’d up and hit using a modern driver. Further experiments are planned using a replica ancient spur-toe iron and it will be interesting to see how the ball performs.

The weights and diameters of the feather balls in the possession of Douglas McEwan in 1896 deserve further investigation. All of these, except the ultra-small souvenir ball, are shown in yellow on the figure below. 

Figure 4. Balls owned by Douglas McEwan added to the study

Despite the diameters being described in 1896 as “only roughly approximate” the graph shows the size and weight of balls owned by Douglas McEwan are broadly in line with all the non-MLC balls that I have personally inspected. The weights clearly increase as the diameters increase, so the observation of the original author that there is “near agreement in their diameters” is demonstrably false, and undermines his conclusion that they were made to a “uniform measure” in terms of diameter.  In line with Figure 3 the balls of unknown makers are below the overall trendline, and are likely to have been used and suffered paint loss. It was noted that ball 55 “seems to be a very old ball”, and although labelled “T.Dunn” this was most likely made by either his father Willie or his Uncle Jamie who had been feather ball makers when they were young. This ball is quite far under the trend line as are balls 46, 47 and 56. If ball 55 had been the only one to be considerably under the trendline it may have indicated that very old balls had lower compression than younger specimens but as this is not the case then the reasons for it can only be speculated – the condition of these balls was not recorded but it is likely they had suffered paint loss.   

One of the MLC balls intrigued me when I checked back through the photographs I had taken because it has good paint coverage and no obvious signs of being covered in linseed oil despite being 6g heavier than expected for its diameter. It had been catalogued as being of unknown maker but I had managed to make out the feint stamp of W&J Gourlay using a magnifying glass and I had also made a note of “possible repaint”. A closer inspection of the photograph reveals all. This ball (MLC04 019) has two distinct paint layers with noticeable small ridges which I believe are palm prints.  This ball provides evidence that ballmakers repainted, restamped and resold balls. Whether the new owner would be aware of the 6 grammes of extra weight he’d have to make fly is another matter.  (As an aside small ridges in the paint caused by fingers and palm prints are not uncommon – investigations could be carried out using today’s technology to match these between balls and other artifacts such as old clubs – in the future it may be a very easy and simple thing to do to prove provenance).  

Figure 4. Over heavy re-painted and restamped ball by W&J Gourlay (Ref MLC 04 019

Conclusion

The processes and materials involved in making feather balls meant that their final weights could unintentionally vary by as much as two pennyweights. The final size, and weight, of a ball was dependant on the dimensions of the cut leather pieces, how much they stretched, and how much paint was applied, providing it was stuffed correctly. All the results indicate that the amount of feathers used was not altered to give a ball of a desired diameter contrary to past belief; they were all stuffed as much as human force applied to a stuffing tool would allow.   

There was clearly demand for balls of significantly different size and associated weight. There is anecdotal evidence that golfers selected balls based on their skill level and the direction of the wind although these historical statements are somewhat confusing. Doleman relates that larger sized balls offered advantages on the putting green at the expense of distance. My own rudimentary experiments confirm that larger balls are easier to hit than smaller ones when the grass is lush. Clubmakers may have tweaked the design of their clubs to suit the most common size of balls being used in their locality and this is something which could be investigated further. Very keen players, as many undoubtedly were, may have purposely selected balls of differing sizes to suit their equipment, the weather, the grass conditions and the layout of the holes they intended to play. It is easy to think that the players of the feathery era approached the playing of the game in a much more relaxed manner than we do today. This to my mind is false-thinking given that we know how often some of them played, how much they spent on their equipment and on wagers, and how fastidiously their scores were recorded – they likely strived to do better and win just as much as some golfers do today.

All the main ballmakers were capable of producing balls that spanned the useful range: 24 to 34 pennyweights. Golfers such as John Dundas, purchased balls of different sizes from different makers; whether he did so deliberately or by happenstance cannot be proven either way. There is no doubt that some extant balls are over-weight as a result of being painted or soaked in linseed oil and this may have confused earlier investigators.  No one ballmaker stands out as making higher density (further flying) balls than any other. The results show that well used balls can retain their roundness and density though their weight will be reduced if the paint becomes detached. There is evidence that the Gourlays repainted, restamped and resold used balls, and it is almost certain that others would have done likewise. In my next article I will examine the proportions of feather balls and some theories about balls made in the 18^th century.

This article was made possible with the support of The Wilson Family Research Fund and the R&A World Golf Museum and the author wishes to express his thanks to these organisations.  Sincere thanks also to fellow BGCS member Peter Grunwell of Fine Golf Books.

Gavin Bottrell

August 2025

References

1.     Golf, A weekly record of “Ye Royal and Ancient” Game. Article 20^th March 1896 “The Weight of The Golf Ball II”. London. 

2.     Thomas Kincaid’s Diary, National Library of Scotland, Adv.MS.32.7.7

3.     Golf Implements and Memorabilia. K. McGimpsey & D. Neech, 1999, Pp20

4.     Antique Golf Ball Reference & Price Guide. Leo M.Kelly Jr. Old Chicago Golf Shop, 1993.  Pp.2

5.     The Obituary of Allan Robertson, Dundee Advertiser, 8^th September, 1859.

6.     Hoyle’s Games 1790 Edition.

7.     The Rubber Ball Controversy. Doleman, A.H. , Golf Illustrated, February 1903.

8. The Science of Golf, Wesson, J. 2009. Oxford University Press