Cylindrical flutes - a History of Holes
We are perhaps inclined to believe that the modern Boehm flute hasn't changed that much since when Boehm released it on an unsuspecting world in 1847. But as you'll see, on the unlikely topic of hole sizes, nothing could be further from the truth.
Size and Gradience
Once we start looking into the matter, it becomes clear that there are two substantial characteristics to be considered - the size of the holes, and the degree to which the size varies as we move down the flute. I've nick-named the later "gradience" for the time being - perhaps a more regular word will recommend itself. The nature of that gradience is also of interest - is it smooth or lumpy. Perhaps we should call that "deviance".
Give me a chart!
An issue like this lends itself easily to a graphical presentation, and we have one for you below. From left to right, we have the main holes of the flute, called by their international note names. You'll recognise them by their position on the flute. Note that only two of the flutes go down to low B, the others terminating at low C.
Vertically, we have the degree of venting, expressed in percent. Venting is measured as the hole diameter divided by the bore diameter at that point. The terminating note (low B or C) achieves 100% venting, as the hole and the bore are the same thing. All the other holes are smaller than the bore, and so get a value less than 100%.
At a glance, the image is bewildering. It seems every likely combination of size and gradience has been pressed into use, and this by the world's most prestigious makers, not back-yard mechanics. We can draw from this that hole size isn't critical, or most of these flutes would never have sold! But it is important, and we can probably learn a lot from studying its history.
A bit about the flutes
I don't have too much information yet about some of these instruments - as it comes to hand I'll add it to the list below. And I can't be sure to get them into precise chronological order, but I'll do my best until we can find out more.
And a hypothesis on development
Can we determine from the evidence of the holes why changes were needed and made? We can try, but we have to be aware that this is theory and could be blown out of the water by subsequent data finds.
Siccama Diatonic flute, released in 1847. Although this page is devoted to hole sizes in cylinder flutes, for reasons that will become obvious it's handy to start with a flute that illustrates best practice just as the cylinder flute is released. Mid blue trace with hollow square markers, we can see that this flute had venting varying fairly evenly from nearly 60% in the top notes to the inevitable 100% at the bottom end. Would we be naive to believe that a similar regime might be empolyed in the new flutes about to be released?
Boehm No 1, ring-key, built in 1847, is in the Dayton C Miller collection - it has essentially uniform holes, but small. Aqua trace, no markers. We would expect a pretty stuffy response from this flute by modern standards, with the notes immediately above the bottom C probably noticeably weaker than it. Clinton was critical of the bottom D of Boehm's first cylinder flutes - could this have been the basis of his concern?
It begs the question: "Why did Boehm think it appropriate to have equally sized holes anyway?" I can imagine two reasons - German tradition and scientific theory. The Germans, in their conical flutes, always aimed for uniformity in the size of finger holes, even when it made for very dodgy tuning - they prized uniformity of response over all other things. But even in those uniformity-ruled days, they permitted the foot holes to be larger, no doubt noting the stultifying effect small foot holes has on a flute.
Indeed, because the degree of venting is proportional to the diameter of the hole divided by the diameter of the bore at that point, uniform holes on a conical bore actually gives increased venting as you go towards the foot. Clearly an argument for increasing sizes of holes as you go down a cylindrical bore! Boehm, in retrospect, seems a little blinded by the apparent scientific elegance of uniform holes. He forgets that venting (or perhaps more relevantly, lack of full venting) is cumulative, and that ideally the lower notes would be on a larger bored tube. So flexibility was called for, but Boehm didn't appear to be listening.
Boehm / Rudall & Rose, a ring-key design made in 1848, (I'm not sure who actually made it). The smallest holes so far, uniform other than for a jump in the foot hole sizes. Is this an attempt to improve the overall balance of the preceding flute by increasing the size of the foot holes? Why decrease the size of the other holes? Perhaps to assist in rebalancing the response, but perhaps simply for playing comfort - open holes larger than this being hard to cover reliably.
Boehm 38 / Rudall & Rose, 1849, made for a Jonathon Grey, and now also in DCM. (Again I'm not sure who actually built this flute). Note essentially the same hole size sequence as the Carte 1851 later on, which was definitely built by Rudall & Carte. Our first sighting of graduated holes - mildly graduated admittedly - but at least some recognition that uniformity isn't the answer. And still relatively small holes.
Boehm "Heindl" flute, No 0099 in the DCM collection. DCM records it as: "Metal flute made by Boehm himself, No. 19, about 1850. Early type of cylinder bore.... Formerly belonged to Edward Martin Heindl, who 'gave' it to Geo. W. Haynes, about 1888, and later took it back! Haynes put on a special thumb key about 1888." Probably the first metal Boehm flute to be brought to America [1864 by Edward Martin Heindl], used by him in the famous Mendelssohn Quintette Club, and in the Boston Symphony Orchestra upon its organization in 1881.
(Hmmm, note that Boehm No 38 seems to be in front of Boehm 19, according to DCM date estimates. Hopefully we can rationalise this.)
Boehm 1851 Claim. In a letter to Louis Lot in 1862, Boehm claimed to have used graduated holes in his silver flutes for 6 years, and to have played a flute with graduated holes when in London in 1851. He gives figures - 12 to 15mm from c down to C# - but confuses this by adding "always a quarter of a millimetre of graduation". These two pieces of information don't quite tally, so in the chart above, I've presented his specific end figures rather than the degree of gradience quoted by Boehm. It equates to a gradience of 0.273mm, a bit greater than Boehm's quoted 0.250.
Now note that this would give a steeper slope than the Boehm 38 / Carte 161 / Boehm 19 gradience. I don't have any data from an actual flute to back up Boehm's claim, but that may well show up. I think at this point we should consider that his memory of the figures is perhaps defective, and he was actually thinking of the slope used in the earlier flutes identified. The period certainly fits.
Boehm goes on to say that in his opinion, such graduated holes are the best, but that he dropped them for greater ease of manufacture. Strange that he did not adopt a several level approximation as we see in modern practice, to achieve the main benefits of graduation without excessively complicating manufacture.
Carte 1851 system is from serial no 161, circa 1853. Note it doesn't have a C5 hole in the same way as a Boehm, so that datum is left out. Clearly the same general approach as Boehm No 38.
Louis Lot 163, also in DCM, was built in 1857. Almost uniform hole size in the 13.5 to 14 region, with the obvious exception of c5, considerably smaller at about 12.3mm. As we'll soon see, similar to but a little larger than Boehm's Schema calls for.
Louis Lot No 208, also 1857, the only flute so far with smaller foot holes (I'm wondering if this is a typo in the information supplied). Still with the smaller c5, but with the generality of holes slightly smaller than the Schema proposes.
Clinton's Boehm flute with Graduated holes is the subject of his 1862 patent and a flute submitted to the 1862 Exhibition in London, where it won a Prize Medal. Substantially bigger holes than we have seen to date, with a substantially bolder gradience. Indeed, Clinton has made the low C# hole as big as he could, nearly 17mm, no doubt remembering that the next and final note, bottom C, is ultimately terminated at 19mm. Rockstro claims that Clinton made the c5 hole one half the size of the C#, which would make it about 8.5mm. A glance at the graph above shows this not to be true, the C5 being 13.4mm, the figure called for in Boehm's Schema, although this might be coincidental.
Notice that this is the first flute we have come across so far that gets close to the gradience of Siccama's Diatonic conical. So, while Clinton's gradience might in the context of other cylinder flutes at the time appear radical, perhaps even foolhardy, he's only restoring a pattern of venting that had been developed over hundreds of years and then abandoned with the new cylindrical bore flutes.
Clinton made several bold deviations from the general gradience too, which we will look into elsewhere when we come to examine this instrument closely. Note that Clinton used this same approach to hole sizes on other cylindrically bored flutes, so it was not limited to his version of the Boehm flute as might otherwise be imagined. Contrary to what you might believe from reading others, Boehm flutes appear to have been a very small proportion of Clinton's output (currently 1 out of 40 flutes found, ie 2.5% of extant Clinton flutes we have located.)
Boehm's Schema, first mentioned in 1862, calls simply for holes of a fixed size of 13.4mm. Considerably bigger than the holes used in Boehm's initial flutes, and no mention of graduation. It's still hard to be sure when Boehm would have developed the Schema.
Rockstro's model (build by Rudall Carte from 1864) - stoutly uniform holes (as Rockstro stridently calls for in his description of the flute), and very large at that, see the Brown line at top of chart. Indeed, here's a chuckle....
Clinton is often ridiculed for the size of his holes, yet the holes on the Rockstro model are on average much greater. While it can be argued that three of Clinton's holes are bigger than Rockstro's, the greatest excess is only 3%, while the smallest of Clinton's holes is 21% smaller. The average of all Clinton's holes (15.2mm) is less than Rockstro's by 8%. Yet, even as late as 1934, seventy years after Clinton's death, the Rudall Carte catalog introduction still comments on Mr Clinton's "extreme" holes, even though they were the manufacturer of the Rockstro model and should have known better! Hmmm, chuckle, or perhaps more likely, incompetence or humbug?
Godfroy No 940, with three hole sizes, built in 1865. Note the relative similarity to the Modern Practice trace. It would be good to have earlier Godfroy data to see when he broke away from the uniform sized holes mould. Was he influenced by Clinton, or is this a path he trod by himself?
Or is this an explanation? Lot travelled to England for the 1862 exhibition. There he would have seen Clinton's graduated hole flute and no doubt become aware of the Patent. We know he discussed graduated holes with Boehm by letter that same year - that's when Boehm told him he'd already tried them, found them better but abandoned them on the grounds of ease of manufacture. Did Lot put two and two together, and come up with his own system of graduation (similar in gradience to Clinton's but with holes of a size between Clinton's and Boehm's, and tiered, to avoid complicating the manufacture as Boehm had complained about)? And did Godfroy then snitch the idea from his countryman Lot? More data on both might clear that up.
Boehm & Mendler, 1877, still sticking rigidly to uniform hole sizes of Schema size, although permitting a significant one-step rise in the foot notes.
Rockstro model, (Duncan), post 1878. Included to show that not all Rockstro models stuck rigidly to Rockstro's formula of uniform holes at 16.3mm. Differences are pretty subtle though, the smaller holes are 15.5mm - only 5% smaller - so unlikely to alter the feel of the flute radically. So why bother? We'll hope to answer that when the flute is playable.
Interestingly, it doesn't alter the gradience - it remains at 0, because the smaller holes are distributed within the larger holes. But the average hole size is now 16mm or 84% of bore diameter (was 16.3mm or 86%). Still larger than the next largest which is Clinton's at 15.2mm (80%).
Rudall Carte Radcliff Model, No 426, made 1904. Two level gradation - body holes 14.1, foot holes 15.4
Rudall Carte Boehm No 5609, made 1914, no actual trace above but same as R.C. Radcliff shown.
Modern Practice - to provide us with an invaluable modern context. Four hole sizes showing a bold gradience generally in the area pioneered by Godfroy (Lot?) and Clinton. These figures provided by Trevor Wye and published by Larry Krantz.
The Chronology of Change
So we've established in the diagram above that massive changes to hole sizes occurred over the life of the Boehm flute to date, but it's a bit hard to see what changed when. So let's break it down into the three components - the average size of the holes, the degree of graduation (or gradience) that they display as we go along the flute, and the degree of variance from that gradience - and look at them in chronological order.
Starting with Average Hole Size, we can see Boehm started out with very small holes, around 11mm, which were quickly jacked up to around 13mm by the makers in England and France. Boehm accepts a figure just over this when he comes to publishing his Schema. Clinton appears to be the first to challenge this with a shift to around 15mm when he starts making cylinder flutes in 1862, only to be exceeded by Rockstro, to over 16mm just 2 years later. Rockstro's extravagance doesn't seem to find favour with anyone else, and other and current makers settle down for a figure about halfway between Clinton's and what had gone before.
Again we start off very conservatively - all equal sized holes - but Rudall Carte obviously felt the need to introduce some graduation immediately. Not so popular on the Continent.
Clinton's 1862 patent and his prize-winning flute certainly stands out boldly, as does Rockstro's total rejection of the "evil" of graduation. Godfroy alone appears interested in the bold graduation, with other makers including Boehm & Mendler and Rudall Carte sticking with a gradience milder than they had used pre Clinton and Rockstro. Finally, Modern Practice went with the degree of gradience that Godfroy used way back in 1865 and that Boehm claimed to have tried, preferred but then abandoned in 1851.
Boehm's later claim that he had tried and preferred boldly graduated holes around 1851 stands out oddly in the context of what else was happening at the time and what happened later. I still suspect he was thinking of the milder levels of graduation as used on the Boehm/Rudall & Rose ring-key of 1848. Is it possible though that his mention of this in a letter to Lot in 1862 might have influenced the French makers to try it out? This would be consistent with the Godfroy flute of 1865. Data from more Lot and Godfroy flutes from both sides of 1862 would help clear this up.
On the data currently available to us, it suggests that Clinton's flute and patent of 1862 is more likely to have been the inspiration for modern practice.
Well, what a bonanza that turned out to be. Strangely, I don't remember any previous publications going into it at all (Not quite true - Boehm says a bit about it, and Miller augments that, but of course they are talking only about flutes of Boehm's own manufacture, not cylindrical flutes by all of the other makers). Which is really a bit odd, given that the bore and mechanism of Boehm's flute has never changed significantly, so the only thing that has are the holes. And as you can see, changed utterly - certainly beyond my wildest imaginings.
With all the modern interest in period flutes by Louis Lot, etc, nobody seems to have bothered to investigate what the big differences are between various period flutes. More fashionable I guess to ascribe differences to matters of incredibly little effect such as use of unusual alloys, seamed tubing, drawn vs. soldered tone holes, minute changes to the head taper and so on! Hole changes such as I have documented above blow all of these out of the water! Hopefully, this will lead to a frenzy of measurement and comparison activity. It should!
So what did we learn?
More to come
There's a lot more to be said on this topic. Much more data is needed to illustrate change over the 150 years of the flute's development. And there is much more of value to be extracted from the data presented above. Some topics to be addressed:
Now if you have access to an historical flute and the tools and skill with which to pull it apart and measure it, I'd be delighted to receive more hole diameter and placement info. What I'd like is the length from the centre of the embouchure to the middle of each hole, and the diameter of these holes. Even if you measure another flute and it turns out to have the same hole sizes as one listed, I'd like to know. Now that we've started to uncover the real story as to why modern flutes of different periods are so different, let's keep the ball rolling.
To the Dayton C Miller collection for provision of data on flutes in their collection. To Robert Bigio for date information on the Rudall Carte flutes, and to Trevor Wye and Larry Krantz for publishing the data on modern practice. If I've forgotten to acknowledge your input, it's through oversight, not intent, so don't remain silent!
Created: Jan 06, Last updated Jun 06.