The Rudall Rose or Carte Models Study



It is far too early to expect to draw final conclusions from the data so far in, but some useful observations can be made already.  

Dating particular flutes more accurately

Combining the address information provided by Langwill's Index of Musical Wind Instrument Makers and the serial numbers and addresses from instruments in the Study, we can interpolate to find the probable date of any flute for which we have a serial number. The blue squares are known address changes or flutes whose dates are known independently.  I am indebted to Robert Bigio for the two final dates.

On the company's productivity

It is also possible to get some idea of the company's rate of production.  We can see it varies greatly:

It is tempting to assume that the introduction of Boehm's 1832 instrument caused the downturn of interest in 8-key flutes from about that time, but we would need to see sales figures for those instruments before such a conclusion could safely be drawn. Indeed, since manufacture by Rudall & Rose of the Boehm instrument didn't start until 1843, perhaps it was more that a downturn in interest in 8-key flutes prompted the company to diversify into other flute types.

It is interesting to note that the Patent Head was patented in 1832 and its manufacture continued until the 60's. Was this an attempt to win back market share?

Again, it is tempting to see the sharp decline at 1850 as a result of the introduction of Boehm's 1847 cylindrical flute. Again an examination of sales figures for the new instrument would be helpful. It was clearly a period of upheaval for the company, with the induction of Carte and several changes of address around that time, in addition to the changes which must have been necessary to take on manufacture of the far more complex instrument.

On Evidence for Models

Lengths of parts and positions of finger holes appear to vary very little between instruments throughout the serial number range.  This makes a certain amount of sense - after all the player's hands define to a certain extent what is possible in terms of reach.  It does suggest that we need to look more to bore profile and fingerhole sizes to differentiate between models.

On Finger hole sizes

The right hand finger holes seem to vary the most, with F# for example varying from around 8 to around 11mm.  Early data seems to suggest strong clusters around 8 and 10mm.  These might be our earliest indications of discrete and popular "models".

Distribution of models

The chart below illustrates how the small, medium and large holed instruments are distributed with time.  Left end starts at Serial No 454, the middle large-hole is No 4260 and the right most flute is 7174.

On body keys

Four distinct key types have been reported so far on the bodies of the instruments surveyed:

  • pewter plugs  (# 655)
  • saltspoon (well distributed)
  • flat disk "keycups" with plain leather "pads" (# 4795)
  • dished cups (#4974)
  • wide cups with vertical sides and a low domed top soldered to forged shafts (# 7174)

On footjoint keys

Footjoint keys for C and C# seem usually to involve pewter plugs closing on metal plates screwed to the wood.   The Eb key varies between the pewter plug variety or being the same as the rest of the keys on the instrument.

On "unusual" keys

Unusual keys (ie. those not part of the "normal" 8-key contingent) reported so far include:

  • a "Brille" (additional upper C pad operated by ring keys on the top two holes) (# 2625)
  • a double Bb key, operated by both LH thumb and RH forefinger (# 4795, # 5088)
  • nine keys, extending the bottom note to B (# 3312)
  • ten keys, extending the bottom note to B and Bb (# 3060)
  • a C#/D trill key (#6239)
  • a roller on short F (# 605)
  • a high E key (# 5356)

On Foot Length

As mentioned above, the lengths of the body sections and placement of finger holes varies only slightly.  A fascinating pattern emerges with the length of the foot joint however.  The earliest instruments were nearly 150mm, gradually making their way down to around 140mm by about No 4000 (c1840).  They settled in this area until about No 7000 (c1885), whereupon they shortened suddenly to just over 130mm.

Two things are probably happening here.  Certainly pitch of the flutes is increasing.  Reduction of "flat foot syndrome" is probably the other contributor.

On Body Scaling and Embouchure to Hole 1 length

During the period covered, 1820 to the end of the century, pitch rose dramatically, then dropped as British high pitch was abandoned (apart from by military bands) circa 1895.  We saw that effect in the rapidly reducing lengths of the foot, above.  We might also expect to see this as a significant shift in lengths of other parts of the flute, such as the distance between holes 1 and 6, and the distance between embouchure and the first hole.  The data does not yield what we might expect.

In the graph below, the distance between the first and sixth hole is given in navy.  While we might expect it to drop over the period and then recover somewhat at the end, it generally increases slightly.  We'll come back to that.

We would expect the distance from embouchure to first hole to drop very markedly in order to accommodate the higher pitch towards the end of the period, but the data shows only a gentle drift of about 5mm downwards.  On the old rule of thumb, 1mm per Hertz, that suggests only a 5Hz variation in pitch, while our other studies suggest a greater than 20 Hz variation occurred.

This can probably be explained easily - the very long tuning slides provided were already capable of covering the full range of pitches likely to be encountered, so there was no need to reduce the distance between embouchure and hole 1 when pitch rose - just push the slide in a little more.  As we'll see later, the size of the embouchure hole increases too, which sharpens the flute considerably.

On hole size balance

But it's not so easy to explain away the slight increase in distance from hole 1 to hole 6 at a time when the pitch is rising and there should be a marked decrease.  What else is going on?

Could Rudalls have used fingerhole sizing rather than fingerhole spacing to allow for increasing pitch?  If so, we'd expect holes at the bottom of the flute, i.e. holes 5 and 6, F# and E to increase, or holes at the top, holes 1 and 2, C# and B, to decrease in diameter, or a bit of both.  It turns out that all holes increase in diameter over the period, although it gets difficult to interpret as there are a mix of large and small hole flutes in the earlier days.  To get around that problem, it's safer to look at the ratios of the hole sizes, rather than their absolute values.  When we do that, in the graph below, we see that the bottom-of-tube holes increase considerably more than the top-of-tube holes. 

In blues, we see the ratio of the diameters of hole 6 and hole 1.  In reds, we see the ratio of holes 5 and 2.  The rise over the period approaches 10%, more than enough to overcome the 2% or so increase in length and make a longer-bodied flute a sharper flute!

Impact of the shorter foot

But there's also something else that is going to sharpen the low notes of the flute as the century progresses - the shortening feet we saw up at "On foot length".  The E note in particular (being produced from a very small hole) is very sensitive to how far away its nearest venting support is.  That venting support will be Eb (if you open that when playing E) or the D hole (where the C# pad is) if you don't.  Both of these move much closer to the E as the century proceeds.  The combination of these various effects probably explain why the C#-E length had to be increased to prevent E going too sharp.

On embouchure size

We can see that the longitudinal length increased from around 11.5mm at the start to between 12 and 12.5 or so after about serial number 4411.  The short axis seems to have generally remained between 10 and 11mm.

On Best Pitch

[Update:  This section needs review in the light of other studies.]

By determining the best pitch of readily available flutes and scaling lengths, it is possible to estimate the design pitch of flutes in the study. This is the pitch for best intonation, not the highest pitch available by pushing the tuning slide fully in. It suggests that, in common with other 2nd generation instruments, they were intended to work best at low and medium pitches (430 and 445) rather than high pitch (450-455) as is generally believed. Note the last two flutes are radically different and follow the tuning of third generation flutes, which seems appropriate given their very late dates.

This picture is confused however by the issue of flat footedness, and more work is needed to work out what is really going on.

Coping with change

Now an interesting puzzle presents itself.  If the finger holes don't move, yet the pitch of the instrument changes with the time, how is that reconciled?  As we've seen the foot gets shorter, but that won't have much impact on the body notes.  The graph below gives us some clues ...

Firstly, focus on the flutes with an F# hole of about 8mm.  Note that the B holes are about the same size.  Note also that these flutes seem more popular in the early part of the century.

Now turn to the large hole flutes and we note that the B hole is generally a mm or so smaller than the F# hole.  But look at 7120 and we see it is much smaller.  What will this do?

Increasing holes at the top of the tube will sharpen the upper tube notes in relationship to the notes lower down the tube.  To find the best pitch again, the slide will have to be compressed.  As it has more effect on the upper tube notes, an extension will be found where the upper and lower tube notes are once more in tune.  That will be at a higher pitch than before.  The flute has been sharpened without recourse to shortening the body.  Some adjustment to the foot length will be required too.

On Bore profile

While the bore dimensions have not been made a formal part of this phase, some bore information is available.  As observed previously, bore information may be needed to confirm the models.


In the graphs above, the vertical dimension (diameter) has been wildly exaggerated.  The vertical scale has been omitted to protect the interests of the museums from which some of the data is taken.  Of particular interest are the "back reamings" at the two junctions between the upper and lower body and the footjoint.  Since these had to be consciously reproduced on each instrument, we can assume the company regarded them as significant.  What result they produce or were intended to produce is not clear but may be determined by experiment.

[Update:  Since the work done on the effects of thread wrapping (see elsewhere on this web site), it is now safe to assume that the bumps at the tenons of the flute bores shown are due largely if not entirely to compression of the bore by the combined action of humidity and thread compression.]

Note that the curve for # 7174 has been left incomplete around the tenons of the upper body because of  recent repairs to those areas.

Note the great similarity between the bores of # 5047 and # 5501.  This is puzzling as the section lengths and fingerhole positions are nearly identical, but their fingerhole sizes are very different.  There remains a mystery to clear up here.

Note also the lengths of the bores which relate directly to the estimated design pitch. The Norman flute is longest, with the 5000 series instruments about the same, and 7120 the shortest.

On Inheritance and Cross-fertilisation

An interesting question is to what extent the company relied on the work of previous makers, and to what extent their instruments were copied by other makers, particularly those who had been employees.  Questions and answers appear together below:

Influence of previous work of Willis (TBD)

Influence of previous work of Rose (TBD)

Influence on Ingram

Ingram was a former employee of Rudall & Rose.  Examination of a flute made by Ingram indicated no sign of influence by the company's designs.

Influence on Wylde (TBD)

Influence on other makers of the time

Comparisons of bore data indicate no points of similarity with Metzler, Pratten's Perfected or B&S flutes.

  Back to the Rudall & Rose Models Study