The doubled treble bells
of the Taylor bellfoundry

In the early 1920s, two of the four major English bellfoundries of the time, Taylor and Gillett & Johnston, discovered that there was a rich market for carillons in North America.  Each installed a 23-bell instrument on this continent in 1922, and for the next several years they competed fiercely in this market.  In the first few years, part of that competition was a "numbers game", a race to see who could produce the largest and/or the heaviest instrument.  Leadership in the race changed hands frequently, but in 1925 G&J emerged triumphant with its very heavy 53-bell carillon for Park Avenue Baptist Church in New York City.  (That was later moved to the new Riverside Church and enlarged to a truly gigantic size and weight.)  Nevertheless, Taylor went on to produce almost a dozen highly-respected concert-class carillons (4 octaves or more) for the USA prior to World War II.

One of the major problems faced by both foundries was how to make small bells which would fit the extended range needed for carillons.  Prior to Taylor's rediscovery (in the 1890s) of the Hemony brothers' principles of tuning, neither foundry had attempted anything of this size.  (Taylor had made a 21-bell chime in 1877, but it was not a success.)  At the bass end, they had made some quite large bells for clock-chimes, as well as Great Paul, still the heaviest bell in the United Kingdom.  But at the treble end, their experience was almost entirely with sets of 10 to 12 bells hung for change-ringing.  And it was obvious from the start that the system of varying profiles used for the trebles of such rings could not be extended to two octaves and beyond.

The foundries addressed this problem by using essentially fixed profiles, i.e., the trebles had approximately the same profiles as basses but were smaller.  Specifically, for two bells an octave apart in pitch, the smaller will be 1/2 the diameter and 1/8 the weight of the larger.  For two octaves apart, the ratios are 1/4 and 1/64; for three octaves, 1/8 and 1/512; for four octaves, 1/16 and 1/4096.  (In practice, this would have made the bells at the top of the fourth octave impossibly small, so some thickening of the profiles of the trebles was necessary to obtain practical sizes.)  With this system, the bells could be tuned to the correct pitches quite accurately.  Unfortunately, this raised a new problem - even with their thickened profiles, the trebles produced a very small sound compared to the basses.  Quite likely this was evident in the very first English-made 4-octave carillon, which was installed in the Rotterdam City Hall in 1920-21.

The Taylor foundry addressed this new problem with an ingenious solution - the topmost octave or so of trebles was given doubled bells.  That is, for each note there were two bells of the same pitch, with their clappers linked together so that they would strike at the same time and (in theory) produce twice as much sound as a single bell.  This scheme was used for the next seven 4-octave carillons produced by Taylor, from 1924 through 1929, and for one more carillon made in 1932, as can be seen in the table below.  (The reason for the production of both doubled and un-doubled trebles in 1929-32 is not yet known.)

In spite of the ingenuity of this approach, it was soon realized that doubled trebles were more trouble than they were worth.  In the first place, they were very difficult to tune to exactly matching pitch.  In the second place, the mechanisms of the day made it very difficult to get a pair of trebles to strike at exactly the same instant.  Taylor soon stopped supplying them, and the carillonneurs fairly soon disconnected the extra bells.  So the sequence of early Taylor carillons in America (plus their four large carillons elsewhere during the same period) looks like this:

 Rotterdam, NL  1920-21  49 bells = 49 notes  (in Netherlands)
 Gloucester        1922  23 bells  (first modern carillon in America)
 Andover           1923  23 bells  (initial installation)
 Birmingham        1924  25 bells
 Gloucester        1924  31 bells  (enlargement from 23 in 1922)
 Morristown        1924  35 bells  (largest carillon in USA for a short time)
+Loughborough, UK  1924  47 notes / 60 bells (in England)
 Andover           1926  37 bells  (enlargement from 23 in 1923)
+Philadelphia      1927  48 notes / 63 bells
+Albany            1927  47 notes / 60 bells
+Cranbrook         1928  48 notes / 61 bells
+Springfield - TM  1928  48 notes / 60 bells
+Lake Wales    Sep.1928  48 notes / 61 bells  (initial installation)
+Sydney, AU        1928  49 notes = 62 bells  (in Australia)
+Indianapolis      1929  49 notes / 63 bells
+Lake Wales        1929  53 notes / 71 bells  (first enlargement)
 Ames              1929  36 notes  (enlargement from 10 in 1899)
-Saint Helens, UK  1929  47 notes = 47 bells  (in England)
-Lincoln           1931  48 notes = 48 bells
-Durham            1932  50 notes = 50 bells  (originally planned as 48)
 Hartford          1932  30 bells
+Richmond          1932  53 notes / 66 bells  (last attempt at doubling)
 Rumson            1934  25 bells
-Ann Arbor         1936  53 notes = 53 bells
-Luray             1937  47 notes = 47 bells
 Springfield - HPC 1937  25 bells
Sites marked "+" were of four octaves or more and had doubled trebles (and are linked to their site data pages, where more details can be found);
sites marked "-" were of comparable size but did not have doubled trebles.

So far as is known, no other bellfounder ever tried this method of getting more sound out of the treble range of a carillon.  All of the others used progressively thicker profiles to compensate for shrinking size, even if that sometimes required using outside hammers because there was insufficient space inside for a conventional clapper.  Not surprisingly, Taylor was much slower to thicken their treble profiles, and has never done so to the extent that some other foundries did.

As mentioned above, all of the duplicate trebles were eventually disconnected, leaving a single working bell for each note throughout the range.  In six of the nine carillons which had them, later work has replaced the doubled trebles (and more) with bells of heavier profile, and the bell frames have been reworked or replaced so that no physical evidence of the original configuration remains.  The original frame (minus duplicate bells) may remain at Philadelphia, while both frame and duplicate bells are believed to remain in place at Loughborough and Springfield - mute evidence of an experiment which failed.

[TowerBells Home Page] [Site data top page] [Credits and Disclaimers] [Feedback]

This page was created 2005/12/01 and last revised on 2009/04/20.

Please send comments or questions to

This page is valid HTML 3.2