The numbers of American square-rigged ships

and the relationship between rig and size

(5th version 28 May 2004)

Introduction

Obviously, these are major topics capable of absorbing a lifetime’s input. However, this page provides a contribution through making available some statistical material not widely covered in general publications, through its partial analysis and through discussion of research sources and strategies and implications. Familiarity with sailing ship rig types is assumed. A parallel coverage of schooners is under construction.

Specifically, this page presents:

● a graph of construction from 1813 to the early 1900’s

● a summary of construction in the leading construction states 1833-1860

● analysis of the relationship between size and rig from 1870 onward

● general discussion of sources and further research

● discussion of the implications for charts illustrating sailing ship rigs

Consent is granted for reproduction of the graph and other content by teaching institutions, libraries and museums for educational and display purposes with acknowledgement to R.J. Lowe, Wellington, N.Z. provided that my own right to republish is not thereby compromised. Please consult regarding any other use besides strictly personal research.

Construction of US square-riggers from 1813

The statistics currently available to me provide only limited details by rig type but are sufficient to indicate the construction cycles and to distinguish the numbers of two and three-masted square-riggers built in each year. This basic pattern can be elaborated with some other fairly readily available sources and analysed further at length using more detailed data sources currently existing and under development. The statistics used were published originally in the annual Report on Commerce and Navigation and from 1884 in the Report of the Commissioner of Navigation.

The graph below shows the numbers of square-riggers constructed annually from 1813 onward. It clearly indicates the predominance of two-masters in the earlier part of the 19^th century changing to a predominance of three-masters at the beginning of the clipper period. This shift undoubtedly reflects the overall increase in average size of ships during the century. Equally undoubtedly, there was much variation in the choice of rig particularly in earlier years when most ships were below 500 tons when the relationship between size and rig may well be much more complex and variable. The source information does not differentiate between three and four-masted rigs but alternative sources show that the four-masted square-riggers were few in number and significantly influence only the last two barely discernible “blips” on the graph. The initial “spike” for 1813 may well reflect only the fact that earlier figures are unavailable (a casualty of the destruction of Washington during the War of 1812?).

The total numbers constructed in the period were 7,283 ships, barks and barkentines and 5,351 brigs and brigantines, a grand total of 12,634 ships to 1910. To these may be added some 35 four and five-masted barkentines (inclusive of auxiliaries) built in the period 1917-1920. 249 three-masters and 360 two-masters were built in the three years 1798-1800 inclusive but figures for 1801-1812 and before 1798 are missing or were never compiled. Comprehensive tonnage statistics for individual rigs were not published as a time series but can be extracted from the annual publications at least from 1868 onward or else constructed from individual ship records or approximated from the available numbers and average sizes derived from region and period-specific samples. It can be inferred that an equivalent graph showing tonnage constructed would show greater divergence between three and two-masted vessels particularly in the middle and right hand side of the graph.

Of these totals, only 1,020 ships, barks and barkentines and 230 brigs and brigantines were built after 1867, the period covered by the List of Merchant Vessels of the United States. For all that the Downeasters of the 1870’s and early 1880’s constitute a highpoint of design and efficiency they appear as an “after-thought to the main event”, numerically speaking.

The graph shows that U.S. construction of square-riggers dwindled to almost nothing, comparatively speaking, during the 1880’s. Generally, only small numbers of specialised categories were built after that, mostly with quite specific trades in mind. In addition to the obvious competition in the last third of the 19^th century from steam ships in both the coastal and foreign trades, from foreign sailing ships in the foreign trades and of rail transport with the coastal and trans-continental trades, increasingly large schooners provided strong competition for the square-riggers in the shorter trades, and ultimately even on some of the longer routes. From the 1850’s onwards three-masted schooners became comparatively numerous and increased in average size and were augmented by substantial numbers of larger four and five-masted schooners from the 1880’s onward. The even larger six-masted schooners and the solitary seven-master built during the early 1900’s are rightly celebrated as the ultimate phase of their line of development but were numerically insignificant. More significant in terms of competition with square-riggers and schooners alike was the rise from the 1880’s onward of purpose-built and converted schooner-barges and large unrigged barges – categories of vessels generally disregarded by all except those who profited from them but undoubtedly an important form of readjustment of shipping types to economic and technological change*. I plan to add fuller documentation of these categories in a later edition but much work requires to be done on schooner-barges, barges and three-masted schooners in order to do so properly.

* The only major study of which I am aware is Schooners and Schooner Barges, Paul C. Morris, Lower Cape Publishing, Orleans, Massachusetts,

1984, 149 pp.

Many of the ships had quite lengthy careers so that a graph of the numbers still in service each year would have a quite different shape. Although very few American square-riggers were built after the mid 1880’s, the Report of the Commissioner of Navigation for 1901 lists 337 square-riggers registered in the US (inclusive of a few foreign-built ships and exclusive of any former square-riggers re-rigged or reduced to barges). The 1903 Report lists 320 surviving, the 1909 Report lists 171 and the1912 Report 129.

Construction by state

The official statistics also provide a breakdown by state of construction from 1833 onwards. I do not have access to any of the detailed primary records earlier than 1868 but a summary for the period 1833-1860 is included as Appendix XV of Albion and Pope’s history of the port of New York*.

* The Rise of New York Port, Robert Greenhalgh Albion and Jennie Barnes Pope, Northeastern University Press and South St Seaport Museum New York, 1984 (originally published 1939 and 1967).

Albion and Pope’s table indicates that

● of 4,407 ships, barks and barkentines built from 1833-1860*

2,061 were built in Maine

1,126 were built in Massachusetts

446 were built in New York City

774 were built elsewhere

● of 2,556 brigs and brigantines built from 1833-1860*

1,575 were built in Maine

298 were built in Massachusetts

116 were built in New York City

567 were built elsewhere

* To be precise, these numbers cover from January 1833 to June 1860 exclusive of the first nine months of 1835 because of shifts in the basis of computation from calendar years to years ended September, to years ended June. Figures for the first nine months of 1835 or a value for the 12 months ended September 1835 may possibly be in the relevant volume of the Report on Commerce and Navigation but I am unable to check (please contact me if you can).

The predominance of Maine is evident in both categories, accounting for 47 percent of the three-masters and 62 percent of the two-masters.

The “Thober list” covering the period from 1870 onward (refer following section), indicates that 533 of his total of 775 three and four-masted square-riggers built from 1870-1902, were built in Maine (69 percent).

Other than for 1868, I do not have access to volumes of the Report on Commerce and Navigation for the period 1861 to 1869 which contain the figures for the years in the nine-year gap between Albion and Pope’s appendix and Thober’s list. I would appreciate assistance with obtaining copies of the relevant pages to enable me to complete the coverage.

The relationship between rig and size

It is a reasonable deduction that the shift from two to three-masted rigs shown in the graph reflects a general increase in average size of ship (though not only that). Even casual inspection of sailing ship records indicates that, at a certain level of generalisation, there is a relationship between ship size, the number of masts and the degree of square-rig. There is a general pattern that within a rig type in a particular period, five-masters tend to be larger than four-masters which tend to be larger than three-masters which tend to be larger than two-masters. Similarly, ships tend to be larger than barks which tend to be larger than barkentines and brigs tend to be larger than brigantines. Three-masted ships tend to be larger than a four-masted barkentine, all other things being equal. However, there is much variation around the averages and size overlap between the classes of ship, which does not of course preclude generalisation or render averages an irrelevant concept.

People who are less familiar with statistics or suspicious of them will find comment and anecdote through the maritime literature that does reflect instinctive recognition of points beyond which a rig was unsuitable for a size of vessel and the occasional very specific statement that the so-and-so was really too big for a three-master, or whatever (you’ll find a number of such quotes in Lubbock for a start).

There is considerable scope for choice in the selection of rig for a particular ship but past certain points practical and economic realities must strongly favour some choices over others in ways that change through time with technology and economics, within limits dictated by physics. To take an extreme hypothetical example, consider the impracticality of scaling up a one-masted rig to a 300-foot hull and handling the resulting vessel. It was through no accident, personal whim or artistic choice that sailing vessels of such a size were built with four to six masts all shorter than the length of the hull.

The following analysis confined to the period after 1869 shows quite strong size differentiation between rig classes but it is quite possible that differentiation was less strong in earlier periods when construction was concentrated in the smaller ship size ranges. Detailed analysis of earlier periods, particularly before the mid 1840’s, may well show greater variability and local differences and additional factors in play that would be fascinating to unravel. In the case of numerically small classes in a limited period, very specific factors may apply. In some cases, many of a class were built by only one or two builders for a few owners and specific trades within a specific investment climate. In such cases the results may well represent only a subset of the range of vessels for which that rig may be technically suitable. Several of the five masted barkentines of the 1917-1920 period reflect the original demand for the class of steamer hull as which they were originally designed although the builders had considerable choice in whether they completed them as schooners or barkentines, and with four, five or six masts. Technically, they could have been finished as four-masted barks. In the economic conditions of the time they could not.

The analysis in the following table is based mainly on the “Thober list” compiled by Frank W. Thober published in Log Chips, Volume II pp 129-32, 142-44 and Volume III pp 11-12, 22-24, 48, 59-60, 71-72, 130-32, 143-44. Gross tonnages are given for the average, smallest and the largest of each rig. Thober’s list covers only square-riggers of three or more masts. I have added comparative figures for brigs and brigantines recorded in the List of Merchant Vessels of the United States for 1893 although these represent only a small proportion of those built from 1870 onward so further analysis of these rigs is required.

The relationship between size and rig, U.S. square-riggers constructed 1870-1920

Number Average Smallest Largest

of ships tons tons tons

1917-1920

5-masted barkentines built 1917-1920 26 2,163 1,600 2,462

4-masted barkentines built 1917-1919 9 1,357 1,216 1,519

1870-1902

4-masted barks built 1874-1902 13 3,122 2,516 3,539

4-masted jackass bark built 1892 1 1,469

4-masted barkentines built 1890-1902 29 977 499 1,554

3-masted full-rigged ships built 1870-1893 302 1,617 349 3,185

3-masted barks built 1870-1899 261 753 177 1,669

3-masted barkentines built 1870-1900 169 561 223 890

Brigs built after 1869* 36 420 128 623

Brigs built before 1870* 39 317 107 507

Brigantines built after 1869* 13 292 152 496

* those surviving in 1893 plus the brigantine Viola built in 1910. These are not necessarily the original rig.

Altogether, 230 brigs and brigantines were built after 1869.

In the following table the average tonnages are repeated in a format that makes it easier to relate both the number of masts and the degree of square-rig simultaneously to size. Reading up and down the columns which show the degree of square-rig and across the rows to compare the number of masts illustrates the generalisations made in the first paragraph of this section. In reading the table, a brig can also be envisaged as a “two-masted full-rigged ship”, a three-masted bark as “the sail plan of a brig stretched over an additional mast” and a four-masted bark as “the sail plan of a three-masted ship stretched over an additional mast” and the generalisations will hold. There were no American five-masted barks or full-rigged ships or four-masted full-rigged ships. For comparative purposes, the average tonnage of the six five-masted barks and the tonnage of the only five-masted full-rigged ship* have been added to the table where they can be seen to be consistent with the generalisations made. No figures have been added for four-masted full-rigged ships as further analysis is required of the particularly complex set of factors involved.**

* The recently constructed cruise ship of this rig is disregarded.

** Only a small number of four-masted full-rigged ships were built, of smaller average tonnage than the totality of four-masted barks which appears

to conflict with the generalisation of the more square-rig the larger the ship (for the same number of masts). However, provisional analysis

suggests that much of the discrepancy is accounted for when the iron vessels of the 1870’s and 1880’s are differentiated from the steel vessels of

the 1890’s and later, and allowance is made for other differences between the requirements of the periods. The statistical effect of adjustment for

blocks of sister- ships among the small number of four-masted full-rigged ships may also be significant. I will add a fuller analysis of four-masted

square-riggers at a later date.

Average tonnage of American square-riggers 1870-1920 with additional comparisons in [ ]

RIG	Masts	Masts	Masts	Masts
	2	3	4	5
Full-rigged ship		1,617		[5,081]
Bark		753	3,122	[4,480]
Jackass Bark			1,469
Barkentine 1917-1920			1,357	2,163
Barkentine 1870-1902		561	960
Brig	420
Brigantine	292

Extending the analysis to other periods and places may be visualised as a series of tables in the same format* representing earlier periods to the left in the same dimension and multiple layers in the third dimension representing localities, regions or nations. Such comparisons would provide a starting point for exploring the reasons for the differences. Deeper statistical analysis would involve comparison with other periods and other places, analysing the variation around the averages and seeking for pattern in the comparisons and the variation and seeking explanations for any such pattern.

* though where only two and three-masted rigs exist a single column will suffice.

A similar analysis has been conducted for Quebec ships over the period 1765-1893 which shows similar relativities. The Quebec analysis for the sub-period from 1867 onward which broadly matches the period of the above tables gives the average for 3-masted ships as 1,252 tons, for 3-masted barks as 694 tons, for 3-masted barkentines as 359 tons, for brigs as 341 tons and for brigantines 214 tons. (There were no four-masters.) The averages for earlier periods are lower but maintain the same general relativity between rigs. The average size of each rig increases through time. For example, the average size of Quebec 3-masted ships was only 403 tons in the period 1805-1835. Fuller details are available in the source, The Charley-man. A History of Wooden Shipbuilding at Quebec 1763-1893, Eileen Reid Marcil, Quarry Press, Kingston, Ontario, 1995, pp 318-319. A similar analysis for Atlantic Canada (not including Quebec) by rig and decade of construction also shows the same general relativities and marked increases in size for each rig except brigs (Sagar, Eric. W. and Gerald E. Panting, 1990, Maritime Capital. The Shipping Industry in Atlantic Canada, 1820-1924, McGill-Queen’s University Press, Table 3.1, page 55).

Note that Thober’s list relates only to ships as originally constructed and disregards post-construction conversions to a rig. This is the appropriate basis for analysis as additional factors apply in the case of conversions and they must certainly be assessed against conditions at the time of conversion rather than original construction and therefore treated separately. The only six-masted barkentines (the City of Sydney and the E. R. Sterling ex Everett G. Griggs ex Lord Wolseley) were conversions from a steamer and a 4-masted ship of 3,017 and 2,577 tons, respectively. All the four-masted jackass barks other than the Olympic (which was built as such and therefore included in Thober’s record) were conversions (only one was American and outside Thober’s period).

Rigs are as generally described except that Thober has, properly, differentiated the four-masted barks which appear in some standard records (including the LMVUS) simply as “ships” along with three-masted ships as the main tabulations of the LMVUS do not directly indicate the number of masts. I have revised Thober’s summary to exclude the four-masted jackass bark Olympic from the four-masted barkentines with which he included her. Particularly in earlier years, barkentines may appear in American records as barks (they often appear in British records as schooners or as “three-masted brigantines”). An element of doubt also hangs over how consistently the brig/brigantine distinction was applied particularly in earlier years when there was much variation in what the term brigantine denoted (refer Chapter IX of E. P. Morris, The Fore-and-Aft Rig in America, Macdonald and Jane’s, London, 1974; previously published 1927, New Haven, and 1970).

Thober’s list as a source

Thober’s list constitutes a major research resource but it appears that it is not absolutely comprehensive (although it is unlikely that any omissions would affect the calculated averages very significantly). Matching against the official construction totals for square-riggers indicates complete coverage from 1894 onwards but suggests that before that date Thober may be missing several dozen ships. (Exact year by year comparison is not possible because Thober’s list and most other records indicate the calendar year of construction while the official statistics in the relevant period are compiled for years ended June consisting of halves of two calendar years.)

There are at least three reasons for omissions:

● Outright omission – easy enough when working with large collections of records in various formats and sources particularly in the case of ships lost or sold foreign early in their careers which leave few traces in their original country. Such vessels might never appear in any volume of the LMVUS and would be traceable only through primary registration, building or shipping departure records, newspaper accounts or personal records.

● The exclusion of vessels built on the Great Lakes which appear to be outside Thober’s scope but which are within the scope of the construction statistics. The detailed construction records for the years ended June 1885, 1887, 1888, 1891 and 1892 (the only ones available to me at present) do not indicate any square-riggers built on the Great Lakes in those years. However, nine square-riggers were built on the Lakes in the year ended June1868 (the only other detailed records directly available to me) so Lakes’ construction may well extend into Thober’s period.

● I have not eliminated the possibility that foreign-built ships added to the US register as reconstructions from wrecks may have been included in the statistics as US construction in the year of rebuilding, which would have a certain administrative logic to it though I think it unlikely. If so, they would have to show up in the detailed construction tables.

I think it is most likely that Thober’s apparent undercount results partly from the exclusion of vessels built on the Lakes and partly from outright omissions. If anyone has access to a set of the annual Report on Commerce and Navigation up to 1883* it would be helpful to have the relevant statistical information to establish the degree to which the exclusion of Lakes’ vessels is a factor (if necessary contact me for a sample page from the 1868 volume indicating what to look for). (It would also be very useful to extract a long statistical series of construction on the Lakes to establish the actual numbers given the widespread neglect of Lakes’ maritime history by those who think only in terms of sea-going vessels.)

* thereafter the tables appear in the Report of the Commissioner of Navigation.

As any outright omissions come to light I will add an “Additions to Thober” page to my site. Readers are invited to submit any known cases for inclusion with acknowledgment. One possible example is the reference to a 540 ton Turkish bark Zigomala in Lloyd’s Register of 1912-13 as built in Philadelphia in 1893 which matches no record in Thober’s list for that year (the only other Philadelphias I can find are in inland Palestine, Tennessee and inland Brazil so unless typographic errors are involved the Zigomala appears to be one of the omissions though not necessarily under its original name). The detailed construction statistics in the Report of the Commissioner of Navigation for 1893 (possibly labelled “1893-94”) would confirm whether any square-riggers were built in Philadelphia in 1893.

Extending the record and the analysis

The sources covered so far provide an overview of the construction of square-riggers in the United States in the 19^th century and a detailed analysis of the connection between rig and size but only during the last phases of construction.

Full analysis of points raised requires dealing with earlier periods in greater depth requiring suitable data in formats in which it can be efficiently analysed.

The computerisation of the List of Merchant Vessels of the United States has the potential to revolutionise such analysis when and if converted into database format (as distinct from its present pdf format which permits individual searching very satisfactorily but is inadequate for efficient statistical analysis). A rough analysis suggests that its record of ships registered in the 1867-1885 period may contain almost three-times as many three-masted square-riggers and four times as many two-masted square-riggers as were built in that period so that it is potentially a means for analysing some 2,500 square-riggers built before 1868, presumably built mostly in the two preceding decades and perhaps constituting around half those built in those two decades. (For those built in the 1867-1885 period it provides an alternative to Thober and a means of finding Thober’s omissions.)

Probably the only other single source that possibly covers so many pre-1868 square-riggers in a single source is the “Holdcamper list” of vessels of all types registered in New York from 1789-1867 (refer to North American sources). This has not been computerised to my knowledge but is in a format that lends itself well to statistical analysis for which a generous sample might well suffice. The majority of ships included in it were built outside New York so it is evidently a major record of construction generally. Albion and Pope’s analysis indicates that New York accounted for a fifth to a quarter of total registered shipping tonnage from the 1820’s through to 1860.*

* The Rise of New York Port, Robert Greenhalgh Albion and Jennie Barnes Pope, Northeastern University Press and South St Seaport Museum New York, 1984 (originally published 1939 and 1967), Appendices I, XI, XII.

However, one could not rely on, say, Maine-built ships registered in New York necessarily being precisely representative of all ships built in Maine. Indeed, it is unlikely. It is therefore also desirable to analyse other listings specific to particular construction ports as a cross-check on any more generalised analysis. Such other listings are of course also valuable for what they themselves contribute to general understanding of variation between localities and through time and of the factors responsible.

Given the predominance of Maine in square-rigger construction the major sources dealing with Maine shipbuilding warrant early attention but smaller ports in other states also need analysis to provide comparisons. Variation between ports within Maine also warrants attention.

I anticipate that fuller analysis will support the interpretation of a link between increasing average size and the shift from two to three-masted rigs but also that the average size of rigs will increase through time and that there will be interesting regional and local variations from/around any general pattern/trend. As I obtain sources and as time permits I will extend my analysis to multiple locations in the earlier periods. Anyone else working on overlapping projects is invited to contact me concerning possible collaboration.

Analysis of the rig-size relationship in the period before 1868 will need to take account of the differences in tonnage measurement method.

Implications for illustrating sailing ship rigs

The demonstrated relationship between rig and ship size has implications for compilers of charts illustrating sailing ship rigs as diagrams in books and as stand-alone posters.

Such charts should reflect that some rigs were suitable for large ships and others for small ships. Otherwise, they miss out on portraying an important part of maritime understanding and education and may actively mislead.

For full discussion click here

Queries and comments to j_lowe@ihug.co.nz To return to main Maritime menu click here