Measuring the
transition from Sail to Steam
(revised 28 January 2008)
Obviously, this topic could fill several hundred pages. The following graphs simply restructure the statistics used in the comparison between Lloyd’s Register and Bureau Veritas’ Répertoire Générale at Comparison in order to extract what they have to say about the transition. A wealth of national shipping statistics available in many sources can be used to address the question more fully but substantial analysis is required to standardise the data as national statistical practices and tonnage thresholds for registration vary enormously. The shipping registers understate the totals but are the best single starting point. An article cited in the last section of this page examines the topic more fully.
Consent is granted for reproduction of the graphs with acknowledgement.
Number of ships
In the above graph The Répertoire Générale figures for the numbers of sailing ships of 50 tons and above (the upper line with the declining trend) are shown as they are larger than the Lloyd’s series. They do not include every small sailing ship in the world even so, but they do represent an absolute lower limit of the number.
The rising trend lines indicate the increasing number of steam ships (inclusive of the few motor vessels built in the period). The Répertoire Générale and Lloyd’s figures for the number of steam ships are both shown as the Répertoire Générale series has the longer coverage in time, while from the mid 1880’s the Lloyd’s figure progressively overtakes it.
The Sail/Steam comparison is distorted by the inclusion of ships of 50-100 tons in the Sail series but neither of the Steam series. However, the most casual comparison between the numbers of ships recorded by the two agencies for national fleets with the numbers registered in those countries is sufficient to indicate that large numbers of small ships, both sail and steam, are not covered by any of the series shown in the graph. It is probable that incorporating all other ships down to five or 15 tons (the cut-off points for registration of the two largest national fleets) would push the true comparison of numbers of ships further in favour of sail, raising the line for sail up the graph and extending the cross-over point beyond World War I by a few years at least.
Tonnage
Tonnage gives a better indication of carrying capacity than numbers of ships.
In this graph, the Répertoire Générale figures are again used for Sail. These represent net tons. The Lloyd’s statistics for the net tonnage of steam ships are shown as the dotted line from 1886 onward. For the period 1870-1886, the Répertoire Générale figures for the gross tonnage of steamships have been adjusted to estimates of net tonnage (the dashed line) using the ratio of net to gross in the summary statistics table of Lloyd’s Universal Register for 1886. (The discrepancy between the Répertoire Générale and Lloyd’s coverage of Steam in the 1880’s is negligible at the scale of the graph.) With full access to the original registers the actual net steam tonnages for 1870-1886 can be substituted but are unlikely to alter the appearance of the graph.
Unqualified, the
above comparison puts the world cross-over point from Sail to Stream in the
early 1890’s. However, as in the case of ship numbers, the comparison is
distorted by counting sailing ships of 50-100 tons but steam from 100 and above
although the discrepancy is not as great as may appear because for steamships,
100 gross represents about 67 net. What’s more important though is that for
both Sail and Steam the tonnages would be increased if the criterion for
inclusion was five or 15 tons (the criteria for
It is likely that including smaller ships would push the comparison slightly in favour of sail and the cross-over point somewhat later in the 1890s and possibly even into the early 20th century – the “mosquito fleets” of the world were tiny, and many of their members had only the carrying capacity of a few modern shipping containers, but they were very numerous and in the aggregate they would have made a measurable, though minor addition to world tonnage.
The measurements
of tonnage for both Sail and Steam are affected by international changes in
tonnage measurement which differed significantly between nations and through
time, particularly in the earlier part of the period covered by the graph. The
basis of the tonnage measurement system introduced by
The mixture of tonnage figures calculated under different systems both nationally and through time could well affect the comparability of the statistics significantly, certainly into the 1880s. This would be unlikely to have much effect on the trend line for steam as most ships covered in the graph would have been measured under the new rules, or on the comparison between Sail and Steam in the later part of the graph. However, the tonnage figures for Sail in the early part of the period could be significantly affected relative to those for Sail itself in the latter part. If that is so, the recorded figures could give a significantly exaggerated impression of the magnitude and rate of decline of Sail.
Further analysis and comparisons of government-produced shipping statistics is necessary to obtain a fuller picture but it is probably safe to conclude that the cross-over point from Sail to Steam, as measured by net tonnage, occurred during the 1890s. An irony is that by the 1890s a major trade for sailing ships was the international carriage of coal – which played a role in the world economy much like petroleum today – of which a major part was for the bunkers of the steam ships that displaced them. They were already part of a steamship world.
The economic
transition
The above comparisons are valid as far as they go. However, the most significant consideration in understanding the transition from Sail to Steam is the question of how much cargo the two fleets (and major sub-categories within them) could shift in a year, and at what cost and return on capital. That is a much bigger story but it is universally accepted that steam ships were carrying more than half of the world’s cargoes well before they constituted half the world’s net tonnage. In terms of economic significance the international cross-over point from Sail to Steam must predate 1890.
It has long been
a common practice in the analysis of historical shipping statistics, dating
back to the period in question, to adjust steam tonnage upward by from 300 to
400 percent when comparing Sail and Steam tonnages, in order to take account of
the ability of steam ships to carry more cargo in a year relative to hold
capacity. Such adjustments are discussed extensively in Maritime Transport and the Integration of the
For the North
Atlantic fleets (which constituted most of the world total), Fischer and
Nordvik put the cross-over point for Belgium before 1880, for Britain, France,
Germany and Spain during the 1880’s, for Italy, Sweden, Russia, Holland, and
Denmark during the 1890’s, and for the United States, Norway, and Greece not
till after 1900.
To return to main Maritime menu click here