of merchantmen and seek the safety
that their speed could offer. H.M.S. _Calgarian_, torpedoed and sinking,
had yet thought for the convoy she escorted. Her last official signal
directed the ships to turn away from the danger.

[Illustration: BUILDING A STANDARD SHIP]

The convoy system did not spring fully served and equipped from the
earlier and less exacting control. Tentative measures had to be devised
and approved, a large staff to be recruited and trained. The clerical
work of administration was not confined to the home ports; similar
adjustment and preparation had to be conducted in friendly ports abroad.
As naval services were adapted to the new control, the system was
extended. The comparatively simple procedure of sending destroyer
escorts to meet homeward-bound convoys became involved with the timing
and dispatch of a mercantile fleet sailing from a home port. The escorts
were ordered out on a time-table that admitted of little derangement.
Sailing from a British port with a convoy of outward-bound vessels, the
destroyers accompanied that fleet to a point in the Atlantic. There the
convoy was dispersed, and the destroyers swung off to rendezvous with a
similar convoy of inward-bound vessels. While the outgoing merchantmen
were allowed to proceed independently after passing through the most
dangerous area, the homeward-bound vessels were grouped to sail in
company from their port abroad. An ocean escort was provided--usually a
cruiser of the older class--and there was opportunity in the longer
voyage for the senior officer to drill the convoy to some unity and
precision in manoeuvre.

The commander of the ocean escort had no easy task in keeping his
charges together. The age-old difficulty of grouping the ships in the
order of their sailing (now steaming) powers has not diminished since
Lord Cochrane, in command of H.M.S. _Speedy_, complained of the
'fourteen sail of merchantmen' he convoyed from Cagliari to Leghorn. In
the first enthusiasm of a new routine, masters were over-sanguine in
estimation of the speed of their ships. The average of former passages
offered a misleading guide. While it was possible to average ten and a
half knots on a voyage from Cardiff to the Plate, proceeding at a speed
that varied with the weather (and the coal), station could not easily be
kept in a ten-knot convoy when--at the cleaning of the fires--the steam
went 'back.' Swinging to the other extreme (after experience of the