confidently taken for granted. The
plate-makers thicken their armour while the gun-makers enlarge the
size and increase the penetrative power of their weapons, until the
weight that has to be carried on a battle-ship renders the attainment
of speed practically impossible.

Meanwhile there is going forward, in the hull of the vessel itself, a
gradual course of evolution which will eventually place the policy of
increasing strength of armour and of guns at a discount. The division
of the air-space of a warship into water-tight compartments will
doubtless prove to be, in actual naval conflict, a more effectual
means of keeping the vessel afloat than the indefinite increase in the
thickness and consequent weight of her armour.

The most advanced naval architects of modern times are bestowing more
and more attention upon this feature, as affording a prospect of
rendering ships unsinkable, whether through accidents or through
injury in warfare. No doubt, for merchant steamers, it will be seen
that development along the lines already laid down in this department
will suffice for all practical purposes. The water-tight bulkheads,
with readily closed or automatically shutting doorways, ensure the
maintenance of buoyancy in case of any ordinary accident from
collision or grounding, while the duplication of engines, shafts and
propellers--without which no steamship of the middle twentieth century
will be passed by marine surveyors as fit for carrying passengers on
long ocean voyages--will make provision against all excepting the most
extremely improbable mishaps to the machinery.

If the numerical estimate of the chance of the disablement of a single
engine and its propeller during a certain voyage be stated at one to
a thousand, then the risk of helplessness through the break down of
both systems in a vessel having twin screws and entirely separate
engines will be represented by the proportion of one to a million.
This mode of reckoning, of course, assumes that the two systems could
be made absolutely independent in relation to all possible disasters;
and some deduction must be made on account of the impossibility of
attaining this ideal. Yet it is evident that when every practicable
device has been adopted for rendering a double accident improbable the
chances against such a disaster will not be far from the proportion
stated.

When we come to consider the evolution of the warship as compared with
that of the merchant stea