men
specially enlisted for that service and kept to that. He would have
some 10 or 20 per cent, only of the fleet built very light and manned
in threes. He does not seem to have contemplated oars three-banked,
and crossbowmen _besides_, as Sanudo does. (See below; and _Muntaner_,
pp. 288, 323, 525, etc.)

In Sanudo we have a glimpse worth noting of the word _soldiers_
advancing towards the modern sense; he expresses a strong preference
for _soldati_ (viz. _paid_ soldiers) over _crusaders_ (viz.
volunteers), p. 74.

[5] _L'Armata Navale_, Roma, 1616, pp. 150-151.

[6] See a work to which I am indebted for a good deal of light and
information, the Engineer Giovanni Casoni's Essay: "_Dei Navigli
Poliremi usati nella Marina dagli Antichi Veneziani_," in
"_Esercitazioni dell' Ateneo Veneto_," vol. ii. p. 338. This great
_Quinquereme_, as it was styled, is stated to have been struck by
a fire-arrow, and blown up, in January 1570.

[7] _Pantera_, p. 22.

[8] _Lazarus Bayfius de Re Navali Veterum_, in _Gronovii Thesaurus_, Ven.
1737, vol. xi. p. 581. This writer also speaks of the Quinquereme
mentioned above (p. 577).

[9] _Marinus Sanutius_, p. 65.

[10] See the woodcuts opposite and at p. 37; also _Pantera_, p. 46
(who is here, however, speaking of the great-oared galleys), and
_Coronelli_, i. 140.

[11] _Casoni_, p. 324. He obtains these particulars from a manuscript work
of the 16th century by Cristoforo Canale.

[12] Signor Casoni (p. 324) expresses his belief that no galley of the
14th century had more than 100 oars. I differ from him with
hesitation, and still more as I find M. Jal agrees in this view. I
will state the grounds on which I came to a different conclusion. (1)
Marino Sanudo assigns 180 rowers for a galley equipped _ai Terzaruoli_
(p. 75). This seemed to imply something near 180 oars, for I do not
find any allusion to reliefs being provided. In the French galleys of
the 18th century there were no reliefs except in this way, that in
long runs without urgency only half the oars were pulled. (See _Mem.
d'un Protestant condamne aux Galeres_, etc., Reimprimes, Paris, 1865,
p. 447.) If four men to a bench were to be employed, then Sanudo seems
to calculate for his smaller galleys 220 men actually rowing (see pp.
75-78). This seems to assume 55 benches, i.e., 28 on one side and 27