Chapter V: V V (1)

W --- cos [phi] and W --- sin [phi] is W --- sec. lb., (3)
g g g

acting along OO´, and so is unaltered.

But the change of the resultant momentum F of the medium as well as of
the body from the vector OF to O´F´ requires an impulse couple,
tending to increase the angle FOO´, of magnitude, in sec. foot-pounds

F·OO´·sin FOO´ = FVt sin ([theta] - [phi]), (4)

equivalent to an incessant couple

N = FV sin ([theta] - [phi])
= (F sin [theta] cos [phi] - F cos [theta] sin [phi])V
= (c2 - c1)(V²/g) sin [phi] cos [phi]
= W´([beta] - [alpha]uv/g). (5)

This N is the couple in foot-pounds changing the momentum of the
medium, the momentum of the body alone remaining the same; the medium
reacts on the body with the same couple N in the opposite direction,
tending when c2-c1 is positive to set the body broadside to the
advance.

An oblate flattened body, like a disk or plate, has c2 - c1 negative,
so that the medium steers the body axially; this may be verified by a
plate dropped in water, and a leaf or disk or rocket-stick or piece of
paper falling in air. A card will show the influence of the couple N
if projected with a spin in its plane, when it will be found to change
its aspect in the air.

An elongated body like a ship has c2-c1 positive, and the couple N
tends to disturb the axial movement and makes it unstable, so that a
steamer requires to be steered by constant attention at the helm.

Consider a submarine boat or airship moving freely with the direction
of the resultant momentum horizontal, and the axis at a slight
inclination [theta]. With no reserve of buoyancy W = W´, and the
couple N, tending to increase [theta], has the effect of diminishing
the metacentric height by h ft. vertical, where

c1 u²
Wh tan[theta] = N = (c2 - c1) -- --- tan [theta], (6)
c2 g

c2 - c1 c1 u² 1 + [alpha] u²
h = ------- -- --- = ([beta] - [alpha] ----------- ---. (7)
W c2 g 1 + [beta] g

51. An elongated shot is made to preserve its axial flight through the air by giving it the spin sufficient for stability, without which it would turn broadside to its advance; a top in the same way is made to stand upright on the point in the position of equilibrium, unstable statically but dynamically stable if the spin is sufficient; and the investigation proceeds in the same way for the two problems (see GYROSCOPE).

The effective angular inertia of the body in the medium is now
required; denote it by C1 about the axis of the figure, and by C2
about a diameter of the mean section. A rotation about the axis of a
figure of revolution does not set the medium in motion, so that C1 is
the moment of inertia of the body about the axis, denoted by Wk1². But
if Wk2² is the moment of inertia of the body about a mean diameter,
and [omega] the angular velocity about it generated by an impulse
couple M, and M´ is the couple required to set the surrounding medium
in motion, supposed of effective radius of gyration k´,

Wk2²[omega] = M - M´, W´k´²[omega] = M´, (1)

Wk2² + W´k´²[omega] = M, (2)

C2 = Wk2² + W´k´² = (W + W´[epsilon])k2², (3)

in which we have put k´² = [epsilon]k², where [epsilon] is a numerical
factor depending on the shape.

If the shot is spinning about its axis with angular velocity p, and is
preceding steadily at a rate [mu] about a line parallel to the
resultant momentum F at an angle [theta], the velocity of the vector
of angular momentum, as in the case of a top, is

C1p[mu] sin [theta] - C2[mu]² sin [theta] cos [theta]; (4)

and equating this to the impressed couple (multiplied by g), that is,
to

c1
gN = (c1 - c2)-- u² tan [theta], (5)
c2

and dividing out sin[theta], which equated to zero would imply perfect
centring, we obtain

c1
C2[mu]² cos [theta] - C1p[mu] + (c2 - c1)-- u² sec [theta] = 0. (6)
c2

The least admissible value of p is that which makes the roots equal of
this quadratic in [mu], and then

C1
[mu] = ½ --p sec [theta], (7)
C2

the roots would be imaginary for a value of p smaller than given by

c1
C1²p² - 4(c2 - c1)-- C2u² = 0, (8)
c2

p² c1 C2
-- = 4 (c2 - c1) -- ---. (9)
u² c2 C1²

_Table of Rifling for Stability of an Elongated Projectile, x Calibres
long, giving [delta] the Angle of Rifling, and n the Pitch of Rifling
in Calibres._

+-----------------------+-----------------+-----------------+-----------------+-----------------+
| | Cast-iron Common| Palliser Shell | Solid Steel | Solid Lead |
| | Shell [f] = 2/3,| [f] = ½, S.G. 8.| Bullet | Bullet [f] = 0, |
| | S.G. 7.2. | | [f] = 0, S.G. 8.| S.G. 10.9. |
+--------+--------------+--------+--------+--------+--------+--------+--------+--------+--------+
| x |[beta]-[alpha]|[delta] | n |[delta] | n |[delta] | n |[delta] | n |
+--------+--------------+--------+--------+--------+--------+--------+--------+--------+--------+
| 1.0 | 0.0000 | 0° 0´|Infinity| 0° 0´|Infinity| 0° 0´|Infinity| 0° 0´|Infinity|
| 2.0 | 0.4942 | 2 49 | 63.87 | 2 32 | 71.08 | 2 29 | 72.21 | 2 08 | 84.29 |
| 2.5 | 0.6056 | 3 46 | 47.91 | 3 23 | 53.32 | 3 19 | 54.17 | 2 51 | 63.24 |
| 3.0 | 0.6819 | 4 41 | 38.45 | 4 13 | 42.79 | 4 09 | 43.47 | 3 38 | 50.74 |
| 3.5 | 0.7370 | 5 35 | 32.13 | 5 02 | 35.75 | 4 58 | 36.33 | 4 15 | 42.40 |
| 4.0 | 0.7782 | 6 30 | 27.60 | 5 51 | 30.72 | 5 45 | 31.21 | 4 56 | 36.43 |
| 4.5 | 0.8100 | 7 24 | 24.20 | 6 40 | 26.93 | 6 32 | 27.36 | 5 37 | 31.94 |
| 5.0 | 0.8351 | 8 16 | 21.56 | 7 28 | 23.98 | 7 21 | 24.36 | 6 18 | 28.44 |
| 6.0 | 0.8721 | 10 05 | 17.67 | 9 04 | 19.67 | 8 56 | 19.98 | 7 40 | 23.33 |
| 10.0 | 0.9395 | 16 57 | 10.31 | 15 19 | 11.47 | 15 05 | 11.65 | 13 00 | 13.60 |
|Infinity| 1.0000 | 90 00 | 0.00 | 90 00 | 0.00 | 90 00 | 0.00 | 90 00 | 0.00 |
+--------+--------------+--------+--------+--------+--------+--------+--------+--------+--------+

If the shot is moving as if fired from a gun of calibre d inches, in
which the rifling makes one turn in a pitch of n calibres or nd
inches, so that the angle [delta] of the rifling is given by

tan [delta] = [pi]d/nd = ½ dp/u, (10)

which is the ratio of the linear velocity of rotation ½dp to u, the
velocity of advance,

c1 C2d²
tan² [delta] = [pi]²/n² = d²p²/4u² = (c2 - c1) -- ----
c2 C1²

W´ / W´ \ / k2 \²
1 + --- [alpha] ( 1 + ---[epsilon] ) ( -- )
W´ W \ W / \ d /
= --- ([beta] - [alpha]) --------------- · -----------------------------. (11)
W W´ / k1 \^4
1 + ---[beta] ( -- )
W \ d /

For a shot in air the ratio W´/W is so small that the square may be
neglected, and formula (11) can be replaced for practical purpose in
artillery by

[pi]² W´ / k2 \² / / k1 \^4
tan²[delta] = ----- = --- ([beta] - [alpha]) ( -- ) / ( -- ), (12)
n² W \ d / / \ d /

if then we can calculate [beta], [alpha], or [beta] - [alpha] for the
external shape of the shot, this equation will give the value of
[delta] and n required for stability of flight in the air.

The ellipsoid is the only shape for which [alpha] and [beta] have so
far been determined analytically, as shown already in § 44, so we must
restrict our calculation to an egg-shaped bullet, bounded by a prolate
ellipsoid of revolution, in which, with b = c,
_ _
/ [oo] ab² d[lambda] / [oo] ab² d[lambda]
A0 = | ------------------------------------------------------- = | -----------------------------------, (13)
_/ 0 (a² + [lambda])[root][4(a² + [lambda])(b² + [lambda])²] _/ 0 2(a² + [lambda])^3/2 (b² + [lambda])

A0 + 2B0 = 1, (14)

A0 B0 1 - A0 1
[alpha] = ------, [beta] = ------ = ------ = ------------. (15)
1 - A0 1 - B0 1 + A0 1 + 2[alpha]

The length of the shot being denoted by l and the calibre by d, and
the length in calibres by x

l/d = 2a/2b = x, (16)

x 1
A0 = ----------- ch^(-1)x - ------, (17)
(x²- 1)^3/2 x² - 1

-x x²
2B0 = ----------- ch^(-1)x + ------, (18)
x² - 1)^3/2 x² + 1

x sh^(-1) [root](x²-1) x
x²A0 + 2B0 = ---------------------- = ------------ log [x + [root](x² - 1)]. (19)
[root](x²-1) [root](x²-1)

If [sigma] denotes the density of the metal, and if the shell has a
cavity homothetic with the external ellipsoidal shape, a fraction f of
the linear scale; then the volume of a round shot being 1/6 [pi] d^3,
and 1/6 [pi] d^3 x of a shot x calibres long

W = 1/6 [pi] d^3 x (i - f^3) [sigma], (20)

d²
Wk1² = 1/6 [pi] d^3 x -- (1 - f^5) [sigma], (21)
10

l² + d²
Wk2² = 1/6 [pi] d^3 x ------- (1 - f^5) [sigma]. (22)
20

If [rho] denotes the density of the air or medium

W´ = 1/6 [pi] d^3 x [rho], (23)

W´ 1 [rho]
--- = ------ -------, (24)
W 1 - f³ [sigma]

k1² 1 1 - f^5 k2² x² + 1
--- = --- -------, ---- = ------, (25)
d² 10 1 - f³ k1² 2

[rho] x² + 1
tan² [delta] = ------- ([beta] - [alpha])-------------, (26)
[sigma] 1/5 (1 - f^5)

in which [sigma]/[rho] may be replaced by 800 times the S.G. of the
metal, taking water as 800 times denser than air on the average, in
round numbers, and formula (10) may be written n tan [delta] = [pi],
or n[delta] = 180, when [delta] is a small angle, and given in
degrees.

From this formula (26) the table following has been calculated by A.
G. Hadcock, and the results are in agreement with practical
experience.

52. In the steady motion the centre of the shot describes a helix,
with axial velocity

/ c1 \
u cos [theta]= v sin [theta] = ( l + -- tan² [theta] ) u cos [theta] [asympt] u sec [theta], (1)
\ c2 /

and transverse velocity

/ c1 \
u sin [theta] - v cos [theta] = ( l - -- ) u sin [theta] [asympt] ([beta] - [alpha]) u sin [theta]; (2)
\ c2 /

and the time of completing a turn of the spiral is 2[pi]/[mu].

When [mu] has the critical value in (7),

2[pi] 4[pi] C2 2[pi]
----- = ----- -- cos [theta] = ----- (x² + 1) cos [theta], (3)
[mu] p C1 p

which makes the circumference of the cylinder on which the helix
is wrapped

2[pi] 2[pi]u
-----(u sin [theta] - v cos [theta] = ------ ([beta] - [alpha]) (x² + 1) sin² [theta] cos [theta]
[mu] p

= nd ([beta] - [alpha]) (x² + 1) sin [theta] cos [theta], (4)

and the length of one turn of the helix

2[pi]
----- (u cos [theta] + v sin [theta] ) = nd(x² + 1); (5)
[mu]

thus for x = 3, the length is 10 times the pitch of the rifling.

53. _The Motion of a Perforated Solid in Liquid._--In the preceding
investigation, the liquid stops dead when the body is brought to rest;
and when the body is in motion the surrounding liquid moves in a
uniform manner with respect to axes fixed in the body, and the force
experienced by the body from the pressure of the liquid on its surface
is the opposite of that required to change the motion of the liquid;
this has been expressed by the dynamical equations given above. But if
the body is perforated, the liquid can circulate through a hole, in
reentrant stream lines linked with the body, even while the body is at
rest; and no reaction from the surface can influence this circulation,
which may be supposed started in the ideal manner described in § 29,
by the application of impulsive pressure across an ideal membrane
closing the hole, by means of ideal mechanism connected with the body.
The body is held fixed, and the reaction of the mechanism and the
resultant of the impulsive pressure on the surface are a measure of
the impulse, linear [xi], [eta], [zeta], and angular [lambda], [mu],
[nu], required to start the circulation.

This impulse will remain of constant magnitude, and fixed relatively
to the body, which thus experiences an additional reaction from the
circulation which is the opposite of the force required to change the
position in space of the circulation impulse; and these extra forces
must be taken into account in the dynamical equations.

An article may be consulted in the _Phil. Mag._, April 1893, by G. H.
Bryan, in which the analytical equations of motion are deduced of a
perforated solid in liquid, from considerations purely hydrodynamical.

The effect of an external circulation of vortex motion on the motion
of a cylinder has been investigated in § 29; a similar procedure will
show the influence of circulation through a hole in a solid, taking as
the simplest illustration a ring-shaped figure, with uniplanar motion,
and denoting by [xi] the resultant axial linear momentum of the
circulation.

As the ring is moved from O to O´ in time t, with velocity Q, and
angular velocity R, the components of liquid momentum change from

[alpha]M´U + [xi] and [beta]M´V along Ox and Oy

to

[alpha]M´U´ + [xi] and [beta]M´V´ along O´x´ and O´y´, (1)

the axis of the ring changing from Ox to O´x´; and

U = Q cos [theta], V = Q sin [theta],

U´ = Q cos ([theta] - Rt), V´ = Q sin ([theta] - Rt), (2)

so that the increase of the components of momentum, X1, Y1, and N1,
linear and angular, are

X1 = ([alpha]M´U´ + [xi])[cos] Rt - [alpha]M´U - [xi] - [beta]M´V´ sin Rt
= ([alpha] - [beta])M´Q sin ([theta] - Rt) sin Rt - [xi] ver Rt (3)

Y1 = ([alpha]M´U´ + [xi]) sin Rt + [beta]M´V´ cos Rt - [beta]M´V
= ([alpha] - [beta])M´Q cos ([theta] - Rt) sin Rt + [xi] sin RT, (4)

N1 = [ -([alpha]M´U´ + [xi]) sin ([theta] - Rt) + [beta]M´V´ cos ([theta] - Rt)OO´
= [ -([alpha] - [beta])´Q cos ([theta] - Rt) sin ([theta] - Rt) - [xi] sin ([theta] - Rt)]Qt. (5)

The components of force, X, Y, and N, acting on the liquid at O, and
reacting on the body, are then

X = lt. X1/t = ([alpha] - [beta])M´QR sin [theta] = ([alpha] - [beta])M´VR, (6)

Y = lt. Y1/t = ([alpha] - [beta])M´QR cos [theta] + [xi]R = ([alpha] - [beta])M´UR + [xi]R, (7)

Z = lt. Z1/t = -([alpha] - [beta])M´Q² sin [theta]cos[theta] - [xi]Q sin [theta]
= [-([alpha] - [beta])M´U + [xi]]V. (8)

Now suppose the cylinder is free; the additional forces acting on the
body are the components of kinetic reaction of the liquid

/ dU \ / dV \ dR
-[alpha]M´ ( -- - VR ), -[beta]M´ ( -- + UR ), [epsilon]C´--, (9)
\ dt / \ dt / dt

so that its equations of motion are

/ dU \ / dU \
M ( -- -VR ) = -[alpha]M´ ( -- - VR ) - ([alpha] - [beta])M´VR, (10)
\ dt / \ dt /

/ dV \ / dV \
M ( -- + UR ) = -[beta]M´ ( -- + UR ) - ([alpha] - [beta])M´UR - [xi]R, (11)
\ dt / \ dt /

dR dR
C-- = -[epsilon]C´-- + ([alpha] - [beta])M´UV + [xi]V; (12)
dt dt

and putting as before

M + [alpha]M´ = c1, M + [beta]M´ = c2, C + [epsilon]C´ = C3, (13)

dU
c1-- - c2 VR = 0, (14)
dt

dV
c2-- + (c1 U + [xi])R = 0, (15)
dt

dR
c3-- - (c1U + [xi] - c2 U)V = 0; (16)
dt

showing the modification of the equations of plane motion, due to the
component [xi] of the circulation.

The integral of (14) and (15) may be written

c1U + [xi] = F cos [theta], c2V = - F sin [theta], (17)

dx F cos² [theta] F sin² [theta] [xi]
-- = U cos [theta] - V sin [theta] = -------------- + -------------- - ---- cos [theta], (18)
dt c1 c2 c1

d[mu] / F F \ [xi]
----- = U sin [theta] + V cos [theta] = ( -- - -- ) sin [theta] cos [theta] - ---- sin [theta], (19)
dt \ c1 c2 / c1

d²[theta] / F² F² \ F[xi] d[mu]
C3--------- = ( -- - -- ) sin [theta] cos [theta] - ----- sin [theta] = F-----, (20)
dt² \ c1 c2 / c1 dt
_ _
d[theta] / | F² cos² [theta] F² sin² [theta] F[xi] |
C3-------- = Fy = / | - --------------- - --------------- + 2----- cos [theta] + H |; (21)
dt \/ |_ c1 c2 c1 _|

so that cos [theta] and y is an elliptic function of the time.

When [xi] is absent, dx/dt is always positive, and the centre of the
body cannot describe loops; but with [xi], the influence may be great
enough to make dx/dt change sign, and so loops occur, as shown in A.
B. Basset's _Hydrodynamics_, i. 192, resembling the trochoidal curves,
which can be looped, investigated in § 29 for the motion of a cylinder
under gravity, when surrounded by a vortex.

The branch of hydrodynamics which discusses wave motion in a liquid or
gas is given now in the articles SOUND and WAVE; while the influence
of viscosity is considered under HYDRAULICS.

REFERENCES.--For the history and references to the original memoirs
see _Report to the British Association_, by G. G. Stokes (1846), and
W. M. Hicks (1882). See also the _Fortschritte der Mathematik_, and A.
E. H. Love, "Hydrodynamik" in the _Encyklöpadie der mathematischen
Wissenschaften_ (1901). (A. G. G.)

HYDROMEDUSAE, a group of marine animals, recognized as belonging to the Hydrozoa (q.v.) by the following characters. (1) The polyp (hydropolyp) is of simple structure, typically much longer than broad, without ectodermal oesophagus or mesenteries, such as are seen in the anthopolyp (see article ANTHOZOA); the mouth is usually raised above the peristome on a short conical elevation or hypostome; the ectoderm is without cilia. (2) With very few exceptions, the polyp is not the only type of individual that occurs, but alternates in the life-cycle of a given species, with a distinct type, the medusa (q.v.), while in other cases the polyp-stage may be absent altogether, so that only medusa-individuals occur in the life-cycle.

The Hydromedusae represent, therefore, a sub-class of the Hydrozoa. The only other sub-class is the Scyphomedusae (q.v.). The Hydromedusae contrast with the Scyphomedusae in the following points. (1) The polyp, when present, is without the strongly developed longitudinal retractor muscles, forming ridges (_taeniolae_) projecting into the digestive cavity, seen in the scyphistoma or scyphopolyp. (2) The medusa, when present, has a velum and is hence said to be _craspedote_; the nervous system forms two continuous rings running above and below the velum; the margin of the umbrella is not lobed (except in Narcomedusae) but entire; there are characteristic differences in the sense-organs (see below, and SCYPHOMEDUSAE); and gastral filaments (phacellae), subgenital pits, &c., are absent. (3) The gonads, whether formed in the polyp or the medusa, are developed in the ectoderm.

The Hydromedusae form a widespread, dominant and highly differentiated group of animals, typically marine, and found in all seas and in all zones of marine life. Fresh-water forms, however, are also known, very few as regards species or genera, but often extremely abundant as individuals. In the British fresh-water fauna only two genera, _Hydra_ and _Cordylophora_, are found; in America occurs an additional genus, _Microhydra_. The paucity of fresh-water forms contrasts sharply, with the great abundance of marine genera common in all seas and on every shore. The species of _Hydra_, however, are extremely common and familiar inhabitants of ponds and ditches.

In fresh-water Hydromedusae the life-cycle is usually secondarily simplified, but in marine forms the life-cycle may be extremely complicated, and a given species often passes in the course of its history through widely different forms adapted to different habitats and modes of life. Apart from larval or embryonic forms there are found typically two types of person, as already stated, the polyp and the medusa, each of which may vary independently of the other, since their environment and life-conditions are usually quite different. Hence both polyp and medusa present characters for classification, and a given species, genus or other taxonomic category may be defined by polyp-characters or medusa-characters or by both combined. If our knowledge of the life-histories of these organisms were perfect, their polymorphism would present no difficulties to classification; but unfortunately this is far from being the case. In the majority of cases we do not know the polyp corresponding to a given medusa, or the medusa that arises from a given polyp.[1] Even when a medusa is seen to be budded, from a polyp under observation in an aquarium, the difficulty is not always solved, since the freshly-liberated, immature medusa may differ greatly from the full-grown, sexually-mature medusa after several months of life on the high seas (see figs. 11, B, C, and 59, a, b, c). To establish the exact relationship it is necessary not only to breed but to rear the medusa, which cannot always be done in confinement. The alternative is to fish all stages of the medusa in its growth in the open sea, a slow and laborious method in which the chance of error is very great, unless the series of stages is very complete.

At present, therefore, classifications of the Hydromedusae have a more or less tentative character, and are liable to revision with increased knowledge of the life-histories of these organisms. Many groups bear at present two names, the one representing the group as defined by polyp-characters, the other as defined by medusa-characters. It is not even possible in all cases to be certain that the polyp-group corresponds exactly to the medusa-group, especially in minor systematic categories, such as families.

The following is the main outline of the classification that is Adopted in the present article. Groups founded on polyp-characters are printed in ordinary type, those founded on medusa-characters in italics. For definitions of the groups see below.

Sub-class Hydromedusae (_Hydrozoa Craspedota_).

Order I. Eleutheroblastea.
" II. Hydroidea (_Leptolinae_).
Sub-order 1. Gymnoblastea (_Anthomedusae_).
" 2. Calyptoblastea (_Leptomedusae_).
Order III. Hydrocorallinae.
" IV. Graptolitoidea.
" V. Trachylinae.
Sub-order 1. _Trachomedusae_.
" 2. _Narcomedusae_.
Order VI. Siphonophora.
Sub-order 1. Chondrophorida.
" 2. Calycophorida.
" 3. Physophorida.
" 4. Cystophorida.

_Organization and Morphology of the Hydromedusae._

As already stated, there occur in the Hydromedusae two distinct types of person, the polyp and the medusa; and either of them is capable of non-sexual reproduction by budding, a process which may lead to the formation of colonies, composed of more or fewer individuals combined and connected together. The morphology of the group thus falls naturally into four sections--(1) the hydropolyp, (2) the polyp-colony, (3) the hydromedusa, (4) the medusa-colonies. Since, however, medusa-colonies occur only in one group, the Siphonophora, and divergent views are held with regard to the morphological interpretation of the members of a siphonophore, only the first three of the above subdivisions of hydromedusa morphology will be dealt with here in a general way, and the morphology of the Siphonophora will be considered under the heading of the group itself.

a, Hydranth;
b, Hydrocaulus;
c, Hydrorhiza;
t, Tentacle;
ps, Perisarc, forming in the region of the hydranth a cup or
hydrotheca (h, t),--which, however, is only found in polyps of the
order Calyptoblastea.]

1. _The Hydropolyp_ (fig. 1)--The general characters of this organism
are described above and in the articles HYDROZOA and POLYP. It is
rarely free, but usually fixed and incapable of locomotion. The foot
by which it is attached often sends out root-like processes--the
_hydrorhiza_ (c). The column (b) is generally long, slender and
stalk-like (_hydrocaulus_). Just below the crown of tentacles,
however, the body widens out to form a "head," termed, the _hydranth_
(a), containing a stomach-like dilatation of the digestive cavity. On
the upper face of the hydranth the crown of tentacles (t) surrounds
the peristome, from which rises the conical hypostome, bearing the
mouth at its extremity. The general ectoderm covering the surface of
the body has entirely lost the cilia present in the earlier larval
stages (planula), and may be naked, or clothed in a cuticle or
exoskeleton, the perisarc (ps), which in its simplest condition is a
chitinous membrane secreted by the ectoderm. The perisarc when present
invests the hydrorhiza and hydrocaulus; it may stop short below the
hydranth, or it may extend farther. In general there are two types of
exoskeleton, characteristic of the two principal divisions of the
Hydroidea. In the Gymnoblastea the perisarc either stops below the
hydranth, or, if continued on to it, forms a closely-fitting
investment extending as a thin cuticle as far as the bases of the
tentacles (e.g. _Bimeria_, see G. J. Allman [1],[2] pl. xii. figs, 1
and 3). In the Calyptoblastea the perisarc is always continued above
the hydrocaulus, and forms a cup, the hydrangium or hydrotheca (h, t),
standing off from the body, into which the hydranth can be retracted
for shelter and protection.

FIG. 2.--_Stauridium productum_, portion of the colony magnified; p,
polyp; rh, hydrorhiza.]

The architecture of the hydropolyp, simple though it be, furnishes a
long series of variations affecting each part of the body. The
greatest variation, however, is seen in the tentacles. As regards
number, we find in the aberrant forms _Protohydra_ and _Microhydra_
tentacles entirely absent. In the curious hydroid _Monobrachium_ a
single tentacle is present, and the same is the case in _Clathrozoon_;
in _Amphibrachium_ and in _Lar_ (fig. 11, A) the polyp bears two
tentacles only. The reduction of the tentacles in all these forms may
be correlated with their mode of life, and especially with living in a
constant current of water, which brings food-particles always from one
direction and renders a complete whorl or circle of tentacles
unnecessary. Thus _Microhydra_ lives amongst Bryozoa, and appears to
utilize the currents produced by these animals. _Protohydra_ occurs in
oyster-banks and _Monobrachium_ also grows on the shells of bivalves,
and both these hydroids probably fish in the currents produced by the
lamellibranchs. _Amphibrachium_ grows in the tissues of a sponge,
_Euplectella_, and protrudes its hydranth into the canal-system of the
sponge; and _Lar_ grows on the tubes of the worm _Sabella_. With the
exception of these forms, reduced for the most part in correlation
with a semi-parasitic mode of life, the tentacles are usually
numerous. It is rare to find in the polyp a regular, symmetrical
disposition of the tentacles as in the medusa. The primitive number of
four in a whorl is seen, however, in _Stauridium_ (fig. 2) and
_Cladonema_ (Allman [1], pl. xvii.), and in _Clavatella_ each whorl
consists regularly of eight (Allman, _loc. cit._ pl. xviii.). As a
rule, however, the number in a whorl is irregular. The tentacles may
form a single whorl, or more than one; thus in _Corymorpha_ (fig. 3)
and _Tubularia_ (fig. 4) there are two circlets; in _Stauridium_ (fig.
2) several; in _Coryne_ and _Cordylophora_ the tentacles are scattered
irregularly over the elongated hydranth.

As regards form, the tentacles show a number of types, of which the
most important are (1) filiform, _i.e._ cylindrical or tapering from
base to extremity, as in _Clava_ (fig. 5); (2) capitate, i.e. knobbed
at the extremity, as in _Coryne_ (see Allman, loc. cit. pl. iv.); (3)
branched, a rare form in the polyp, but seen in _Cladocoryne_ (see
Allman, loc. cit. p. 380, fig. 82). Sometimes more than one type of
form is found in the same polyp; in _Pennaria_ and _Stauridium_ (fig.
2) the upper whorls are capitate, the lower filiform. Finally, as
regards structure, the tentacles may retain their primitive hollow
nature, or become solid by obliteration of the axial cavity.

The hypostome of the hydropolyp may be small, or, on the other hand,
as in _Eudendrium_ (Allman, loc. cit. pls. xiii., xiv.), large and
trumpet-shaped. In the curious polyp _Myriothela_ the body of the
polyp is differentiated into nutritive and reproductive portions.

_Histology._--The ectoderm of the hydropolyp is chiefly sensory,
contractile and protective in function. It may also be glandular in
places. It consists of two regions, an external epithelial layer and a
more internal sub-epithelial layer.

The epithelial layer consists of (1) so-called "indifferent" cells
secreting the perisarc or cuticle and modified to form glandular cells
in places; for example, the adhesive cells in the foot. (2) Sensory
cells, which may be fairly numerous in places, especially on the
tentacles, but which occur always scattered and isolated, never
aggregated to form sense-organs as in the medusa. (3) Contractile or
myo-epithelial cells, with the cell prolonged at the base into a
contractile muscle-fibre (fig. 6, B). In the hydropolyp the ectodermal
muscle-fibres are always directed longitudinally. Belonging primarily
to the epithelial layer, the muscular cells may become secondarily
sub-epithelial.

FIG. 5.--Colonies of _Clava_. A, _Clava squamata_, magnified. B, _C.
multicornis_, natural size; p, _polyp_; _gon_, gonophores; rh,
hydrorhiza.]

The sub-epithelial layer consists primarily of the so-called
interstitial cells, lodged between the narrowed basal portions of the
epithelial cells. From them are developed two distinct types of
histological elements; the genital cells and the cnidoblasts or
mother-cells of the nematocysts. The sub-epithelial layer thus
primarily constituted may be recruited by immigration from without of
other elements, more especially by nervous (ganglion) cells and
muscle-cells derived from the epithelial layer. In its fullest
development, therefore, the sub-epithelial layer consists of four
classes of cell-elements.

The genital cells are simple wandering cells (archaeocytes), at first
minute and without any specially distinctive features, until they
begin to develop into germ-cells. According to Wulfert [60] the
primitive germ-cells of _Gonothyraea_ can be distinguished soon after
the fixation of the planula, appearing amongst the interstitial cells
of the ectoderm. The germ-cells are capable of extensive migrations,
not only in the body of the same polyp, but also from parent to bud
through many non-sexual generations of polyps in a colony (A. Weismann
[58]).

The cnidoblasts are the mother-cells of the nematocysts, each cell
producing one nematocyst in its interior. The complete nematocyst
(fig. 7) is a spherical or oval capsule containing a hollow thread,
usually barbed, coiled in its interior. The capsule has a double wall,
an outer one (o.c.), tough and rigid in nature, and an inner one
(i.c.) of more flexible consistence. The outer wall of the capsule is
incomplete at one pole, leaving an aperture through which the thread
is discharged. The inner membrane is continuous with the wall of the
hollow thread at a spot immediately below the aperture in the outer
wall, so that the thread itself (f) is simply a hollow prolongation of
the wall of the inner capsule inverted and pushed into its cavity. The
entire nematocyst is enclosed in the cnidoblast which formed it. When
the nematocyst is completely developed, the cnidoblast passes outwards
so as to occupy a superficial position in the ectoderm, and a delicate
protoplasmic process of sensory nature, termed the _cnidocil_ (cn)
projects from the cnidoblast like a fine hair or cilium. Many points
in the development and mechanism of the nematocyst are disputed, but
it is tolerably certain (1) that the cnidocil is of sensory nature,
and that stimulation, by contact with prey or in other ways, causes a
reflex discharge of the nematocyst; (2) that the discharge is an
explosive change whereby the in-turned thread is suddenly everted and
turned inside out, being thus shot through the opening in the outer
wall of the capsule, and forced violently into the tissues of the
prey, or, it may be, of an enemy; (3) that the thread inflicts not
merely a mechanical wound, but instils an irritant poison, numbing and
paralysing in its action. The points most in dispute are, first, how
the explosive discharge is brought about, whether by pressure exerted
external to the capsule (i.e. by contraction of the cnidoblast) or by
internal pressure. N. Iwanzov [27] has brought forward strong grounds
for the latter view, pointing out that the cnidoblast has no
contractile mechanism and that measurements show discharged capsules
to be on the average slightly larger than undischarged ones. He
believes that the capsule contains a substance which swells very
rapidly when brought into contact with water, and that in the
undischarged condition the capsule has its opening closed by a plug of
protoplasm (x, fig. 7) which prevents access of water to the contents;
when the cnidocil is stimulated it sets in action a mechanism or
perhaps a series of chemical changes by which the plug is dissolved or
removed; as a result water penetrates into the capsule and causes its
contents to swell, with the result that the thread is everted
violently. A second point of dispute concerns the spot at which the
poison is lodged. Iwanzov believes it to be contained within the
thread itself before discharge, and to be introduced into the tissues
of the prey by the eversion of the thread. A third point of dispute is
whether the nematocysts are formed _in situ_, or whether the
cnidoblasts migrate with them to the region where they are most
needed; the fact that in _Hydra_, for example, there are no
interstitial cells in the tentacles, where nematocysts are very
abundant, is certainly in favour of the view that the cnidoblasts
migrate on to the tentacles from the body, and that like the genital
cells the cnidoblasts are wandering cells.

a, Undischarged nematocyst.
b, Commencing discharge.
c, Discharge complete.
cn, Cnidocil.
N, Nucleus of cnidoblast.
o.c, Outer capsule.
x, Plug closing the opening of the outer capsule.
i.c., Inner capsule, continuous with the wall of the filament, f.
b, Barbs.]

The muscular tissue consists primarily of processes from the bases of
the epithelial cells, processes which are contractile in nature and
may be distinctly striated. A further stage in evolution is that the
muscle-cells lose their connexion with the epithelium and come to lie
entirely beneath it, forming a sub-epithelial contractile layer,
developed chiefly in the tentacles of the polyp. The evolution of the
ganglion-cells, is probably similar; an epithelial cell develops
processes of nervous nature from the base, which come into connexion
with the bases of the sensory cells, with the muscular cells, and with
the similar processes of other nerve-cells; next the nerve-cell loses
its connexion with the outer epithelium and becomes a sub-epithelial
ganglion-cell which is closely connected with the muscular layer,
conveying stimuli from the sensory cells to the contractile elements.
The ganglion-cells of Hydromedusae are generally very small. In the
polyp the nervous tissue is always in the form of a scattered plexus,
never concentrated to form a definite nervous system as in the medusa.

FIG. 8.--Vacuolated Endoderm Cells of cartilaginous consistence from
the axis of the tentacle of a Medusa (_Cunina_).]

The endoderm of the polyp is typically a flagellated epithelium of
large cells (fig. 6), from the bases of which arise contractile
muscular processes lying in the plane of the transverse section of the
body. In different parts of the coelenteron the endoderm may be of
three principal types--(1) digestive endoderm, the primitive type,
with cells of large size and considerably vacuolated, found in the
hydranth; some of these cells may become special glandular cells,
without flagella or contractile processes; (2) circulatory endoderm,
without vacuoles and without basal contractile processes, found in the
hydrorhiza and hydrocaulus; (3) supporting endoderm (fig. 8), seen in
solid tentacles as a row of cubical vacuolated cells, occupying the
axis of the tentacle, greatly resembling notochordal tissue,
particularly that of _Amphioxus_ at a certain stage of development; as
a fourth variety of endodermal cells excretory cells should perhaps be
reckoned, as seen in the pores in the foot of _Hydra_ and elsewhere
(cf. C. Chun, HYDROZOA [1], pp. 314, 315).

The mesogloea in the hydropolyp is a thin elastic layer, in which may
be lodged the muscular fibres and ganglion cells mentioned above, but
which never contains any connective tissue or skeletogenous cells or
any other kind of special mesogloeal corpuscles.

FIG. 9.--Colony of _Hydractinia echinata_, growing on the Shell of a
Whelk. Natural size.]

FIG. 10.--Polyps from a Colony of _Hydractinia_, magnified. dz,
dactylozoid; gz, gastrozoid: b, blastostyle; gon, gonophores; rh,
hydrorhiza.]

2. _The Polyp-colony._--All known hydropolyps possess the power of
reproduction by budding, and the buds produced may become either
polyps or medusae. The buds may all become detached after a time and
give rise to separate and independent individuals, as in the common
_Hydra_, in which only polyp-individuals are produced and sexual
elements are developed upon the polyps themselves; or, on the other
hand, the polyp-individuals produced by budding may remain permanently
in connexion with the parent polyp, in which case sexual elements are
never developed on polyp-individuals but only on medusa-individuals,
and a true colony is formed. Thus the typical hydroid colony starts
from a "founder" polyp, which in the vast majority of cases is fixed,
but which may be floating, as in _Nemopsis_, _Pelagohydra_, &c. The
founder-polyp usually produces by budding polyp-individuals, and these
in their turn produce other buds. The polyps are all non-sexual
individuals whose function is purely nutritive. After a time the
polyps, or certain of them, produce by budding medusa-individuals,
which sooner or later develop sexual elements; in some cases, however,
the founder-polyp remains solitary, that is to say, does not produce
polyp-buds, but only medusa-buds, from the first (_Corymorpha_, fig.
3, _Myriothela_, &c.). In primitive forms the medusa-individuals are
set free before reaching sexual maturity and do not contribute
anything to the colony. In other cases, however, the
medusa-individuals become sexually mature while still attached to the
parent polyp, and are then not set free at all, but become appanages
of the hydroid colony and undergo degenerative changes leading to
reduction and even to complete obliteration of their original medusan
structure. In this way the hydroid colony becomes composed of two
portions of different function, the nutritive "trophosome," composed
of non-sexual polyps, and the reproductive "gonosome," composed of
sexual medusa-individuals, which never exercise a nutritive function
while attached to the colony. As a general rule polyp-buds are
produced from the hydrorhiza and hydrocaulus, while medusa-buds are
formed on the hydranth. In some cases, however, medusa-buds are formed
on the hydrorhiza, as in Hydrocorallines.

In such a colony of connected individuals, the exact limits of the
separate "persons" are not always clearly marked out. Hence it is
necessary to distinguish between, first, the "zooids," indicated in
the case of the polyps by the hydranths, each with mouth and
tentacles; and, secondly, the "coenosarc," or common flesh, which
cannot be assigned more to one individual than another, but consists
of a more or less complicated network of tubes, corresponding to the
hydrocaulus and hydrorhiza of the primitive independent
polyp-individual. The coenosarc constitutes a system by which the
digestive cavity of any one polyp is put into communication with that
of any other individual either of the trophosome or gonosome. In this
manner the food absorbed by one individual contributes to the welfare
of the whole colony, and the coenosarc has the function of circulating
and distributing nutriment through the colony.

The hydroid colony shows many variations in form and architecture
which depend simply upon differences in the methods in which polyps
are budded.

FIG. 11.--_Lar sabellarum_ and two stages of its Medusa, _Willia
stellata_. A, colony of _Lar_; B and C, young and adult medusae.]

In the first place, buds may be produced only from the hydrorhiza,
which grows out and branches to form a basal _stolon_, typically
net-like, spreading over the substratum to which the founder-polyp
attached itself. From the stolon the daughter-polyps grow up
vertically. The result is a spreading or creeping colony, with the
coenosarc in the form of a root-like horizontal network (fig. 5, B;
11, A). Such a colony may undergo two principal modifications. The
meshes of the basal network may become very small or virtually
obliterated, so that the coenosarc becomes a crust of tubes tending to
fuse together, and covered over by a common perisarc. Encrusting
colonies of this kind are seen in _Clava squamata_ (fig. 5, A) and
_Hydractinia_ (figs. 9, 10), the latter having the perisarc calcified.
A further very important modification is seen when the tubes of the
basal perisarc do not remain spread out in one plane, but grow in all
planes forming a felt-work; the result is a massive colony, such as is
seen in the so-called Hydrocorallines (fig. 60), where the interspaces
between the coenosarcal tubes are filled up with calcareous matter, or
_coenosteum_, replacing the chitinous perisarc. The result is a stony,
solid mass, which contributes to the building up of coral reefs. In
massive colonies of this kind no sharp distinction can be drawn
between hydrorhiza and hydrocaulus in the coenosarc; it is practically
all hydrorhiza. Massive colonies may assume various forms and are
often branching or tree-like. A further peculiarity of this type of
colony is that the entire coenosarcal complex is covered externally by
a common layer of ectoderm; it is not clear how this covering layer is
developed.

In the second place, the buds may be produced from the hydrocaulus,
growing out laterally from it; the result is an arborescent, tree-like
colony (figs. 12, 13). Budding from the hydrocaulus may be combined
with budding from the hydrorhiza, so that numerous branching colonies
arise from a common basal stolon. In the formation of arborescent
colonies, two sharply distinct types of budding are found, which are
best described in botanical terminology as the monopodial or racemose,
and the sympodial or cymose types respectively; each is characteristic
of one of the two sub-orders of the Hydroidea, the Gymnoblastea and
Calyptoblastea.

In the monopodial method (figs. 12, 14) the founder-polyp is,
theoretically, of unlimited growth in a vertical direction, and as it
grows up it throws out buds right and left alternately, so that the
first bud produced by it is the lowest down, the second bud is above
the first, the third above this again, and so on. Each bud produced by
the founder proceeds to grow and to bud in the same way as the founder
did, producing a side branch of the main stem. Hence, in a colony of
gymnoblastic hydroids, the oldest polyp of each system, that is to
say, of the main stem or of a branch, is the topmost polyp; the
youngest polyp of the system is the one nearest to the topmost polyp;
and the axis of the system is a true axis.

In the sympodial method of budding, on the other hand, the
founder-polyp is of limited growth, and forms a bud from its side,
which is also of limited growth, and forms a bud in its turn, and so
on (figs. 15, 16). Hence, in a colony of calyptoblastic hydroids, the
oldest polyp of a system is the lowest; the youngest polyp is the
topmost one; and the axis of the system is a false axis composed of
portions of each of the consecutive polyps. In this method of budding
there are two types. In one, the biserial type (fig. 15), the polyps
produce buds right and left alternately, so that the hydranths are
arranged in a zigzag fashion, forming a "scorpioid cyme," as in
_Obelia_ and _Sertularia_. In the other, the uniserial type (fig. 16),
the buds are formed always on the same side, forming a "helicoid
cyme," as in _Hydrallmania_, according to H. Driesch, in which,
however, the primitively uniserial arrangement becomes masked later by
secondary torsions of the hydranths.

In a colony formed by sympodial budding, a polyp always produces first
a bud, which contributes to the system to which it belongs, i.e.
continues the stem or branch of which its parent forms a part. The
polyp may then form a second bud, which becomes the starting point of
a new system, the beginning, that is, of a new branch; and even a
third bud, starting yet another system, may be produced from the same
polyp. Hence the colonies of Calyptoblastea may be complexly branched,
and the budding may be biserial throughout, uniserial throughout, or
partly one, partly the other. Thus in _Plumularidae_ (figs. 17, 18)
there is formed a main stem by biserial budding; each polyp on the
main stem forms a second bud, which usually forms a side branch or
_pinnule_ by uniserial budding. In this way are formed the familiar
feathery colonies of _Plumularia_, in which the pinnules are all in
one plane, while in the allied _Antennularia_ the pinnules are
arranged in whorls round the main biserial stem. The pinnules never
branch again, since in the uniserial mode of budding a polyp never
forms a second polyp-bud. On the other hand, a polyp on the main stem
may form a second bud which, instead of forming a pinnule by uniserial
budding, produces by biserial budding a branch, from which pinnules
arise as from the main stem (fig. 18--3, 6). Or a polyp on the main
stem, after having budded a second time to form a pinnule, may give
rise to a third bud, which starts a new biserial system, from which
uniserial pinnules arise as from the main stem--type of _Aglaophenia_
(fig. 19). The laws of budding in hydroids have been worked out in an
interesting manner by H. Driesch [13], to whose memoirs the reader
must be referred for further details.

_Individualization of Polyp-Colonies._--As in other cases where animal
colonies are formed by organic union of separate individuals, there is
ever a tendency for the polyp-colony as a whole to act as a single
individual, and for the members to become subordinated to the needs of
the colony and to undergo specialization for particular functions,
with the result that they simulate organs and their individuality
becomes masked to a greater or less degree. Perhaps the earliest of
such specializations is connected with the reproductive function.
Whereas primitively any polyp in a colony may produce medusa-buds, in
many hydroid colonies medusae are budded only by certain polyps termed
_blastostyles_ (fig. 10, b). At first not differing in any way from
other polyps (fig. 5), the blastostyles gradually lose their nutritive
function and the organs connected with it; the mouth and tentacles
disappear, and the blastostyle obtains the nutriment necessary for its
activity by way of the coenosarc. In the Calyptoblastea, where the
polyps are protected by special capsules of the perisarc, the
_gonothecae_ enclosing the blastostyles differ from the hydrothecae
protecting the hydranths (fig. 54).

In other colonies the two functions of the nutritive polyp, namely,
capture and digestion of food, may be shared between different polyps
(fig. 10). One class of polyps, the _dactylozoids_ (dz), lose their
mouth and stomach, and become elongated and tentacle-like, showing
great activity of movement. Another class, the _gastrozoids_ (gz),
have the tentacles reduced or absent, but have the mouth and stomach
enlarged. The dactylozoids capture food, and pass it on to the
gastrozoids, which swallow and digest it.

Besides the three types of individual above mentioned, there are other
appendages of hydroid colonies, of which the individuality is
doubtful. Such are the "guard-polyps" (machopolyps) of _Plumularidae_,
which are often regarded as individuals of the nature of dactylozoids,
but from a study of the mode of budding in this hydroid family Driesch
concluded that the guard-polyps were not true polyp-individuals,
although each is enclosed in a small protecting cup of the perisarc,
known as a nematophore. Again, the spines arising from the basal crust
of _Podocoryne_ have been interpreted by some authors as reduced
polyps.

3. _The Medusa._--In the Hydromedusae the medusa-individual occurs, as
already stated, in one of two conditions, either as an independent
organism leading a true life in the open seas, or as a subordinate
individuality in the hydroid colony, from which it is never set free;
it then becomes a mere reproductive appendage or _gonophore_, losing
successively its organs of sense, locomotion and nutrition, until its
medusoid nature and organization become scarcely recognizable. Hence
it is convenient to consider the morphology of the medusa from these
two aspects.

(a) _The Medusa as an Independent Organism._--The general structure
and characteristics of the medusa are described elsewhere (see
articles HYDROZOA and MEDUSA), and it is only necessary here to deal
with the peculiarities of the Hydromedusa.

FIG. 20.--_Cladonema radiatum_, the medusa walking on the basal
branches of its tentacles (t), which are turned up over the body.]

FIG. 21.--_Clavatella prolifera_, ambulatory medusa. t, tentacles; oc,
ocelli.]

As regards habit of life the vast majority of Hydromedusae are pelagic
organisms, floating on the surface of the open sea, propelling
themselves feebly by the pumping movements of the umbrella produced by
contraction of the sub-umbral musculature, and capturing their prey
with their tentacles. The genera _Cladonema_ (fig. 20) and
_Clavatella_ (fig. 21), however, are ambulatory, creeping forms,
living in rock-pools and walking, as it were, on the tips of the
proximal branches of each of the tentacles, while the remaining
branches serve for capture of food. _Cladonema_ still has the typical
medusan structure, and is able to swim about, but in _Clavatella_ the
umbrella is so much reduced, that swimming is no longer possible. The
remarkable medusa _Mnestra parasites_ is ecto-parasitic throughout
life on the pelagic mollusc _Phyllirrhoe_, attached to it by the
sub-umbral surface, and its tentacles have become rudimentary or
absent. It is interesting to note that _Mnestra_ has been shown by J.
W. Fewkes [15] and R. T. Günther [19] to belong to the same family
(_Cladonemidae_) as _Cladonema_ and _Clavatella_, and it is reasonable
to suppose that the non-parasitic ancestor of _Mnestra_ was, like the
other two genera, an ambulatory medusa which acquired louse-like
habits. In some species of the genus _Cunina_ (Narcomedusae) the
youngest individuals (actinulae) are parasitic on other medusae (see
below), but in later life the parasitic habit is abandoned. No other
instances are known of sessile habit in Hydromedusae.

FIG. 22.--_Corymorpha nutans_, adult female Medusa. Magnified 10
diameters.]

The external form of the Hydromedusae varies from that of a deep bell
or thimble, characteristic of the Anthomedusae, to the shallow
saucer-like form characteristic of the Leptomedusae. It is usual for
the umbrella to have an even, circular, uninterrupted margin; but in
the order Narcomedusae secondary down-growths between the tentacles
produce a lobed, indented margin to the umbrella. The marginal
tentacles are rarely absent in non-parasitic forms, and are typically
four in number, corresponding to the four perradii marked by the
radial canals. Interradial tentacles may be also developed, so that
the total number present may be increased to eight or to an
indefinitely large number. In _Willia_, _Geryonia_, &c., however, the
tentacles and radial canals are on the plan of six instead of four
(figs. 11 and 26). On the other hand, in some cases the tentacles are
less in number than the perradii; in _Corymorpha_ (figs. 3 and 22)
there is but a single tentacle, while two are found in _Amphinema_ and
_Gemmaria_ (Anthomedusae), and in _Solmundella bitentaculata_ (fig.
67) and _Aeginopsis hensenii_ (fig. 23) (Narcomedusae). The tentacles
also vary considerably in other ways than in number: first, in form,
being usually simple, with a basal bulb, but in _Cladonemidae_ they
are branched, often in complicated fashion; secondly, in grouping,
being usually given off singly, and at regular intervals from the
margin of the umbrella, but in _Margelidae_ and in some Trachomedusae
they are given off in tufts or bunches (fig. 24); thirdly, in position
and origin, being usually implanted on the extreme edge of the
umbrella, but in Narcomedusae they become secondarily shifted and are
given off high up on the ex-umbrella (figs. 23 and 25); and, fourthly,
in structure, being hollow or solid, as in the polyp. In some medusae,
for instance, the remarkable deep-sea family _Pectyllidae_, the
tentacles may bear suckers, by which the animal may attach itself
temporarily. It should be mentioned finally that the tentacles are
very contractile and extensible, and may therefore present themselves,
in one and the same individual, as long, drawn-out threads, or in the
form of short corkscrew-like ringlets; they may stream downwards from
the sub-umbrella, or be held out horizontally, or be directed upwards
over the ex-umbrella (fig. 23). Each species of medusa usually has a
characteristic method of carrying its tentacles.

FIG. 23.--_Aeginopsis hensenii_, slightly magnified, showing the
manner in which the tentacles are carried in life.]

FIG. 24.--_Rathkea octonemalis._]