Professional Documents
Culture Documents
Textbook Fatigue and Fracture of Weldments The Ibess Approach For The Determination of The Fatigue Life and Strength of Weldments by Fracture Mechanics Analysis Uwe Zerbst Ebook All Chapter PDF
Textbook Fatigue and Fracture of Weldments The Ibess Approach For The Determination of The Fatigue Life and Strength of Weldments by Fracture Mechanics Analysis Uwe Zerbst Ebook All Chapter PDF
https://textbookfull.com/product/fatigue-and-fracture-of-
nanostructured-materials-pasquale-cavaliere/
https://textbookfull.com/product/the-welding-engineer-s-guide-to-
fracture-and-fatigue-1st-edition-philippa-l-moore/
https://textbookfull.com/product/proceedings-of-the-8th-
international-conference-on-fracture-fatigue-and-wear-
ffw-2020-august-26-27-2020-magd-abdel-wahab/
https://textbookfull.com/product/the-fracture-of-brittle-
materials-testing-and-analysis-stephen-w-freiman/
The fracture of brittle materials : testing and
analysis Second Edition Freiman
https://textbookfull.com/product/the-fracture-of-brittle-
materials-testing-and-analysis-second-edition-freiman/
https://textbookfull.com/product/the-fracture-of-brittle-
materials-testing-and-analysis-second-edition-freiman-2/
https://textbookfull.com/product/a-practical-approach-to-
fracture-mechanics-jorge-luis-gonzalez-velazquez/
https://textbookfull.com/product/fracture-mechanics-fundamentals-
and-applications-fourth-edition-ted-l-anderson/
https://textbookfull.com/product/geologic-fracture-mechanics-1st-
edition-richard-a-schultz/
Uwe Zerbst · Mauro Madia · Benjamin Schork
Jonas Hensel · Pawel Kucharczyk
DesireTchoffo Ngoula · Didi Tchuindjang
Julian Bernhard · Carla Beckmann
Fatigue and
Fracture of
Weldments
The IBESS Approach for the
Determination of the Fatigue Life and
Strength of Weldments by Fracture
Mechanics Analysis
Fatigue and Fracture of Weldments
Uwe Zerbst Mauro Madia
•
Carla Beckmann
123
Uwe Zerbst DesireTchoffo Ngoula
Bundesanstalt für Materialforschung Materials Mechanics Group
und -prüfung (BAM) Technische Universität Darmstadt
Berlin, Germany Darmstadt, Hessen, Germany
Pawel Kucharczyk
IWT Solutions AG
Aachen, Nordrhein-Westfalen, Germany
This Springer imprint is published by the registered company Springer Nature Switzerland AG
The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland
Foreword
This book gives a comprehensive and thorough guide to the readers who are lost in
the forest of fatigue of weld. It presents the clear fracture mechanics and material
background of fatigue of weld by starting first from fatigue crack initiation, short
crack and next by discussing long cracks, crack closure, crack growth and
threshold, residual stress, stress concentration, stress intensity factor, J-integral,
multiple cracks, weld geometries and defects, microstructural parameters including
HAZ and cyclic stress–strain behavior. The authors have succeeded in treating all
these mutually interacting important parameters in a unique distinct way of anal-
ysis. The elaborate works performed by the deep scientific knowledge of the
authors should be admired.
This is the timely and excellent publication in the field of weld.
v
Notice
While the authors believe that the information and guidance given in this book are
correct, all parties making use of it must rely on their own skill and judgement. The
authors cannot assume any liability for loss or damage caused by any error or
omission in the application of the IBESS method. Any and all such liabilities are
disclaimed.
The authors do not give any warranty or guarantee whatsoever that the infor-
mation and guidance given in this book do not infringe the rights of any third party
or can be used for any particular purpose at all. Any person intending to use the
same should satisfy himself as to accuracy and the suitability for the purpose for
which it is intended to be used.
vii
Contents
ix
x Contents
xi
Abbreviations
xiii
xiv Abbreviations
s Secondary
SC Short crack
SE(B) Single notch bending specimen (fracture mechanics testing)
T Transverse
TIG Tungsten inert gas (welding)
UM Strength undermatching (rYW < rYB)
WM Weld metal
Nomenclature
xv
xvi Nomenclature
i Counting parameter; i ¼ 1, 2, 3, …
J J-integral (monotonic loading)
J0.2;BL Resistance against stable crack initiation (monotonic loading)
J0.2 Resistance against stable crack initiation (alternative definition)
k Depth of a secondary notch, e.g., an undercut
k Slope of finite life (high cycle) fatigue S-N curve in double
logarithmic scaling, Eq. (107)
K′, n′ Coefficients of the cyclic stress–strain curve (Ramberg–Osgood)
K Stress intensity factor (K-factor)
KJc Monotonic fracture resistance (formally derived from J-integral)
K0 Scale parameter in three-parameter Weibull distribution
Kmat Fracture resistance, monotonic loading (general term)
Kmax Maximum K-factor in a loading cycle
Kmed 50% percentile value of fracture resistance in the Master Curve
approach
Kmin Minimum K-factor in a loading cycle
Kmin Shift parameter in three-parameter Weibull distribution
Kop K-factor at crack opening
Kr K-factor due to residual stresses
Kr Ordinate of the FAD (¼K/Kmat)
KI K-factor for mode I crack opening (normal to the crack faces)
K Mean value of the K-factor (in the loading cycle)
‘ Section width along the weld toe (IBESS approach)
L Weld width, Fig. 77
Lr Ligament yielding parameter (monotonic loading)
m Shape parameter in three-parameter Weibull distribution
m(z,a) Weight function (K-factor determination)
M Strength mis-match ratio (commonly rYW/rYB)
N Number of loading cycles
N Number of specimens in a statistical test set (Figs. 29, 30, 41)
Nc Number of loading cycles up to fracture
p,q Fit parameters of the threshold and fracture regions of the da/dN–DK
curve, Eqs. (4) and (80)
P Probability
Pf Failure probability
Pt Roughness parameter used for specifying the secondary notch depth
k in the present study, Annex D
R Loading ratio (¼rmin/rmax or Kmin/Kmax)
ReL Lower yield strength (materials showing a Lüders’ plateau)
Rp0.2 0.2% proof strength (materials without Lüders’ plateau)
Rm Uniaxial tensile strength
SCF (Elastic) Stress concentration factor (at weld toe surface)
T Plate thickness
Ti Components of the traction vector for determining the J-integral
Nomenclature xvii
xix
xx Abstract
• The method has been validated with a large number of S-N curves including two
materials, three weldment types with two weld geometries, each referring to
different manufacturing technologies and the as-welded and stress-relieved state.
• Although not finally solved, an extended discussion is provided on the issue of
welding residual stresses including their redistribution under cyclic loading.
• A number of simplifications is proposed at lower analyses levels which, how-
ever, partly lack complete validation by now.
Introduction
The basic topic of the present study is the determination of the fatigue strength by
fracture mechanics. Common fracture mechanics-based fatigue considerations, e.g.,
within the frame of damage tolerance considerations, are usually limited to residual
lifetime determination of so-called long cracks which are often defined in con-
junction with nondestructive testing. The extension of this concept to the total
lifetime, as in the S-N curve approach, requires an adequate description of so-called
short crack propagation which cannot be based on the DK concept and must
consider the crack closure phenomenon as well as its gradual buildup at the short
crack stage. The terms long and short cracks will be explained in Sect. 1.1.3.
Further, it has to provide a meaningful definition of the initial crack dimensions
needed in fracture mechanics and a solution for the multiple crack problems at
stress levels higher than the fatigue limit as this is specific for configurations such as
weldments. A fundamental discussion of these aspects beyond its application to
welds is provided by Zerbst et al. (2018b, c).
The idea to apply fracture mechanics to the determination of the fatigue strength of
weldments is anything but new. First attempts date to around 1970 (Maddox 1970)
who wrote in 1974: “It is now widely recognized that flaws will inevitably exist in
welded structures and the old idea of removing all detectable defects must be replaced
by the ‘fitness for purpose’ design philosophy. This makes it necessary to define
reliable methods of assessing the significance of flaws, particularly in the context of
fatigue, …. The most promising approach to this problem lies in the use of the fracture
mechanics based description of fatigue crack propagation.” (Maddox 1974). At that
time, a major problem was to develop appropriate K-factor solutions. Ever since effort
was spent to further develop the approach. No detailed overview shall be given here;
see, however, Radaj et al. (2006), Chap. 6, for a review. Note that even today, fracture
mechanics besides some promising applications (see, e.g., Nykänen 2005, repre-
sentative for many other studies) is still not widely used as a tool for fatigue strength
determination of weldments. One reason for this is certainly that the approaches
usually fail to meet the above-formulated requirements. They are based on linear
elastic fracture mechanics even for short cracks and on fixed or (from S-N curves)
xxi
xxii Introduction
back-calculated initial crack sizes, they usually consider one crack only, neglect the
variation in the local weld geometry along the weld toe, etc.
The present study aims at a discussion of all these points and at offering possible
solutions obtained within the frame of the German research cluster IBESS. It
combines rather basic notes with the introduction of a methodology for the fracture
mechanics-based determination of the fatigue strength of welds, but also with a
critical discussion of hitherto not satisfactorily resolved problems.
Partners Involved in the IBESS Cluster Project
xxiii
xxiv Partners Involved in the IBESS Cluster Project
1 Background Information
Fatigue cracks are initiated at the smooth surface in plain specimens (i.e., at surface
extrusions and intrusions caused by persistent slip bands) or—more frequently—at
geometrical discontinuities (which act as geometric stress concentrators) and/or
material defects such as non-metallic inclusions (which cause strain concentration
zones due to their different stiffness compared to the matrix material and sometimes
also chemical mis-match), see Zerbst et al. (2018b). Polak (2003) notes with respect
to the crack nucleation stage: “Numerous studies have shown that in the majority
of materials and under normal loading conditions, the period of crack initiation in
smooth specimens without defects amounts to less than 5–20% of the fatigue life. In
materials containing defects, the fraction of life spent in crack initiation is even lower.
The major part of the life is spent in the growth of cracks, namely in the growth of
short cracks.” That this statement usually is also true with respect to welds has been
confirmed, e.g., by Verreman and Nie (1996) for manual fillet welds of structural
steel.
Subsequent to crack initiation, fatigue crack propagation is characterized by dis-
tinct successive stages, Fig. 1, see, e.g., Tanaka (2003), Davidson et al. (2003).
As long as its size is still in the order of microstructural features such as the grain
size, the crack is designated as microstructurally short. The crack and the plastic
zone ahead of its tip is completely embedded within a single or a few grains. As the
© Springer Nature Switzerland AG 2019 1
U. Zerbst et al., Fatigue and Fracture of Weldments,
https://doi.org/10.1007/978-3-030-04073-4_1
2 Fatigue and Fracture of Weldments
consequence, crack growth is strongly affected by the local microstructure and char-
acterized by (temporary) acceleration and deceleration phases or even crack arrest.
Typical growth barriers are grain boundaries between adjacent grains of different
crystal orientation where the generation of the plastic zone is inferred. An example
is shown in Fig. 2. The microstructural crack propagation stage is terminated when the
crack depth (of the initially semi-circular or semi-elliptical crack) reaches the order
of one or a few grain sizes (Murakami 2002); a few, since not every grain boundary
will act as a crack stopper when, e.g., the crystal orientation of two adjacent grains
is not significantly different. Note that, in general terms, it is not the crack depth
but the crack front length which is the critical parameter. E.g., Capelli et al. (2008),
investigating single phase aluminium alloy, restrict the microstructurally short crack
stage to a crack front length intersecting 15 grains. That also means that the above
statement about the critical crack depth no longer applies for through surface cracks
or long shallow surface cracks showing a small crack aspect ratio a/c. The applica-
tion of classical fracture mechanics to microstructurally short cracks is not possible
(Pippan and Riemelmoser 2003).
A microstructurally short crack which is not arrested will finally become a mechan-
ically short crack. The general definition of this is a size up to the order of the plastic
zone size ahead of its tip. As the consequence, the linear elastic fracture mechanics
concept based on the K-factor range K is not applicable but elastic-plastic concepts
based on the cyclic J-integral, J, or a cyclic crack tip opening displacement, δ,
have to be used. This will be discussed in Sect. 2.2.
The term “physically short crack” refers to the observation that during that stage
the crack closure phenomenon is being gradually built-up. It does not exist at the very
beginning, but then the effect intensifies with increasing crack size (and crack wake
1 Background Information 3
Fig. 2 Non-propagating crack in EA1N railway axle steel. The crack stopped within a grain. It
experienced more than 108 loading cycles; according to Beretta et al. (2009)
length) and finally it reaches a stable, nearly crack size independent state when the
crack becomes a long one. The crack closure phenomenon is the topic of Sect. 1.2.
Note that the difference in the terms “mechanically” and “physically” short cracks
is not different crack sizes (these at least will overlap) but the different phenomena.
A mechanically/physically short crack becomes a long one when the crack closure
effect stabilizes, i.e. it becomes independent of the crack size and when the linear
elastic K concept is applicable (see above). An exception is low cycle fatigue (LCF)
which occurs at stress levels close to the yield strength. Excluding this exception,
long crack propagation can be described by the da/dN–K curve and the long fatigue
crack propagation (FCP) threshold Kth,LC . The long crack stage is terminated by
the failure of the structure, i.e., by monotonic fracture but there might also be other
failure criteria associated with loss of functionality.
Figure 3 provides a scheme of the crack sizes referring to the stages discussed
above. Note, however, that the numbers are strongly material dependent. Therefore,
the figure should be regarded as a first impression only which roughly describes the
relations for steel.
4 Fatigue and Fracture of Weldments
Fig. 3 Crack length scales of the life cycle of a component subjected to cyclic loading; according
to Tanaka (2003)
1.1.5 Failure
Failure occurs when the crack reaches a critical size such that the maximum crack
driving force in the loading cycle, e.g. Kmax , reaches or exceeds the monotonic
fracture resistance of the material. Besides this, other failure criteria are possible as
well when the functionality of the component is impaired even if the component does
not break yet.
1 Background Information 5
Since it can be assumed that a fatigue crack does not grow when the crack is closed,
i.e., its faces are in contact during the lower part of the loading cycle at negative
R ratios (R Kmin /Kmax ), only part of the stress intensity factor range K (= Kmax
− Kmin ) will contribute to its propagation. This is designated by an effective value
Keff (= Kmax − Kop ) with Kop being the K-factor at crack opening, for the definition
see Fig. 4a. As a first approximation, Keff should be ½ K for R −1, however,
it is observed to be even smaller, i.e., the crack is open only at stress levels above
zero.1 This is designated as “crack closure phenomenon”. It is commonly described
by a crack closure parameter U such that
Keff U · K (1)
1 Sometimes also below, depending on the R-ratio and the applied stress ranges
6 Fatigue and Fracture of Weldments
Several mechanisms are responsible for the crack closure phenomenon. The three
most important ones are the plasticity-induced, the roughness-induced and the oxide
debris-induced crack closure mechanisms (Fig. 4b–d). For these and further ones
see, e.g., Tanaka and Akiniwa (2003) and Suresh (2003).
Under load, a plastic zone is formed ahead of the crack tip. When the crack prop-
agates, this will remain as plastically stretched material at the surfaces of the wake
of the crack (Fig. 4b). The consequence is geometrical mis-match particularly at the
specimen sides where the plastic zone dimension is largest due to prevailing plane
stress conditions. Since the plastic zone size increases quadratically with the applied
load and decreases quadratically with the yield strength of the material (for small
scale yielding conditions) the plasticity-induced mechanism becomes stronger for
higher maximum loads in the loading cycle and for smaller yield strengths. This is
at least the case for long cracks.
The roughness of the crack faces causes asperity contact (Fig. 4c) which is enhanced
by mixed mode loading components (shear) due to the kinking or branching of the
crack. Usually it will be affected by the grain size in that a larger grain size tends to
a rougher surface. Note that the roughness-induced effect plays a major role in the
threshold regime of the da/dN–K curve.
The authors have already pointed to the importance of crack arrest for fatigue life
and strength considerations. It is immediately evident that only cracks which do not
arrest, negatively affect structural integrity. Crack arrest can happen at any of the
crack propagation stages.
As mentioned above, microstructurally short cracks arrest at microstructural bar-
riers, usually grain boundaries. It is important to understand that the fatigue limit is
not a critical stress for crack initiation but a critical stress for avoiding or overcoming
crack arrest (Miller 1993; Murakami 2002). With respect to the material fatigue limit
(determined with smooth, unnotched specimens) it is that stress below which none
of the numerous microstructural short cracks will grow permanently. Note, however,
that in real (usually notched) structures other arrest mechanisms will dominate as
will be discussed below. Note further that the material fatigue limit will disappear
when mechanisms such as corrosion as a time dependent phenomenon are active
which will overcome the crack arrest barriers.
Crack arrest of physically short cracks occurs due to the gradual build-up of the
crack closure effects when the increase of K with crack depth is smaller than the
decrease of Keff due to an increasing Kop .
Long cracks arrest when the applied K is below the fatigue crack propagation
threshold Kth .
This is probably the most common crack arrest scenario in components (Murakami
2002, 2003; Tanaka 2003). Due to the stress gradient away from the notch the rate
of increase in K becomes smaller, an effect which is additionally superimposed
by the gradual build-up of the crack closure phenomenon. When the resultant Keff
falls below the fatigue crack propagation threshold Kth , crack arrest will occur.
8 Fatigue and Fracture of Weldments
Further crack arrest cases are possible. Three options are: (i) Crack arrest can occur
when the crack grows into a stress shielded area, e.g., near a geometrical stiffener
as it is typical for thin-wall structures. (ii) Arrest also occurs when the crack grows
towards an area of compressive residual stresses (e.g., in deep-rolled crankshafts)
(iii) It also can happen when the crack propagates into a material with a higher
threshold Kth than that from where it started. An example is crack transition from
a fine to a coarse grain zone, e.g. a heat affected zone in a weld. Note that the coarser
grain is usually associated with a higher threshold (Plekhov et al. 2011). A plausible
explanation for this is a stronger roughness-induced crack closure effect when the
larger grain is associated with rougher crack faces, see Sect. 1.2.3.
Whilst, as mentioned above, the crack arrest mechanism of the microstructurally
short cracks (Sect. 1.3.1) determines the material’s fatigue limit, the component
fatigue limit is usually controlled by the other mechanisms.
Fig. 5 Crack growth diagram da/dN–K for long fatigue crack propagation; schematic view
Another random document with
no related content on Scribd:
Family 1. Syllidae see p. 306 Family 8. Amphinomidae see p. 318
" 2. Hesionidae " 308 " 9. Eunicidae " 318
" 3. Aphroditidae " 309 " 10. Glyceridae " 320
" 4. Phyllodocidae " 313 " 11. Sphaerodoridae " 320
" 5. Tomopteridae " 315 " 12. Ariciidae " 321
" 6. Nereidae " 315 " 13. Typhloscolecidae " 321
" 7. Nephthydidae " 317
Sub-Order 2. Spioniformia.
Family 1. Spionidae see p. 321 Family 4. Magelonidae see p. 325
" 2. Polydoridae " 323 " 5. Ammocharidae " 325
" 3. Chaetopteridae " 323
Sub-Order 3. Terebelliformia.
Family 1. Cirratulidae see p. 325 Family 3. Ampharetidae see p. 330
" 2. Terebellidae " 327 " 4. Amphictenidae " 330
Sub-Order 4. Capitelliformia.
Family. Capitellidae, see p. 331.
Sub-Order 5. Scoleciformia.
Family 1. Opheliidae see p. 331 Family 4. Scalibregmidae see p. 334
" 2. Maldanidae " 332 " 5. Chlorhaemidae " 334
" 3. Arenicolidae " 333 " 6. Sternaspidae " 335
Branch B. Cryptocephala.
Sub-Order 1. Sabelliformia.
Family 1. Sabellidae see p. 336 Family 3. Amphicorinidae see p. 339
" 2. Eriographidae " 338 " 4. Serpulidae " 339
Sub-Order 2. Hermelliformia.
Family. Hermellidae, see p. 341.
In Aphroditidae and certain Amphinomidae the body is more or less oval in shape.
In Lipobranchius and Sternaspis it is grub-like, short, and cylindrical, with rounded
ends; in the former it is difficult to distinguish head and tail, or dorsal and ventral
surfaces.
The segments composing the trunk may be all alike, or may constitute two more or
less sharply marked regions, the thorax and abdomen, differing in the character of
the chaetae, or in their arrangement, or in some other way, as in the Sabelliformia
and the Capitelliformia.
The posterior extremity is generally more or less narrowed, and most of the
Nereidiformia are provided with special elongated cirri, borne by the anal segment.
In the Maldanidae and others the body terminates in a funnel, at the bottom of
which is placed the anus. Only in a few cases is the anus not terminal; in
Notopygos and other Amphinomidae, as well as in some species of Polynoë, it is
dorsal.[313] In Sabellaria and Pectinaria the hinder end of the body undergoes
great degeneration; in the former it is achaetous, but cylindrical and bent forwards
alongside the body (Fig. 131). In Pectinaria (Fig. 177), this region, which is called
the "scapha," is leaf-like, and serves to close the narrower end of the tube in which
the worm lives. Arenicola marina, and some Terebellids have no chaetae in the
hinder, narrower part of the body.
In the family Serpulidae one (rarely two) of the most dorsally placed gill filaments
is enlarged terminally, and acts as a stopper or "operculum," which closes the
mouth of the tube when the animal withdraws into it. Further, in Spirorbis this
operculum is grooved on one side, and serves as a brood pouch in which the eggs
undergo development (Fig. 184, p. 341). It will be seen, therefore, that the palps
may be very important organs for the life of the worm, and they are no less
interesting to the comparative anatomist, serving as they do as an excellent
illustration of the various uses which Nature finds for one and the same organ.
In the other sub-Orders the prostomium carries neither palps nor tentacles.
In the Cryptocephala there is never more than a single pair of tentacles, and these
are generally reduced to a group of sensory cells, though in Sabellaria they retain
a considerable size.
In a few genera, such as Aphrodite, Nephthys, Capitella, the first postoral segment
is distinguished from the succeeding segments only by its position with regard to
the mouth (Fig. 132) and by its smaller size. But in the remainder of the
Polychaeta, with here and there an exception, the peristomium is achaetous in the
adult.[319]
In a few cases, such as the Chlorhaemids and Sternaspis, and to a slight degree
in Arenicola, the "head" and even the anterior part of the worm is capable of being
withdrawn into the body.
In most Annelids the chaetae are in two bundles on each side, but there are
certain families in which the dorsal bundle, and even the notopodium itself, is
absent, as in the Eunicidae, Syllidae, and Phyllodocidae; or the dorsal bundle may
be absent only in certain regions of the body, as in the hind-body of Terebellids. In
some Amphinomidae and Aphroditidae the notopodium is scarcely distinct as a
separate lobe, being a slight tubercle on the upper surface of the neuropodium;
but the notopodial chaetae are present, and indeed particularly well developed in
many cases.
But whilst, in the Nereidiformia, the parapodia, whether consisting of two lobes or
only one, are always well developed, and project to a more or less pronounced
degree from the sides of the body, it is otherwise in the rest of the group, where
the chaetigerous lobes are usually reduced to mere tubercles or ridges, no doubt
in relation to their burrowing or tubicolous habits. In Sternaspis the chaetae issue
directly from the body-wall.
Amongst the Nereidiformia we find examples in which the parapodia, instead of
being more or less conical "legs," are flattened fore and aft so as to serve as
efficient "fins," as in the active swimmers, Nereis virens and Nephthys caeca, and
in the pelagic Phyllodocids, Alciopids, Typhloscolecids, and Tomopteris.
Of the typical dorsal and ventral cirri, the ventral is only absent in some
Amphinomids amongst the Nereidiformia; the dorsal is absent in Nephthys and
degenerate in Glycera, whilst in a very large number of families of the other sub-
Orders neither cirrus is present. These cirri, though originally filamentous and
sensory, may, by virtue of special blood supply, become "gills," and this occurs in
several families of different sub-Orders. Thus in Eunice this gill is comb-like; in
Amphinome and in Arenicola (on certain segments) it is arborescent, as it is also
in one to three segments in Terebellids; whilst in Ariciidae, Spioniformia,
Cirratulidae, Opheliidae, and Sabellaria it remains more or less finger-shaped or
filamentous. In the family Serpulidae the thoracic cirri, both dorsal and ventral,
become flattened and extended antero-posteriorly, and unite with one another to
form the "thoracic membrane."[320] In Phyllodocidae the cirri are foliaceous and
natatory, and they contain a great quantity of glands of a peculiar character. The
Aphroditidae are distinguished from other Annelids by the possession of "elytra" or
dorsal scales, which appear to be the dorso-ventrally flattened cirri, retaining their
sensory nature, but adding to this function several others.[321]
The chaetae or bristles are mainly used in locomotion, but it is not unreasonable
to believe that some of the stronger, serrated kinds may be used as weapons of
offence and defence; certainly the Polynoids, bristling as they do with stiff chaetae
along each side, must be rather unpleasant to their smaller enemies.
The various bristles may be placed in three chief groups, viz. (1) simple; (2)
jointed; (3) uncini (see Fig. 138).
(1) The simple chaetae may be smooth and hair-shaped, i.e. "capillary," such as
are present in nearly all families: or they may be forked (Amphinomidae), comb-
shaped (Eunice), notched or serrated, or provided with a series of frills at right
angles to their length, as in Aphroditidae; or fringed along one or both sides with a
membranous expansion, as in Terebellids and Sabellids. The simple chaetae may
also be short and spine-like, as in the ventral bundles of Arenicola; or they may be
slightly curved at the end and notched, forming what are generally termed
"crotchets," such as are common amongst Oligochaeta. These "crotchets" may be
simple, or have numerous denticulations at the end (Maldanidae), or be provided
with a membranous hood (Spioniformia, Capitelliformia). In Hermione peculiar
sheathed, spear-like bristles occur (Fig. 138, N).
(2) Jointed chaetae have already been described (p. 246); they are confined to the
sub-Order Nereidiformia, and occur only in certain families.
(3) The uncini are very short chaetae, which are simply embedded in the skin, and
do not extend beyond the body-wall into the body-cavity. An uncinus is a sharply
curved hook, which may have more or less numerous secondary teeth on it. They
are characteristic of the Sabelliformia and the Terebelliformia.
While the chaetae in the Nereidiformia and others are grouped in bundles, those
of many other families are in vertical, transverse rows, as in Maldanidae and in
Arenicola. The uncini are always embedded in such rows, usually slightly raised
from the general level of the body surface, each being termed a "torus
uncinigerus." These tori are usually limited to the sides of the body, but in Myxicola
and in Notomastus they encroach upon the dorsal surface, and in Chaetozone,
also upon the ventral, so as nearly to encircle the body, recalling the
"perichaetous" condition of some earth-worms.
Gills.—We have already seen that several different organs, e.g. the palps in
Sabelliformia, the prostomial tentacles of Chlorhaemidae, and the notopodial cirri
of sundry other Polychaetes, may take on a respiratory function. There are,
however, certain "gills" developed either on the parapodium itself or elsewhere on
the body which it is difficult to homologise. Such are the retractile gills on the
parapodia of the Glyceridae (Fig. 136, C); those of Dasybranchus, near the
abdominal neuropodia; those of Mastobranchus, near the notopodia. Nephthys
has a sickle-shaped gill on the under surface of the notopodium. The long gill
filaments at the posterior end of Sternaspis, again, are only doubtfully interpreted
as the dorsal cirri of some of the posterior segments.
Since primitively the whole skin of the worm is respiratory, any part of the skin may
become more or less specialised for this function, and chiefly, of course, on the
more actively moving parapodia. The blood-vessels constituting the essential part
of the "gill" may make use of any already existing outgrowth (such as a cirrus or a
tentacle), or may push the body-wall out on their own account.
Internal Anatomy.
Probably those organs which have the greatest effect in modifying the shape of
the body are the septa, for we find in the long, free-swimming worms that these
are regularly present throughout the body, and external "segmentation" of the
body is well marked. In burrowing and tubicolous forms the septa are frequently
incompletely developed, or more or fewer may be absent; and the body becomes
less distinctly segmented externally, tends to vary greatly in diameter during
movement, or becomes plumper. With the disappearance of the septa there is also
a diminution in the number of nephridia, as in Arenicola, with only six pairs.
Further, there is frequently a dimorphism of these organs; instead of all of them
serving equally as excretory organs and as genital ducts, some of the most
anterior in the Sabelliformia and Terebelliformia become greatly enlarged, and
take on practically the whole of the former function; whilst more or fewer of the
posterior nephridia dwindle in size, and become genital ducts. The absence of
septa allows a free communication between the successive segments, and thus a
freer flow of coelomic fluid for the distension of the anterior end of the worm during
burrowing.
In the Scoleciformia and Capitelliformia the buccal region exists, but there are no
jaws. In the Sabelliformia and Terebelliformia eversion does not take place and
jaws are absent.
Amongst the Nereidiformia the jaws are absent in the Phyllodocidae and
Hesionidae; when present they are usually set in the direct course of the food.
There may be one small tooth used for stabbing, as in some Syllids (Fig. 141, A);
or a circle of such denticles (Autolytus, Fig. 140, D). To these are added powerful
grasping jaws in Nereis (E); or the latter may alone be present, as in Glycera (F).
In Polynoë the four jaws are carried by hard pieces, to which the muscles are
attached (C and G). In Nephthys there is a dorsal and a ventral jaw.
In the Eunicidae, however, the numerous denticles are carried in a special pouch
below the food tract, with which it communicates anteriorly.[323] They are arranged
in an upper and lower series. The lower series (L) consists of a pair of flat plates
(k) on each side partially embedded in and acted upon by muscles, with a harder
enamelled piece—the actual lower "tooth" (j)—at its anterior end. The upper series
(U) consists of several pieces, varying in shape and size in the various genera of
this family; but developmentally they result from modifications of two rows of small,
similar pieces.[324]
The intestine is generally straight and cylindrical, and is usually constricted by the
septa, if these are present. In the Polynoids the intervening sacculations become
so long as to receive the name of "caeca," which, in Aphrodite, become
enormously elongated (Fig. 142); there are eighteen pairs of them (c), each being
a slender tube bent upon itself, giving off short branches and dilated distally,
where it lies in the base of the parapodium.
Fig. 141.—A, Alimentary canal of Syllid: B, transverse section of pharynx of the
same; b, buccal region; d, oesophageal outgrowth; g, salivary glands; i,
intestine; j, tooth; p, pharynx; s, gizzard: C, alimentary canal of Petta (after
Wirén); i, intestine; o, oesophagus; r, rectum; s, stomach.
Otocysts are rare. Arenicola possesses a pair at the base of the prostomium, each
of which in some species retains an opening to the exterior.[327] They probably
serve as "organs of direction" rather than of "hearing." Aricia and Polyophthalmus
likewise have such organs on the prostomium; whilst Fabricia, Myxicola, Terebella,
and a few others possess them in the peristomium, or in some other segment of
the body.
Fig. 144.—Ammotrypane aulogaster Rathke, enlarged. (From Cuningham.) Anterior
end. a, Prostomium; b, everted buccal region; c, notopodial cirrus; X, ciliated
organ everted; I, II, III, first three segments.
The eggs and spermatozoa in the Polychaeta are discharged into the sea either
by rupture of the body-wall or through the nephridia; the male and female
elements unite, and the resulting fertilised eggs undergo development, either
floating separately in the water, or embedded in jelly, or attached to the body or to
the tube of the worm.
The little animal is thus equipped for an independent life: the provisional chaetae
help in keeping it balanced; and in some cases (Spionidae) serve to protect the
little soft creature, for when it is touched it curls up, and its chaetae stick out at the
sides, so that it looks like a hairy caterpillar. But the larva is quite at the mercy of
the sea, for it is carried hither and thither by currents, and in this way the species
is disseminated. The larvae of the Polychaetes, like those of other animals, occur
at certain periods of the year in large quantities at the surface of the sea, and
serve as food for various larger animals.
These larvae are at first very different from the adult animal, and the necessary
changes to be passed through are more or less great according to the species. It
is not our intention to describe these changes in detail.[329] The larva increases in
size, the permanent chaetae make their appearance in regular order, and the body
exhibits segmentation, the new segments always appearing just in front of the anal
segment. The internal organs gradually develop, and the prostomial and
parapodial appendages grow out in their turn. In the Sabelliformia the
multifilamentous "gills" arise by the continued branching of an at first simple
process (the palp) arising from the latero-ventral surface of each side of the
preoral lobe.[330] These gradually encroach dorsally and ventrally till the
prostomium is more or less encircled; meanwhile the peristomium grows forwards
so as to conceal the prostomium, which no longer increases at the same rate as
does the rest of the body.
Although most worms appear to discharge their ova directly into the sea and take
no further care of them, some make provision for their offspring either by laying
the eggs in a jelly, which will serve as food for the young larvae—Aricia, Ophelia,
Protula, Phyllodoce—or by attaching them to their body. In certain Polynoids the
eggs are attached by means of a secretion to the back, under the elytra, where
they undergo development up to a certain stage. In Exogone and some other
Syllids they are attached to the ventral cirri, or in Grubea limbata, all over the
back. In the female Autolytus (Sacconereis) a ventrally-placed brood sac is formed
by the hardening of a secretion; the eggs develop into embryos inside the brood
sac, and then become free, with head appendages and three pairs of parapodia.
Enormous numbers of such embryos may occur; for instance, some 300 were
counted in a brood sac of Autolytus ebiensis. In the case of tubicolous worms, the
eggs are frequently attached to the tube, either inside or outside. In Spirorbis and
Salmacina the operculum serves as a brood pouch.
Only a very few species are known to be viviparous, viz. Syllis vivipara Kr.,
Cirratulus chrysoderma Clap., Marphysa sanguinea Mont., and Nereis diversicolor
Müll.
In most genera there is no external difference between a mature worm filled with
generative products and an immature one, except, it may be, in the colour; for the
yolk of the eggs is frequently tinted yellow, or pink, or bluish, while the
spermatozoa in mass are white; so that the normal colouring of the worm may be
modified when filled with these elements. But in a few instances striking
anatomical peculiarities are exhibited by the mature worm.[331] In many species of
Nereis, for instance, those segments containing the generative products undergo
more or less extensive changes, while the anterior ones remain unaltered. The
body of the ripe Nereis is then distinguishable into an anterior non-sexual region
and a posterior sexual region; and so great are these changes in certain species
that the mature worms were for a long time believed to belong to a different genus,
and received the name Heteronereis. But we now know their true relations, thanks
to the work of Claparède and others. The males in the Heteronereid phase have
fewer unaltered anterior segments than the females, so that there is a sexual
dimorphism.
The changes which Nereis undergoes in its transformation affect chiefly (a) the
shape of the parapodia, and (b) the form of the chaetae of these parapodia. Other
organs may also be affected, though less noticeably; thus the eyes become
enlarged, the intestine may become so compressed by the generative products as
to be functionless, and the tail develops special sensory papillae.[332]
In the parapodia an increase in size and a sharper delineation of the various parts
take place; then flattened foliaceous outgrowths (Fig. 147, x, y) arise from certain
lobes of the feet, in which, too, the blood supply becomes greatly increased. The
old chaetae are pushed out by the development of new ones of quite a different
shape; these are jointed like the old ones, but the appendix is, in many species at
least, flattened and oar-shaped (Fig. 123, C, p. 246); and the chaetae are
arranged in a fan-like manner. Both these modifications are in evident relation to
the free-swimming habit which the Heteronereid now adopts. The new foot serves
as a swimming organ, the old one was a walking appendage.
But in other genera the hinder genital region of the body becomes separated, on
maturity, from the anterior non-sexual region. Various stages of this "schizogamy,"
or fission into a sexual and a non-sexual zooid, have been observed in different
genera. In the genus Syllis the first segment of the sexual zooid, after its
separation from the asexual zooid, proceeds to bud forth a head. The character of
the head is alike in both sexes, though different species present heads of different
shapes; and as the worms were originally described as distinct genera, the names
then given are retained as descriptive terms. Thus the "Chaetosyllis" form has only
two tentacles; the "Ioda" form has three tentacles and a pair of palps. One and the
same species (e.g. S. hyalina) may successively pass through these stages.
With regard to the asexual portion, there is a regeneration of the tail segments
after the sexual zooid has separated; and the number of segments so regenerated
is usually equal to those that have become sexual. After a time these newly
formed segments will produce generative organs, and take on the characteristic
natatory chaetae, and this region will in its turn separate.
One original "stock," or asexual zooid, thus produces several sexual zooids, but
these are only of one sex for a given stock. The males differ in several important
characters from the females; so different, indeed, are the two sexes that before
their history was worked out by Agassiz[335] they were placed in different genera.
The male zooid has thus come to be known as Polybostrichus (Fig. 149, B). It has
three tentacles and two bifid palps; there are two pairs of peristomial cirri; the
testes are confined to the four anterior segments, which are without natatory
chaetae. The female is termed Sacconereis, owing to the possession of a great
ventral brood sac; its head possesses no separate palps; the peristomium carries
only one cirrus on each side; ova occur in every segment of the body, and may
even extend into the hinder segments of the asexual zooid (Fig. 149, C).
Fig. 149.—Myrianida fasciata. (From Malaquin.) The bright red markings of the living
animal are here represented black. A, An asexual individual which has
produced by budding from the zone (z) a chain of twenty-nine zooids, the
oldest being labelled 1, the youngest 29. B, A ripe male zooid (Polybostrichus),
with three tentacles and a pair of forked palps (p). There are five unaltered
anterior segments. C, A ripe female zooid (Sacconereis) with the palps fused
with the prostomium; s, the ventral brood pouch projecting on each side; t,
tentacles.