“‘2“ lllll"Hill"IIIIUIUHHIllWIN"lllllllHlHlllHlllHlll

“7 _ 1293 01706 5545

This is to certify that the

dissertation entitled

LIFE HISTORIES AND POPULATION DYNAMICS
OF THREE EARTHWORM SPECIES (OLIGOCHAETA:LUMBRICIDAE)
IN A NORTHERN MICHIGAN HARDWOOD FOREST

presented by

Mark Timothy Thogerson

has been accepted towards fulfillment
of the requirements for

Ph.D . degree in Zoology

 

 

Dat / 9

MSU is an Affirmative Action/Equal Opportunity Institution 0-12771

 

 

LIBRARY

Michigan State
University

 

 

 

PLACE IN RETURN BOX
to remove this checkout from your record.
TO AVOID FINES return on or before date due.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

1/99 c:/CiRC/DateDue.p65-p.14

 

 

 

LIFE HISTORIES AND POPULATION DYNAMICS
OF THREE EARTHWORM SPECIES (OLIGOCHAETAzLUMBRICIDAE)
IN A NORTHERN MICHIGAN HARDWOOD FOREST
By

Mark Timothy Thogerson

A DISSERTATION

Submitted to
Michigan State University
in partial fiilfillment of the requirements
for the degree of

DOCTOR OF PHILOSOPHY

Department of Zoology

1997

0F TH

Life
Holimeiste
transition 11
mOdels are
c0nditions ;
t0 other ear
“'ermicultur

Bot]
year malel
two years in
rapidly, 80m

MOSI Ofthe

ABSTRACT
LIFE HISTORIES AND POPULATION DYNAMICS
OF THREE EARTHWORM SPECIES (OLIGOCHAETAzLUMBRICIDAE)
IN A NORTHERN MICHIGAN HARDWOOD FOREST
By

Mark Timothy Thogerson

Life histories of Dendrobaena octaedra (Savigny), Lumbricus rubellus
Hoflineister and Aporrectodea tuberculata (Eisen) are presented, based on
transition matrix population models and original field observations. The matrix
models are of a new type, being dynamic in nature and using environmental
conditions as driving variables. The models themselves are intended to be adaptable
to other earthworm Species, and may be useful to both soil ecologists and
vermiculturists.

Both D. octaedra and L. rubellus were found to have an approximate three-
year maximum lifespan in northern Michigan, averaging a life cycle approximately
two years in length. Small immatures hatching early in the warm season grow
rapidly, some of them becoming reproductive near the end of their first summer.

Most of the cocoon production takes place during the second year of life, with

 

about 75°
classified

adult size.

uith maxi:
productim
years. witl
being l\'-at
constant 11
adult size.
the maxim

The
6fleets of e
the Operati
Significant
between 01'.
mOchly l6]
Clilellates r

A ne
invertebratt

for so“ moi

about 75% of the cocoons produced during this year. These Species can be
classified as r-adapted, with high juvenile mortality, rapid growth, relatively small
adult size, high cocoon production and a short life cycle.

A. tuberculata grows more slowly, reaching maturity in its second year,
with maximum cocoon production in the third year after hatching. Cocoon
production continues for several years. The maximum lifespan is about seven
years, with an approximate four-year average life cycle. This species tends toward
being K- adapted, with substantially lower juvenile mortality, slower growth, a
constant mortality rate throughout its adult life, lower cocoon production, larger
adult size, and a longer lifespan with a noticeable proportion of individuals living to
the maximum physiological age.

The A. tuberculata model is also used to explore possible population-level
effects of extremely low frequency (ELF) electromagnetic fields associated with
the operation of the United States Navy’s ELF antenna in northern Michigan.
Significant decreases in clitellate earthworm densities were found (p = 0.001)
between observed field populations and predicted model values, given mean
monthly temperature and moisture data; however, higher fecundities of those
clitellates remaining may offset the lower clitellate densities.

A new technique for permanently marking earthworms and other soft-bodied
invertebrates using tattoos is presented, as is a modified and automated technique

for soil moisture determination via time-domain reflectometry.

This digs

This dissertation is dedicated to the memory of my father, Rev. D. D. Thogerson

 

Ec
many peo]
Snider. Tl
guidance 2
me many t
Donald Di
cocoons b

Sex
fieldwork
assisted m

occasions

ACKNOWLEDGEMENTS

Ecological research of this scope is impossible without the assistance of
many people. I wish to thank my graduate committee, Drs. Richard and Renate
Snider, Thomas Burton, Ralph Pax, and Donald Dickmann for their support,
guidance and patience throughout the years of this project. Renate Snider taught
me many things about soil biology and ecology over the years of fieldwork, and
Donald Dickmann made it possible for me to complete my field observations on
cocoons by allowing me to place them in an experimental forest near campus.

Several undergraduates, technicians and high schoolers helped with
fieldwork, but three stand out: Chad Schaedig, Amy Ottoson and Dana F eak
assisted me with collection and processing of my bucket microcosms on many
occasions.

Finally, I wish to acknowledge the love and support of my family and friends
throughout my graduate study. Without the knowledge that they were behind me, I

would not have made it.

LIST OF
LIST OF Z
Chapter I
SYSTEM.
EARTHW

Sys

No
Eat

Ean

TABLE OF CONTENTS

LIST OF TABLES ........................................................ x
LIST OF FIGURES ..................................................... xiii
Chapter 1

SYSTEMATIC S, BIOGEOGRAPHY, BIOLOGY AND ECOLOGY OF
BART HW ORM S ......................................................... l
Systematics ........................................................ 1
The caliginosa Problem ....................................... 4
North American Distribution and Zoogeography ....................... 5
Earthworm Anatomy and Biology .................................... 8
Organ Systems and Function ................................. 10
Digestive System, Intestinal Flora and Enzymes ......... 10
Circulatory System and Respiration ..................... 11
Excretory System .................................... l3
Nervous System ...................................... 13
Reproductive System ................................. 14
Muscular System and Locomotion ..................... 15
Life Cycle ................................................. 16
External Causes of Mortality .......................... 18
Earthworm Ecology ............................................... 20
Guild Classification Systems ................................. 20
Life History Characters ............................... 20
Feeding Ecology and Internal Anatomy ................. 21
Vertical Stratification and Ecological Function .......... 22
General Requirements and Limiting Factors ................... 24
Food ................................................ 25
Soil Moisture and Water Relations ..................... 26
Temperature Ranges .................................. 27
Soil Properties ....................................... 30

Organic Matter and Chemical Nutrient Composition
........................................ 3O
Texture, Porosity, Compaction, Water-holding
Capacity, Clay Content and Other Physical

3 1

Factors .................................

vi

 

MI

Chapter 2
OBJECTl
lnt'

Sp:
Eat

Chapter 3

METHOD:

Site

Fiel.

Coll

pH, Alkalinity and Other Chemical Properties ..... 32

Light ...................................................... 33
Effects of Earthworms on Their Environment .................. 33

Litter Decomposition and Turnover .................... 34

Soil Organic Matter Turnover and Nutrient Cycling ...... 34

Effects on Soil Texture, Porosity, and Water-holding Capacity
.............................................. 36

Modelling of Earthworm Populations ............................... 37
Reichle (1971) - Carbon Flux in a Deciduous Forest ............ 38

Bouché and Kretschmar (1977), Bouché (1980) -- R.E.A.L., a
Descriptive Model of Earthworm POpulation Dynamics in

Agroecosystems ...................................... 38
Lavelle and Meyer (1977, 1982) - Allez les Vers, a Complex

Population Dynamics Model ofMillsonia anomala Omodeo,

Based on Individuals .................................. 38
Martin and Lavelle ( 1992) - an Elaboration of the 1982 Model,
Taking Vertical Distribution into Account ............... 39

Mitchell (1983) - WORM.F OR, a Model of Production, Growth and
Population Dynamics for Eiseniafetida in Sewage Sludge

.................................................... 40

Chapter 2
OBJECTIVES AND RATIONALE ........................................ 41
Introduction ..................................................... 41
Research Objectives .............................................. 43
Species Studied .................................................. 43
Earthworm Population Modelling Approach ......................... 45
Model Construction ........................................ 48
Sources and Use of Data Collected ........................... 49
Incubator Rearing under Controlled Conditions .......... 49
Field Microcosm Rearing under Near-natural Conditions . 50
Periodic Censuses of Natural Populations ............... 50

Building and Validating the Models, and Testing for ELF

Effects ........................................ 50

Chapter 3
METHODS ............................................................ 52
Site Descriptions ................................................. 52
ELF Field Sites ............................................. 52
Field Microcosm Site ....................................... 54
Field POpulation Sampling Methods ................................. 54
Earthworm Censuses ........................................ 54
Environmental Data Collection for Field Populations ........... 5 5
Collection of Earthworms for Rearing Experiments ................... 55

 

 

St

E2

Flt

lnr

Chapter 4
VALIDAT
Tin

ChElpter 5
DEVELor
POPUtau
Ger
C or
C or
The
The
The
Lift

Soil and Litter Preparation for Incubator and Field Microcosm Rearings

........................................................... 56
Physical Characteristics of Prep ared Soil ...................... 58
Earthworm Tattooing Procedure ................................... 60
Anesthesia ................................................. 60
Tattooing Procedure ........................................ 62
Viewing of Tattoos ......................................... 64
Field Microcosm Rearings ......................................... 66
Microcosm Construction .................................... 66
Field Placement of Microcosms .............................. 68
Temperature and Moisture Monitoring ........................ 70
Microcosm sampling ........................................ 72
Incubator Rearings ................................................ 73
Incubator Experimental Design .............................. 73
Response Surface Methodology and Bootstrapping Techniques
Employed ........................................... 76
Model Development .............................................. 81
Determination of F ate Probabilities for Inclusion in Matrices . . . . 83
Model Testing and Validation ...................................... 87
Data Collection and Analysis, Model Building, and Other Computer
Programs Used ............................................. 89
Chapter 4
VALIDATION OF SPECIALIZED TECHNIQUES ......................... 91
Time Domain Reflectometry vs. Gravimetric Moisture Methods ........ 91
Testing the Tattooing Technique .................................... 92
Chapter 5
DEVELOPMENT AND VALIDATION OF EXPERIMENTALLY DERIVED
POPULATION MODELS ................................................ 98
General Growth Pattern ........................................... 98
Comparison of Incubator and Field Microcosm Models .............. 103
Comparison of Composite Models with Pre—ELF Subset .............. 106
The D. octaedra Model ........................................... 110
The L. rubellus Model ............................................ 112
The A. tuberculata Model ........................................ 115
Life Cycle Inferences and Comparisons Using Models ................ 120
Temperature-related cocoon development .................... 120
Phenology of Earthworms after Hatching ..................... 127
Effects of hatching time on survival and development of
worms ....................................... 128
Phenological and life history comparisons between species
132

ooooooooooooooooooooooooooooooooooooooooooooo

 

 

Chapter t
USING 1

Chapter '
SUMMA

Appendix
MULTIPI
MIC ROC

Appendix
MODEL-

Appendix
DATA SU

Chapter 6

USING THE A. tuberculata MODEL TO TEST FOR ELF EFFECTS ........ 143 ‘

Chapter 7

SUMMARY AND CONCLUSIONS ...................................... 149
Summary of Modelling Techniques and Approach ................... 150
Life Cycles and Life Histories of Individual Species .................. 152
Effects of ELF Exposure on A. tuberculata ......................... 155
Directions of Future Research ..................................... 156

Appendix A

MULTIPLE REGRESSION COEFFICIENTS FOR IN CUBATOR, FIELD

MICROCOSM, AND COMBINED MODELS ............................. 158

Appendix B

MODEL- GENERATED MONTHLY POPULATION STRUCTURES ........ 167

Appendix C

DATA SUMMARIES FOR IN CUBATOR AND FIELD MICROCOSM STUDIES

....................................................................... 174

1 82

 

 

Table I.

Table 2.
Table 3.
Table 4.
Table 5_
Table 6.
Table 7,
Table 8,
Table 9,

Table 10.

Table 11.

1

Table 12‘ l

E

Table 1.

Table 2.

Table 3.

Table 4.

Table 5.

Table 6.

Table 7.

Table 8.

Table 9.

Table 10.

Table 11.

Table 12.

LIST OF TABLES

Endemic North American earthworms, adapted from Gates (1972). . . . . 6

Distribution of introduced and native earthworm families, genera, and
species in glaciated and unglaciated areas of North America. ......... 7

Physical parameters of soil from Test and Control sites, and prepared

soil used in microcosm studies. ................................. 53
Percent organics and size distribution of inorganic fraction by mass in
three lots of experimental soil mixture. ........................... 59
Pairs of temperature and moisture levels used in the incubator rearing
experiments. .................................................. 74
Incubator experiment demographic breakdowns for all earthworm
species and stages used to calculate multiple regressions ............ 80
Maximum mass in each size class for the three earthworm species
studied. ....................................................... 82

Lilliefors probabilities for Kolmogorov- Smirnov goodness- of—fit tests on
monthly size increments for three earthworm species. .............. 84

Parameters of the von Bertalanfly growth firnction for three earthworm

species from experimental rearings in field microcosms. ............ 99
ANOVA results from comparison of incubator and field microcosm
models for all three earthworm species. n.s. = not significant at 0.05
level. ........................................................ 105

Number of cocoons and worm transitions used in model construction, by
model type and developmental stage for the entire study. .......... 105

Effects of environmental variables on D. octaedra individuals of various
sizes and developmental stages. ................................ 111

 

Table I3.
Table I4.
Table l5.
Table 16.

Table 17.

Table 18.

Table 19

Table 20,

Table 2 l _

Table 22.

Table 23.
Table 24_
Table 2 5_

Table 26_ i

Table 13.

Table 14.

Table 15.

Table 16.

Table 17.

Table 18.

Table 19.

Table 20.

Table 21.

Table 22.

Table 23.
Table 24.
Table 25.

Table 26.

Efi‘ects of environmental variables on L. rubellus individuals of various
sizes and developmental stages. ................................ 114

Slopes and confidence intervals of regressions of observed on modelled

populations of A. tuberculata after adjustment ................... 119
Effects of environmental variables on A. tuberculata individuals of
various sizes and developmental stages .......................... 119

Parameters for cocoon development equations for three lumbricids,
derived from combined incubator and field microcosm data ........ 124

Regressions of proportion of fertile cocoons on temperature shortly after
cocoon deposition in field microcosms, incubators, and combined
microcosms and incubators for three lumbricid species. ........... 125

Temperature and A horizon soil moisture means for each of thirteen
28-day months in a typical year, employed for phenological analysis
using earthworm models. ...................................... 128

Comparison of cohorts of 10,000 class 1 individuals started at difi‘erent
times (Month 1 = May, Month 6 = late September to mid October) for

three lumbricid species at the end of Year 2. ..................... 130
Modelled maximum cocoons deposited per clitellate in any given month
of each year for three modelled lumbricid populations. ........... 137

Summary of proportions of each earthworm developmental stage
experiencing stage change or mortality during a sampling period, for
three lumbricid species. ....................................... 139

t-tests of model prediction vs. field observations between pre-ELF and
operational periods, for the entire population and for each

developmental stage separately. ................................ 144
Multiple regression coefficients for D. octaedra incubator model. . . 158
Multiple regression coefficients for L. rubellus incubator model. . . . 159

Multiple regression coeflicients for A. tuberculata incubator model. 160

Multiple regression coefficients for D. octaedra field microcosm model.
16 l

ooooooooooooooo
nnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnnn
ooooooooooooo

Table 27.
Table 28.
Table 29.
Table 30.
Table 3 I.
Table 32.
Table 33.
Table 34,
Table 35.
Table 36,
Table 37,

Table 38.
Table 39,
Table 40_
Table 41‘
Table 42.

Table 43_

Table 27.

Table 28.

Table 29.

Table 30.

Table 3 1.

Table 32.

Table 33.

Table 34.

Table 35.

Table 36.

Table 37.

Table 38.
Table 39.
Table 40.
Table 41.
Table 42.

Table 43.

Multiple regression coefficients for L. rubellus field microcosm

model. ....................................................... 162
Multiple regression coefficients for A. tuberculata field microcosm
model. ....................................................... 163

Multiple regression coeflicients for D. octaedra combined incubator and
microcosm model. ............................................ 164

Multiple regression coeflicients for L. rubellus combined incubator and
microcosm model. ............................................ 165

Multiple regression coeflicients for A. tuberculata combined incubator
and microcosm model. 166

OOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOOO

Monthly changes in modelled population structure of a cohort of class 1
D. octaedra, starting on May 1 (day 1, month 1) of a typical year. . . 167

Monthly changes in modelled population structure of a cohort of class 1
L. rubellus, starting on May 1 (day 1, month 1) of a typical year. . . . 169

Monthly changes in modelled population structure of a cohort of class 1
A. tuberculata, starting on May 1 (day 1, month 1) of a typical year. 171

L. rubellus incubator cocoon development summary for each of five
temp eratures. ................................................ 174

D. octaea’ra incubator cocoon development summary for each of five
temperatures. ................................................ 175

A. tuberculata incubator cocoon development summary for each of five

temperatures. ................................................ 175

D. octaedra incubator worm summary by developmental stage. . . . . 17 6
L. rubellus incubator worm summary by developmental stage. ..... 17 7
A. tuberculata incubator worm summary by developmental stage. . . 178
D. octaedra field microcosm summary by date. .................. 17 9
L. rubellus field microcosm summary by date ..................... 180
A. tuberculata field microcosm summary by date ................. 181

Figure I.

Figure 4.

“gm 5.

Figure 6.

Figure 7,

Time 8.

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

Figure 6.

Figure 7.

Figure 8.

LIST OF FIGURES

Generalized lumbricid life cycle. Some longer-lived species with resting
stages alternate between reproductive and non-reproductive states
(dashed arrow); others have only a single extended reproductive
period. ...................................................... 17

Changes in Eiseniafetida cocoon parameters with temperature.
A: Incubation time. Bars indicate range. B: Hatchability and number of
hatchlings/cocoon. Adapted from Tsukamoto and Watanabe (1977). 29

Diagram of the tube apparatus used to determine field capacity of the
prepared soil used in the incubator and field microcosm experiments.
See text for description. ....................................... 61

Diagram of tattooing tool. A -— Minutien Nadel with bent tip;

B -- aluminum tube, crimmd to hold needle in place (can be opened to
allow tip replacement); C -- wooden handle. Total length of tool is
approximately 12 cm. ......................................... 63

Anterior 2/ 3 of sexually mature (clitellate) A. tuberculata, showing
placement of tattoos. T -- tattoos; C -- clitellum; P -- prostomium.
Earthworm illustration by Catherine Nerbonne. ................... 63

Aspirator apparatus used for viewing earthworms (adapted from
Thielemann 1986). A -- Glass tube for examining earthworms;
B -— rubber tubing; C -- jar with rubber stopper; D -- mouthpiece. . . 65

Diagram of bucket microcosms used in the field portion of the study.

A -- Screened lid, Shown here without the TDR moisture probe;

B -- bucket with screened bottom; C -- bucket with bottom removed,
used as a sleeve for easy placement and removal from the ground. . . 67

Layout of the bucket microcosm site. Buckets in each group are 0.5 m
apart on center, with 1.0 m aisles between groups. Codes for
treatments: OCT -- D. octaedra buckets, CON —- control buckets with
no worms (not part of this project), RUB -- L. rubellus buckets,

TUB -- A. tuberculata buckets. ................................. 69

 

 

 

 

Figure 9.

Figure l(

Figure l 1

Figure 12

Figure 13

Figure 14

TTEUIe 15

Figure 16_

Flgm‘e l7

ITigure 18.

Figure 19.

iigure 9.

i‘igure 10.

iigure 11.

iigure 12.

iigure 13.

i‘igure 14.

iigure 15.

 

 

igure 18.

.gure 19.

"igure 16.

igure 17.

Time domain reflectometry (T DR) apparatus used in this study.
A -- coaxial soil moisture probe; B -- Tektronix 1502C TDR meter;
C -- laptOp computer connected to TDR meter via a serial cable. . . . 71

Design of incubator rearing experiments. Numbers in parentheses are
the factor level codings used in the rotatable central composite design.
Moisture is gravimetric moisture. ............................... 78

Hypothetical cumulative normal curve adjusted right to a mean change
of 0. 5. Labelled are the Z- scores which represent growth, no change in
size, and shrinkage. A and B denote the probabilities associated with
shrinkage and [no change + shrinkage], respectively. .............. 85

Representation of the D. ocz‘aedra matrix. Details in text. ......... 88

Regression of gravimetric on volumetric soil moisture in field
microcosms filled with prepared soil. ............................ 93

Survival rate vs. initial mass of tattooed L. rubellus and A. tuberculata
following their first month after introduction to the field microcosms.
Points are actual survival rate; lines are regression lines. See text for
the equations. ................................................ 95

Marking duration distribution of worms of the two species studied.
Removed worms were those removed due to loss of markings during the
experiment, and were used to calculate the mean marking duration time
(numbers next to arrows at top of graphs). The marked and active
worms retained their marks throughout the experiment. ........... 96

Von Bertalanfl’y growth curves for (A) a representative individual, and
(B) the field microcosm population of D. octaedra. Construction
details in text; VBGF parameters for (B) are found in Table 9. . . . . 100

Von Bertalanffy growth curves for (A) a representative individual and
(B) the field microcosm population of L. rubellus. Construction details
in text; VBGF parameters for (b) are found in Table 9. ........... 101

Von Bertalanfi‘y growth curves for (A) a representative individual and
(B) the field microcosm population of A. tuberculata. Construction
details in text; VBGF parameters for (B) are found in Table 9. . . . . 102

Comparison of total size of modelled populations (worms and cocoons)
to actual observed populations of D. octaedra at the CONTROL site
during the preoperational phase of the ELF project. The solid line is a

xiv

 

Figure 21

Figure 22

Figure 23

Figure 24,

Figllre 2 5.

Figure 26,

 

11".)“

f (C —-( jam 1
o

least-squares regression line with slope = 1.0196 and r2 = 0.884;
intersection of this line with the upper right corner would indicate a 1:1
correspondence between modelled and observed p0pulations. The
dashed lines bound the 95% confidence interval about the slope of the

regression line. .............................................. 107

Comparison of total size of modelled populations to actual observed
pOpulations of L. rubellus at the TEST site during the preoperational
phase ofthe ELF project. Slope = 1.0724 and r2 = 0.892 for the
regression line. See Figure 19 for line and significance details. . . . . 108

Comparison of total size of modelled pOpulations to actual observed
populations of A. tuberculata at the TEST site during the
preoperational phase of the ELF project. Slope = 1.0221 and r2 = 0.490
for the regression line. See Figure 19 for line and significance
details. ..................................................... 109

A: 3-dimensiona1 response surface plot of the population growth rate
for D. octaedra from modelled data. B: Contour plot of the same
surface. The shaded portion indicates the temperature and moisture
conditions under which population growth occurs. ............... 113

. A: 3-dimensional response surface plot of the pepulation growth rate

for L. rubellus from modelled data. B: Contour plot of the same
surface. The shaded portion indicates the temperature and moisture
conditions under which population growth occurs. ............... 116

Slopes of observed vs. modelled A. tuberculata populations for each of
the immature size classes individually, from data initially modelled with
one set of equations for all immature size classes. Slope > 1.0 indicates
that the model underestimated the true population; slope < 1.0 shows
that the observed population was overestimated by the model. Pooled
slope of all six classes is shown for comparison. ................. 117

A: 3-dimensional response surface plot of the p0pulation growth rate
for A. tuberculata from modelled data. B: Contour plot of the same
surface. The shaded area indicates the temperature and moisture

5 conditions under which population growth occurs. ............... 121

. Combined field microcosm and incubator cocoon development times vs.

temperature. A: D. octaedra; B: L. rubellus; C: A. tuberculata.
Dashed curves in graphs A, B and C are 95% confidence intervals about
the regression lines. D: comparison between the three regressions;

dashed horizontal line is 365 days. ............................. 123

Figure 2

Figure 2

Figure 2

Figure 3

3 27.

:28.

:29.

:30.

:31.

Fertility rates of the cocoons of three lumbricids with respect to
temperature at time of deposition in field microcosms, incubator
rearings, and combined microcosm and incubator studies. ........ 126

(A) Total cohort survivorship and (B) phenology of selected D.

octaedra stages, based on a modelled cohort of individuals started May
1. .......................................................... 133

(A) Total cohort survivorship and (B) phenology of selected L.
rubellus stages, based on a modelled cohort of 10,000 individuals
started May 1. ............................................... 134

(A) Total cohort survivorship and (B) phenology of selected A.
tuberculata stages, based on a modelled cohort of 10,000 individuals
started May 1. ............................................... 135

Observed vs. modelled p0pulation slopes and 95% confidence intervals
by year for (A) clitellate numbers, (B) total cocoon numbers, and

(C) clitellate fecundity. Dashed lines in (A) and (B) indicate a slope of
1.0, where observed and modelled agree exactly. ................ 145

 

 

 

 

 

 

 

E;
Fossils 3]
the terres
allied mar
650-570 ]
worm, Pr
not know
ancestmt

An
Stries 0“]
metaneph]
A Ventral,
Circlllaton

posterior]!

Chapter 1
SYSTEMATICS, BIOGEOGRAPHY, BIOLOGY AND

ECOLOGY OF EARTHWORMS

Systematics
Earthworms are probably among the oldest of the terrestrial animals.
sils are uncommon, since earthworms are soft-bodied, decompose quickly, and
terrestrial environment is not especially good for fossil preservation. Closely
3d marine polychaetes are known from Australian pre-Cambrian strata some

-570 million years old (Glaessner et al. 1969). An Ordovician fossil segmented

 

Zm, Protoscolex latus (Bather 1920), has been placed in the Oligochaeta. It is
Lurown whether oligochaetes were derived from polychaetes, or had a similar
ester (Lee 1972).
Annelids are segmented worms, divided into segments by septa creating a
:s of hydrostatically isolated compartments, each containing a pair of
tnephridia, paired ganglia, and a number of external setae used for locomotion.
ntral, fused double nerve cord runs the entire length of the animal. The
;1atory system is closed, with a ventral vessel in which the blood flows

eriorly, and a dorsal vessel in which blood flows anteriorly.

 

 

 

The
number of
specialized
take place 1
ova. Oligo
nithstand e
almost emit
Commensal

Som
Ruppen ant
annelids: Pt
groups the l
Meglitsch a]
Bmsca 199(
aseries ofsl
Which Was e
Oligochaeteg
freSTIWaters

There
oligochaetes
APOWCtOda
This latter g,

 

2

The Oligochaeta are set apart from other annelids by the presence of a small
lumber of setae, usually eight (arranged in four pairs) per segment, and absence of
specialized outgrowths of the body wall. Fertilization and embryonic development
cake place within a “cocoon” formed over the clitellum of the parent producing the
)va. Oligochaetes are either terrestrial or aquatic, with only a few species able to
vithstand estuarine or intertidal habitats. Polychaetes, on the other hand, are
IIIHOSt entirely marine, the Hirudinea are strictly aquatic, and Branchiobdellids are
tommensal or parasitic on aquatic invertebrates.

Some confirsion exists about the higher taxonomic groups of the Annelida.
Luppert and Barnes (1994) and Brusca and Brusca (1990) list three classes of
nnelids: Polychaeta, Oligochaeta, and Hirudinida. In contrast, Clark (1978)
roups the last two into a single class, Clitellata, with three subclasses, as do
Ieglitsch and Schram (1991). Cladistic analysis of the major groups (Brusca and
rusca 1990) suggests that polychaetes diverged from the ancestral stock first, and
series of small changes from the basic annelid plan produced a prom-clitellate,
hich was essentially identical to the modern oligochaete. Polychaetes and
igochaetes then developed as sister clades, one in saltwater and the other in
ashwater sediments.

There is similar confusion even within the most intensively studied family of
.gochaetes, the Lumbricidae. For instance, several members of the genus
torrectodea Orley were at times included in another genus, Allolobophora Eisen.
is latter genus became a “catch-all” for many diverse species (Sims 1983),

inly because a type species was not designated. Omodeo (1956) rectified this

 

oversight,
resurrectet
trapezolde
formerly p
count does
synonym o
.iporrectm
Comprising
Aporrector
have differ
original na
this "Specit
Usage (Ens

Alrr
011300haett
Benham, C
vedTOVSky
holotypes a
MacNab an
informatim
appointmer
Reynolds, 1

hithemto

3

oversight, thus forcing a revision of the genus sensu stricto. Gates (1975)
resurrected the genus Aporrectodea, and designated the type species as A.
trapezoides (Dugés). Other species now recognized as part oprorrectodea were
formerly placed in no less than six other genera, some of them now defunct. This
count does not include Nicodrilus Bouché 1972, which is now considered a junior
synonym oprorrectodea (Reynolds 1977a). Within this genus, the “species”
4p0rrect0dea caliginosa (Savigny) is considered by many to be a complex
comprising Aporrectodea trapezoides, Aporrectodea tuberculaz‘a (Eisen),
4p0rrectodea turgida (Eisen), and Aporrectodea nocturna Evans, all of which
rave difi‘erent diagnostic features, habits, and phenologies; nonetheless, the
)riginal name persists, especially in Europe, because the biology and ecology of
his "species" have been widely studied and “caliginosa” has become established by
[sage (Easton 1983, Sims 1983).

Almost all taxonomic literature treating North American terrestrial
'ligochaetes up to the mid-19005 was penned by European workers: Beddard,
Ienham, Cernosvitov, Cognetti, Eisen, Pickford, Rosa, Stephenson, Ude and
’edjovsky (Gates 1982). AS a result, many of the extant North American
olotypes are in European collections. In the 1940's, Gates in the east, and
[acNab and McKey-Fender on the west coast began to publish distributional
formation about terrestrial species in their respective areas. With Gates’
tpointment as a research fellow at Tall Timbers Research Station in Florida, and
Bynolds’ association with the same institution in the 197 0'5, much work was done

'them to characterize the earthworm fauna of the eastern United States and

 

 

Canada. 1
and introd
systematic
publicatio:
their ecolc

NOIIII Aim

The caIigi
Ear
the “Specie
ar.Iélltnents
in Eumpe‘
Cocoon mo
There are at
IWgIda) T
Snider T981
eCOli’gicalr
011 this Spec
refers to “A.
Work is A. 1]
reasonable 3
the species1

(1982): The

 

4

Canada. Reynolds’ primary contribution was to define the distributions of native
and introduced species in the eastern U. S. and Canada. Gates worked on the
systematics of North American oligochaetes during this time, and his final
publication (Gates 1982) is a compendium of the known North American species,
their ecology, biology, and distribution. Little systematics work has been done on

North American endemics since.

The caliginosa Problem
Earthworm workers have, for many years, published investigations including

the “species” Aporrectodea [A llolobOphora] caliginosa. Despite quite convincing
arguments by Reynolds (1977a) and Gates (1982), many workers, primarily those
in Europe, continue to use this designation, although it is obvious by differences in
icocoon morphology, adult size, coloration, external genitals, and behavior that
there are actually four species (A. nocturna, trapezoides, tuberculata, and
turgida). Three of these are sympatric in Upper Michigan forests (Snider and
Snider 1988), lending further credence to the assertion that they are not merely
ecological morphs of one species. Since many ecological studies have been done

on this species group, it is important to define clearly what is meant when one

 

refers to “A. caliginosa”, especially since one of the species being examined in this
:work is A. tuberculata. When the species “A. caligz‘nosa” is referenced here, a

:easonable assumption using evidence presented in the cited literature is made that

   

e species being treated is A. tuberculata sensu Reynolds (1977 a) and Gates

31982). The former species designation is used only to preserve the historical

 

 

 

reference.
currently I

retain orig

The
33 species
families. 11
megadrile
North Am
Pleistocen.
(Reynolds

All
result of"
Post-Pleist
distributior
Center of It
in Origin, a:
Permit then
disperSa] m
these taXa t
appmxhnat

appr0Ximat

 

_II‘

3"—

(I

-——

Q'

T's. “MTG

5

reference. Generic and specific names within the text are spelled according to
currently accepted usage (e.g., “Octolasion” and Vanda”); bibliographic citations

retain original spellings.

North American Distribution and Zoogeography

The currently recognized taxa of native North American earthworms include
83 species in seven genera, all of which are endemic. These are contained in five
families, three of which are only found in North America (Table 1). Seven other
megadrile families (excluding the Enchytraeidae), all introduced, are also found in
North America. All of these families have been only collected south of the
Pleistocene glacial limit, in glacial refirgia, or near large p0pulation centers
(Reynolds 1995).

All other taxa are presumed to have been introduced to North America as a
result of transp ortation by man. Reynolds et al. (1974) proposed his theory of
post-Pleistocene introduction as the only rational way to explain the present
distribution of earthworms in North America. Many species of Lumbricidae, whose
center of radiation is in Europe, and many Megascolecidae, which are Australasian
in origin, are found scattered throughout North America, wherever conditions
ermit them to maintain viable populations. Continental drift, proposed as a
dispersal mechanism by Omodeo (1963) is too slow a process to have introduced
these taxa to North America since the most recent glaciation, which ended

approximately 11,000 years ago, since North America and Europe began to split

approximately 190 million years ago (Smith 1973). There is also no evidence of a

 

 

Table l.

 

Fa
Acanthor

Komarelt

Lumbrici

LUIOdIlIll
Spargano

land bridge
Predominar
there been
e’ilianse an
diherenceg
OTSuch diff
(Gates 1971
The
Pleistocene
islands adjat
1982). on(
all of Canad
Undi:

faunas °Omp

6

 

Table 1. Endemic North American earthworms, adapted from Gates (1972).

 

 

Family Genus Species count Distribution
Acanthodrilidae Argilophilus * 19 Extreme northwest US
Diplocardia * 38 South and central US
Komarekionidae * Komarekiona 1 southwestern
Appalachia
Lumbricidae Bimastos * 9 South-central US
Eisenoides * 2 Southeastern US
Lutodrilidae * Lutodrilus 1 Coastal Louisiana
Sparganophilidae * Sparganophilus 13 Southeastern US

 

* Entire taxon endemic to North America.

land bridge across the north Atlantic since the last glaciation to allow the
predominantly European Lumbricidae to cross (Wright and Frey 1965). Even had
there been such a land bridge, the time necessary for earthworms to cross such an
expanse and colonize American soils would surely have been long enough for
differences to arise between the American and European populations. The absence

of such differences alone is a convincing argument against natural colonization

 

(Gates 1970).

.- The present distributions of all endemic taxa are closely associated with
EPleistocene glacial refugia, either south of the limit of glaciation (Gates 1970), or
:slands adjacent to the Pacific coast of North America (McKey-Fender and Fender
T982). Glaciation extirpated all earthworms from the northern United States and
£111 of Canada, and the native species failed to recolonize (Table 2).

Undisturbed areas in the southern Appalachians tend to support earthworm

aunas composed of a high percentage of endemics, while sites that have been

 

Table 2.
species in

LOCA

\

Glaciate

Capt

 

Nov
Pr.

hiassa.
Rhod
Nonh
30th
Upper M
U"glacial
Dr

M:

Ke

 

 

 

 

Table 2. Distribution of introduced and native earthworm families, genera, and
species in glaciated and unglaciated areas of North America.

 

 

 

INTRODUCED NATIVE
LOCATION FA GEN SPP FAM GEN SPP REFERENCE
M
Glaciated
Cape Breton 1 8 14 Reynolds 197 5a
Ontario 1 8 l7 2 2 21 Reynolds 1977 3
Nova Scotia 1 8 15 Reynolds 197 6
Pr. Edward 1 6 11 Reynolds 197 5b
Island
Massachusetts 2 10 16 1 2 22 Reynolds 1977b
Rho de Island 1 8 13 Reynolds 1973 a
North Dakota 2 5 5 Reynolds 19783
South Dakota 1 3 4 Gates 197 9
Upp er Michigan 1 5 10 Snider 1991
Lower Michigan 1 10 19 2 2 23 Snider 199 1
Ungla cia ted
Delaware 1 6 10 2 3 4 Reynolds 1 973b
Maryland 3 8 14 3 4 8 Reynolds 1 974
Kentucky 1 1 1 3 4 6 Dotson and
Kalisz 1989
Tennessee 2 9 2 3 3 4 l4

 

Reynolds 1977c,
1977d,l978b,
Reynolds et al.
1974

1 One endemic species known only from a botanical garden, the other is limicolous
and restricted to the Great Lakes shoreline.

: Both species known only from arboretums or botanical gardens.
Confined to southernmost tier of counties; largely untouched by Wisconsin

_glaciation.

 

 

 

 

cleared. c
Lumbrtcu
Otto/0310
the native
IMichaels
exotics. u
Ear
earthworm
States and
introductit
earthworm
EllTopean t
It is

North Ame
has Perforn
Caro/1119113,
depths of 1
Climate is n

The,

COHCentric i'

8

cleared, cultivated or otherwise severely disturbed have exotic earthworms, such as
Lumbricus terrestris L., L. castaneus (Savigny), L. rubellus Hoffmeister,
Octolasion tyrtaeum (Savigny), and Pheretz'ma spp. (Kalisz and Dotson 1989). Of
the native taxa, Komarekiona eatom’ Gates and Eisenoides carolinensis
(Michaelsen) seem the most susceptible to disturbance and competition with
exotics, whereas the genus Diplocardia tends to persist.

Early settlers north of the limit of Pleistocene glaciation reported a lack of
earthworms, yet lumbricids are now found widely throughout the northern United
States and Canada (Table 3). Gates (1982) intercepted a variety of potential
introductions from all over the world, demonstrating that it is indeed probable that
earthworms were introduced to North America accidentally subsequent to
European colonization.

It is not known why endemic species have failed to colonize the areas of
North America affected by the Wisconsin glaciation. S.W. James (p ers. comm.)
has performed transplant experiments with native Diplocardz‘a spp. and E.
carolinensis in northern and western Minnesota, where frost annually extends to
depths of 1.5 m. After three years the populations persisted, demonstrating that
climate is not a factor in halting the northward expansion of their range into

previously glaciated areas.

Earthworm Anatomy and Biology
The general body plan of the Annelida is cylindrical, consisting of two

concentric tubes. The outer layer consists of the integument and outer

 

muscular
organs.a
C3YH)’b01
creating f
oppodng
Th

‘ Thect
diagon
hdkm

hydros

‘ The ep
cohage

All out
TOngitu
the lllllt

.

A perit.
The
SPhiflcters 1
°°mpletely
between tw
The
aIIanged in

as a key 0113

9

musculature, and the inner is composed of the alimentary tract and its associated
organs, as well as another muscle layer. These layers are separated by a coelomic
cavity bounded by a peritoneum and divided longitudinally by a series of septa,
creating fluid-filled compartments whose dimensions can be changed by sets of
opposing muscles.
The body wall consists of four layers (Seymour 1978):
0 The cuticle, consisting of laminated layers of collagen fibers, running roughly
diagonal to the long axis of the worm and alternating left- and right-handed
helices in adjacent layers. This gives strength and flexibility to the animal's

hydrostatic skeleton.

0 The epidermis, mainly a supportive layer of columnar cells that produce the
collagen fibers.

0 An outer layer of circular muscle fibers and an interior layer of opposing
longitudinal muscle fibers. The gut is also surrounded by two layers of muscle,
the inner layer of circular and the outer transverse muscle fibers.

9 A peritoneal membrane that defines the inner boundary of the body wall.

The septa, which divide the coelom into segments, have pores with
sphincters that can allow the passage of small amounts of coelomic fluid or
completely isolate the segments. They consist of a layer of connective tissue
between two layers of peritoneal cells (Stephenson 1930).

The typical earthworm has eight setae per segment (sometimes more),

arranged in four pairs. The spacing of these setal pairs around the segment is used

as a key character to distinguish species (Reynolds 1977a).

 

Organ S;
Di

TI
esophagu
intestines
procure a;
gizzard ar.
Th1

which. by
absorption
enlimes tl
Edwards a
litter-feedE
Caligt‘nosa
decompose
down 00111;
organics w(
Alth
eartthrm‘
091mm that
little ofthe]

“Wu and as

10

Organ Systems and Function

Digestive System, Intestinal Flora and Enzymes

The lumbricid digestive system consists of a buccal cavity, pharynx,
esophagus, crop, gizzard, and anterior secretory and posterior absorptive
intestines. The anterior portion, from the buccal cavity through the crop, is used to
procure and store food prior to processing; material is fragmented in the muscular
gizzard and passed on to the intestine.

The intestine is basically a tube with a more or less convoluted typhlosole
which, by nature of its increased surface area, aids in both enzyme secretion and
absorption of nutrients. The anterior portion secretes an acid mucus and various
enzymes that break down proteins, chitin and carbohydrates (Laverack 1963,
Edwards and Fletcher 1988). Cellulase and chitinase are present in the gut of the
litter-feeder Dendrobaena octaedra Savigny, but not in the geophagous Species A.
caliginosa (Nielsen 1962). As epigeic species feed on raw litter and little-
decomposed humus, it seems reasonable that they would have a means of breaking
down complex structural molecules; those feeding in the soil on well-decomposed
organics would have less need for such enzymes.

Although a variety of extracellular enzymes have been found in the
earthworm gut and surrounding tissue, many of these enzymes have specific pH
optima that are not met in the earthworm gut (Laverack 1963). It seems that very
little of the plant tissue and detritus ingested by earthworms is actually broken

down and assimilated; indeed, the digestive processes of earthworms may enhance

 

 

polymeri
1985).

IT
organics
the gut is
water (Le

M
which is s
has been 1
stages ( Se
OTIOXIIIS,
1992). Hi

Individual
Chloragoc)
aIlllnonia a

‘heflephrid

Cirt
The
closed. It c
museumr an

Which dist r11

ll

polymerization of aromatic compounds, resulting in more complex humins (Lee
1985)

The posterior portions of the intestine absorb the low molecular weight
organics resulting from the chemical reactions in the anterior intestine. This part of
the gut is also very important in osmoregulation, as it absorbs a variety of ions and
water (Lee 1985), forming castings which are eliminated via the anus.

Much of the gut tract is surrounded by layers of chloragogenous tissue,
which is similar in function to the vertebrate liver. It is able to store glycogen, and
has been implicated in the ability of certain earthworm species to undergo resting
stages (Semenova 1967). It has also been shown to absorb and sequester a variety
of toxins, such as heavy metals, pesticides, and herbicides (Fischer and Molnar
1992). High levels of certain toxins can deplete the chloragogenous tissue.
Individual cells become detached from the tissue as a whole, and these
chloragocytes float free in the coelomic fluid. Senescent cells autolyze, liberating
ammonia and other waste products into the coelom where they are eliminated via

the nephridia and dorsal pores (Laverack 1963, Edwards and Lofty 1972).

Circulatory System and Respiration

The oligochaete circulatory system, unlike that of most invertebrates, is
closed. It consists of two to five pairs of esophageal vessels which are strongly
muscular and provided with valves and function as hearts, an efferent ventral vessel

which distributes the blood via segmental branches to the somatic vessels,

networks
which pu
B1
onygenai
areas of t
ofsegmer
cells and r
worm uti
caudal reg
Waterlogg
0x:
erYlhrocru
molecule e
much TOWe
“01 most e1
As 1

t1131111 Or 81
The Water h,
ahoVe (Lee
nC for Seve;
Ean]
that “ofmall

IUleIated f0]

12
networks of capillaries in the gut and the body wall, and an afferent dorsal vessel
which pumps the blood forward by peristalsis.

Blood is oxygenated in the subcuticular capillaries and is mixed with non-
oxygenated blood from the gut in the dorsal vessel. Some species have particular
areas of the cuticle modified for gas exchange. In A. caliginosa, the lateral regions
of segments IX-XIII have a thinner than normal cuticle with flattened epithelial

cells and numerous capillaries (Stephenson 1930). An East African glossoscolecid

 

worm utilizes a similar modification together with specialized musculature in the
caudal region to form a roughly conical “lung” which is protruded above the
waterlogged, anoxic soils and sediments which it occupies (Beadle 1957).

Oxygen is carried in both the plasma and in the respiratory pigment
erythrocruorin, analogous to vertebrate hemoglobin. Unlike vertebrates, this
molecule exists free in the plasma instead of in erythrocytes. Erythrocruorin has a
much lower oxygen binding potential than mammalian hemoglobin, and seems to
act most efficiently at low oxygen tensions (Weber 197 8).

As long as the cuticle remains moist, oxygen can be readily absorbed from
the air or soil atmosphere. Oxygen uptake from water is also possible, as long as
the water has sufficient surface area to permit adequate diffusion from the air
above (Lee 1985). Immatures of Aporrectodea turgida can be kept in water at 5-6
°C for several months, although the worms do not grow or mature (pers. obs).

Earthworms can tolerate very high CO2 tensions, substantially higher than
that normally found in the soil atmosphere (Lee 1985). Anaerobiosis can also be

tolerated for short periods, energy being derived from glycogen stores (Weber

 

 

1978). I

may beco

Er

Th
and last.
coelomo:
on the vet
nephridia
outside, '
anteriorp

Nit
C0Piousa;
Stress, Sa
aCivaly TE

IlePhridial

Net
The
Segment, a
muscles, 6
complete I

The firgt f(

13

197 8). This is advantageous especially during periods of heavy rain, when burrows

may become flooded and available oxygen depleted.

Excretory System

The typical earthworm has paired nephridia in each segment, save the first
and last. These are true metanephridia, with the nephrostome Opening into the
coelom of the immediately anterior segment and the bladder opening to the outside
on the ventral surface. In some earthworms (Pheretima 5.1. group), several pairs of
nephridia in the anterior portion of the worm open into the gut rather than to the
outside. This may serve to decrease water loss, and probably changes the pH in the
anterior portion of the gut (Edwards and Lofty 1972, Oglesby 1978, Lee 1972).

Nitrogenous wastes are eliminated primarily as ammonia, diluted in a
copious amount of urine. Some species are able to produce urea when under water
stress. Salts are resorbed as the urine passes through the nephridium, Na+ being

actively removed, while Cl‘ as well as other ions passively diffuse out of the

nephridia] lumen.

Nervous System

The nervous system consists of a ventral nerve cord with ganglia in each
segment, and three pairs of nerve branches per segment which innervate the
muscles, epidermis, gut, and the posterior septum. The first two pairs form nearly
complete nerve rings meeting at the mid-dorsal line (Edwards and Lofty 197 2).

The first four segments deviate from this: segment 111 contains the cerebral

 

ganglion
segment
segment
subphary
nerves or
contains ‘

stimuli ( l

Rt

Th
Prevent St
Parthenog
Observed
P311111 em
0bServatit

Ge
Male: Pa
trallsferrm
TOSIeTTOIh
gono110re.
msegmem

been 110th

l4

ganglion and is innervated from the ganglia of the fourth segment, the second
segment nerves arise fiom the junction of the circumpharyngeal connectives in
segment III, and the first segment is innervated by a pair of nerves arising from the
subpharyngeal connectives in segment II. The pro stomium is innervated by two
nerves originating at the cerebral ganglion in segment III, and its epidermis
contains many sensory organs capable of receiving light, chemical, and tactile

stimuli (Laverack 1963 ).

Reproductive System

The typical earthworm is hermaphroditic and often possesses mechanisms to
prevent self-fertilization, insuring amphimixis (Reynolds 1977a). However,
parthenogenesis, together with reduction of the male reproductive organs has been
observed in some species (Lee 1972). Pseudogamy, in which spermatozoa play no
part in embryonic development other than as a stimulant, is also known from a few
observations (Reynolds 197 7 a).

Generalized sexual organs are as follows:
Male: Paired testes in segments X and XI; seminal vesicles in IX-XII. Sperm
transferred via sperm funnels and sperm ducts to a vas deferens which may extend
posteriorly for several segments before opening to the outside via the male
gonopore. A prostate gland is also generally present. Paired saclike spermathecae
in segments ix and x are present to store transferred sperm. SpermatOphores have

been noted from several species (Lee 1972).

 

 

Female:
ovisacs 11

IT
mucous s
of the sto
and a nun
develops

fimctiona

Ml

Ear
Provide 1h
Interior to
Perform [h
hydrostatit
a Peristalti
diTeCtlon 0
longitudin;
segments 6

TIDchmnm

15

Female: Paired ovaries in XIII. Oocytes travel through the coelomic fluid into
ovisacs which lead into an oviduct to the female genital pore.

The clitellum (often segments XXX-XXXV in the Lumbricidae) produces a
mucous sheath which slips forward, receiving at least one mature oocyte and some
of the stored sperm. As the sheath slips off the prostomium, the ends are sealed
and a nutritive substance fills this cocoon. The oocyte is fertilized and the embryo
develops within the cocoon. Upon eclosion, the immature earthworm is fully

functional.

Muscular System and Locomotion

Each segment has two sets of muscles. The outer circular muscle fibers
provide the contractile force to make each segment longer and smaller in diameter.
Interior to the circular muscles are opposing longitudinal muscle fibers which
perform the opposite function. Each segment acts as a more or less sealed
hydrostatic system in close coordination with adjacent segments. Worms move via
a peristaltic wave of deformation (Dobrolyubov 1986) running retrograde to the
direction of movement. The setae are extended as the segments contract
longitudinally, providing a firm grip within the burrow, and are retracted as the
segments extend. A similar set of muscles surrounds the gut, and acts

synchronously with the muscles of the body wall.

Life Cyc
T1

(1978 ) b1
breeding
exhibited
event per
Sp
ESpecially
1972). for
11315. rem.
550-600 (1
Sliecies f0
i70 days,
(Perrier) a
8hides in 1
Mo
SatChell 19
temperatur
Causing inc
believed to
TUde) and 1

16
Life Cycle

The life cycle of lumbricids is termed semi-continuous by Olive and Clark
(197 8) because they produce several cocoons ("broods") over an extended
breeding season, and may do so for several years. This differs from the polytelism
exhibited by many polychaetes, which generally have only one intensive breeding
event per year.

Specific life cycles for many earthworm species are poorly known,
especially regarding life expectancy. Michon (1954, quoted in Edwards and Lofty
1972), found that Dendrodrilus rubidus (Savigny) became reproductive in 100-140
days, remained clitellate for 200-250 days, with death occurring at approximately
55 0-600 days. Reinecke et al. (1992), investigating the suitability of three epigeic
species for vermicomposting, listed time to maturity for Eiseniafetida (Savigny) as

Z :t:70 days, for Eudrilus eugeniae (Kinberg) as :1:60 days, and for Perionyx excavatus
(Perrier) as 146 days at 25 °C. Maximum life expectancies for the latter two
’ species in outdoor beds was 120 and 90 days, respectively.

Most earthworms mature in roughly one year (Evans and Guild 1948,
Satchell 1967), although environmental conditions such as seasonally low
temperatures or periods of drought may considerably lengthen this period, thereby
causing increased mortality of immatures. Life expectancy for many species is
believed to be 11/2-2 years, although large anecics such as Aporrectodea longa
:Ude) and L. terrestris may live as long as 10 years (Satchell 1967). A generalized

earthworm life cycle diagram is presented in Figure 1.

 

COt

CLIT

Figure 1_ (
altemate he

 

17

C00 ON IMMATURE

/
CLITELLATE ACLITELLATE

, Figure 1. Generalized lumbricid life cycle. Some longer—lived species with resting stages
alternate between reproductive and non-reproductive states (dashed arrow); others have
only a single extended reproductive period.

E
British 11
daily ten
diapause

T
species.
deep bun
(cndogei.
produced
degree of

the risk 0

Ex

Ma
amPhibian
Shrews [0
eanhWOm
(Scozopax
1970), shr.
1980), and
review 0“

Judas (198

earthworm

l8

Evans and Guild (1948) found that cocoon production of several species of
British lumbricids varied throughout the year, peaking with the maximum mean
daily temperature in midsummer, except in species with an obligatory summer
diapause.
There is great variation in the rate of cocoon production by different
species. Satchell (1967), re-analyzing data from Evans and Guild (1948) noted that
deep burrowers (anecics) produced 3 to 13 cocoons per year, topsoil-dwellers
(endogeics) produced 25 to 27 cocoons each year, and litter (epigeic) species
produced 42 to 106 per year. He suggested that these differences are related to the
degree of environmental variation each species is likely to encounter; the greater

the risk of early mortality, the higher the cocoon production.

External Causes of Mortality
Many animals utilize earthworms as prey. Among the vertebrates are
amphibians of all orders, snakes, lizards, birds, and a variety of mammals, from
shrews to bears. In the Upper Peninsula, animals which are known to eat
earthworms include garter snakes (T hamnophis spp.) (Reynolds 1977a), woodcock
(Scolopax minor Gmelin) (Liscinsky 1965, Reynolds 197 7a), moles (Skoczen
1970), shrews (Sorex spp.) (Judas 1989), red foxes (Vulpes vulpes L.) (Macdonald
1980), and robins and blackbirds (T urdus spp.) (Granval and Aliaga 1988). In a
review of vertebrate predators, as well as an exclusion experiment of his own,
Judas (1989) concluded that vertebrates generally do not seriously reduce

earthworm standing stocks, do not affect their vertical distribution, and are not an

 

 

 

 

 

 

 

 

 

importar

certain b

beetles (
litter. Iu
Imortalit
effect “'3
E2
HISIIOSOII
cocoons i
Michigan
Observed
[Immature
arOllnd wl
males 19',
(personal ‘
eanthm
attack 18 115
paTdSIIIZe (

in Europe.

 

19

: important factor in population control, although he did mention studies in which
I certain birds were shown to have a significant effect on earthworm populations.
Invertebrate predators include chilopods (Judas 1989) and large carabid
beetles (Loreau 1988), which feed on earthworms on the soil surface or in the leaf
litter. Judas (1989) found that chilopods are important earthworm predators
(mortality was twice as high in chilopod treatments than in controls), but their
effect was confined to small size classes.

Earthworms and their cocoons are hosts to a variety of parasites.
Histiosoma murchiei Hughes and Jackson, an anoetid mite, is known to infest
cocoons in Denmark ((‘y’elstrup and Hendriksen 1991) and northern lower
Michigan (Oliver 1962). An unknown nematoceran fly larva has also been
observed in cocoons of several earthworm species (personal observation).

ature and adult worms may become infested with monocystid gregarines,
a round which the worms form a fibrous capsule which eventually becomes calcified

iDales 197 8). These white nodules are easily seen through the integument

”fl-1111

personal observation). The cluster fly, Pollem‘a rudis Fabricius, lays its eggs on

arthworms or in moist soil, and the larva parasitizes the mature earthworm. The
ttack is usually fatal (Yahnke and George 1972). At present, it is known to
arasitize only Eisenia rosea (Savigny) in North America, but attacks other species

1 Europe.

Guild C

St
which ca
environrr
character

ecologicr

Li
Sa
1948) ant
history Ch
Worms pr
qUiCkly‘ a
generally
A I
generally 1
and/ or 1in
burrowing
developed
not red‘Pit

dorsal su r1

20

Earthworm Ecology
Guild Classification Systems
Several attempts have been made to group earthworm species into guilds
which can be used in generalizations about how worms interact with their
environment. Various workers have taken different tacks: grouping by life history
characters, feeding ecology and internal anatomy, and by vertical stratification and

ecological function.

Life History Characters

Satchell (1980), summarizing information from Evans and Guild (1947,
1948) and Grafl ( 195 3), deveIOped a system of classification based primarily on life
Listory characters, Sp ecifically, adaptation along the r-K continuum. r-adapted
vorms produce many cocoons, experience high mortality early in life, mature
_uickly, and are often small, whereas K—adapted species produce few cocoons, are
enerally large and mature more slowly, and have a longer life span.

A suite of other characters follow this dichotomy as well. r-worms
enerally do not aestivate, are red-pigmented, respond weakly to light, consume
1d/or live in the litter layer, have a thin cuticle and are not well suited to
rrrowing. K-worms, on the other hand, live in the mineral soil and have well-
:veloped musculature for burrowing; they also possess a thicker cuticle which is
it red-pigmented, although they may be dark brown or gray, particularly on. the

trsal surface.

and/1.!

tending

P
based or
between
and root
digesting

H
epilobic 1
arrangem
ability to
burrowml
Prostomjl
bundled 11
and are w,

De

A‘ tubercn

21

Of the three species treated in this work, D. octaedra is an r-adapted species

and A. tuberculaz‘a is K-adapted. Lumbricus rubellus Hoffmeister is intermediate,

tending toward the former.

Feeding Ecology and Internal Anatomy

Perel’ (197 7 ) divided earthworms into morpho-ecological associations
based on feeding habits, musculature and intestinal morphology. She distinguished
between humus formers, those taxa which feed on largely undecomposed litter
and roots, and humus feeders, species which ingest large amounts of soil,
digesting well-decomposed humus and associated microbes.

Humus formers have a simple typhlo sole and a moniliform gut tract, a closed
epilobic or tanylobic cephalic lobe for grasping food items, and a complex pennate
arrangement of longitudinal muscle fibers, to which Perel’ attributed these worms’
ability to respond quickly to external stimuli, but which are poorly suited for
burrowing. Humus feeders, on the other hand, have a prolobic or epilobic
pro stomium, a highly convoluted typhlosole within a straight tubular gut, and
bundled longitudinal muscle fibers which facilitate strong longitudinal contractions
and are well suited for burrowing.

Dendrobaena octaedra and L. rubellus are considered humus formers, and

A. tuberculata is a humus feeder under this method of guild classification.

 

using a r
role. ant
classifier
consume
anéciqu.
He realiz
and \‘isuz
three bag
SUbgrOUp
Vertical 5
ability to
Stage, ant
a
Straminict
(Tillertem1
(COPTopha
and moistr
temPOrary
00000113 ((

83101161118 1

22

Vertical Stratification and Ecological Function

Bouché (1977) developed and defined the concept of classifying earthworms
using a combination of morphological characters which indicate their ecological
role, and their ”preferred" placement in the litter/ soil horizons. His three basic
classifications were épigées, worms which live above the soil horizons and
consume litter, endogées, those living and feeding within the mineral soil, and
anéciques, which live in the mineral soil and come to the surface to feed on litter.
He realized that each species may embody some characters of all three basic types,
and visualized each species being placed within a triangular region with one of the
three basic types at each vertex. Thus, each of the three basic types is divided into
subgroups which are composites or adapted to specific habitats. He used not only
vertical stratification, but also factors such as presence of “digging muscles”,
ability to keep the cuticle moist, reproductive rates, presence and type of resting
stage, and gut transit time to delineate his groupings.

Epigées typically live in the litter layer of forests. True litter species are
straminicoles, but there are also species which specialize in living in compost or
other temporary surface organic accumulations (détritiphages), mammalian dung
(coprophages), or under tree bark (corticoles). Due to wide ranges of temperature
and moisture conditions at the surface, many of these species exploit rich
temporary organic matter sources, are small, short-lived, and produce many
cocoons (or several worms per cocoon); therefore, they fall into the same class as

Satchell's r—adapted group.

 

 

permane
more or
matrix.
epiendog
specializ
of these
classifiec
A
them actl

Some of

this schel
éPigée, a]
adult, ren

Sa
”Product
humus f0]
[W0 SChen
gUt morph

comprehe]

23

Endoge’es live within the mineral soil, constructing temporary to semi-
permanent horizontal burrows, ofien with several surface openings. They feed on
more or less enriched pockets of organic matter incorporated into the inorganic soil
matrix. Subgroups include hypoendoge'es, living in the deep horizons, and
epiendogées, which live closer to the surface. Members of the latter group which
specialize on dead and senescent plant roots are termed saprorhizophages. Many
of these earthworms are large and long-lived relative to épigées, and can be
classified within Satchell's K-adapted guild.

Anéciques often construct deep, permanent vertical burrows, and some of
them actually pull large leaf fragments down into the soil, where they feed on them.
Some of the largest lumbricids, including L. terrestris, are members of this guild.

The three species studied in this work each fall into different classes under
this scheme: A. tuberculata is a hypoendogée, D. octaedra is a straminicolous
épigée, and L. rubellus switches from an épigée early in life to an epiendogée as an
adult, remaining mostly within the A—horizon.

Satchell's r-K continuum is an interesting way of examining megadrile
reproductive strategies, but is incorporated largely within the Bouché scheme. The
humus former/ humus feeder dichotomy proposed by Perel’ cuts across the other
two schemes, largely because it depends more on functional feeding strategies and
gut morphology. Of the three classifications discussed above, Bouché's is the most

comprehensive, and is used most extensively in this text.

 

Genera

(I5

castings
of their I
sources
IemperaI
Megadri
of suflici
R
activity~ '
examinec
(bottoml;
(Munsell
Palatabili'
Character
Were tem]
(negative;
Soil teXtu
importam
as a grOUp
eaIThWOIn
factms 111a

ImPOI‘tam

24

General Requirements and Limiting Factors

Since earthworms lose moisture readily through their integument, urine and
castings, sufficient substrate moisture is of primary importance. Similarly, because
of their limited capacity for movement, they must live close to suitable food
sources (Lee 1985). All ectotherms are more or less at the mercy of the ambient
temperature regime to provide acceptable temperatures for metabolic activity.
Megadriles have additional requirements including soil texture, pH, and presence
of sufficient quantities of nutrients such as calcium.

Reynolds and Jordan (1975) postulated a conceptual model for megadrile
activity, based on environmental and edaphic characters. Among the characters
examined were landscape slepe and aspect, elevation, physiographic position
(bottomland, terrace, upland), soil pH, soil temperature, soil moisture, soil color
(Munsell notation), soil texture (percent of various fractions), and vegetation
palatability on a subjective scale. Absolute values of Pearson's correlation for all
characters were low (below 0.2), but several were greater than 0.1. Among these
were temperature (positive for aclitellates), soil color characters, percent sand
(negative), percent silt (positive), and palatability (positive for aclitellates only).
Soil texture correlations were the strongest, suggesting that soil composition is an
important regulator of earthworm distribution. Since all earthworms were treated
as a group, one would expect to see low correlation values. Had they divided
earthworms into functional groups, or treated single species, correlations for some
factors may have been substantially higher; nevertheless, their findings identify

11nportant factors determining earthworm activity.

 

commur
forests:
factors;
similar t
which fe

endogei<

F.

F1
of decay
microorg
(Edwards

Rir
able to de
°f3mino 2
than any a
exceed eq]
eanhWOm

mslgnifica]

25

Fragoso and Lavelle (1992) found that species distribution and earthworm
community structure were determined by a hierarchy of variables in tropical rain
forests: temperature was most important, followed by edaphic (nutrient-related)
factors; seasonal effects such as rainfall patterns comprised a third level. Given
similar temperature regimes, nutrient-poor soils favor anecics and epigeics, both of
which feed on surface litter, whereas rich soils of neutral pH favor geophagous

endogeics.

Food
Food for earthworms consists of nonliving organic matter in various stages

of decay and free-living microflora and fauna. Experimental evidence shows that

 

microorganisms, particularly protozoa and fungi, are of major importance
(Edwards and Fletcher 1988).

Richards and Anne (1982) have demonstrated that some earthworms are
able to derive a small portion of their nutrition via transintegumentary absorption
of amino acids, hexoses, and short-chain fatty acids. Uptake is by difiusion rather
than any active transport mechanisms; therefore, concentration ratios cannot
exceed equilibrium. Since absorption rates are low, the contribution to the
earthworms‘ nutritional budget via absorption through the cuticle is probably

insignificant.

 

ionscon
earthwo
flunh.ai
SOHSOIU'
minahna
leuate
capacht'

B.
eaflh\¥0r
humidity
auayexc
System ar
rePlaced i
Wonnsth
that are la

his
eXteIlded ]
linedwith
Metabolic

1955) haw

still be rev

26

Soil Moisture and Water Relations

The cuticle is permeable to water, and selectively permeable to a range of
ions commonly found in the soil environment (Laverack 1963). As a result,
earthworms have only a limited ability to control the osmotic pressure of their body
fluids, and are sensitive to changes in soil moisture and ionic concentrations in the
soil solution. The cuticle, nephridia, calciferous glands, and gut wall all play a part
in maintaining ionic and osmotic equilibrium. Most earthworms are confined to
soil water tensions in the range of pF 2.0 to 4.7, approximately from soil field
capacity to close to the wilting point (Lee 1985).

Because of the need to keep their cuticle moist to aid in respiration,
earthworms tend to lose water to the environment except under conditions of high
humidity. They also lose water via the copious hypotonic urine needed to flush
away excreted ammonia. Total losses through the integument and excretory
system are probably in the range of 10-20% of body weight per day, which must be
replaced if the worm is to survive (Lee 1985). Moisture losses are greater for
worms that are small, thin and active (juveniles and small species) than for those
that are large, thick, and inactive (Piearce 1981).

Many earthworms will enter a quiescent state if they encounter conditions of
extended low water availability. They form a ball, encase themselves in a chamber
lined with mucus and castings, lose much of their body water and slow their
metabolic activity drastically. Several workers (Schmidt 1918, Hall 1922, Grant
1955) have found that many species can lose 70% to 80% of their body water and

still be revived.

 

decrease
than fie]
increasir
norms h

P

producti

fresh anc‘

Tl

Tl
One of In
usually d-
temperati
few studi
p OPUlatio

Ur
iliability 0
re(illireme
(Rigby 19
body, resu

861

27

Aporrectodea longa, exposed to differing soil moistures, showed little
decrease in body water until soil suction was increased to pF 2.78, somewhat less
than field capacity (pF 2). Live weight decreased from that point as a function of
increasing soil suction. At pF > 3.79, somewhat above the wilting point of plants,
worms initiated diapause (Kretschmar and Bruchou 1991).

Production of cocoons also depends on soil moisture. The highest cocoon

production occurs at high soil moisture (30-40% gravimetric). Cocoon mass, both

fresh and dry, also increases with soil moisture (Evans and Guild 1948).

Temperature Ranges

The study of temp erature effects on earthworms has traditionally followed
one of two paths: (1) vital ranges or lethal limits, or (2) temperature preferences,
usually determined by placing earthworms in a long soil-filled trough with a
temperature gradient and allowing them to redistribute. There have also been a
few studies which examined the effects of temperature on reproduction in field
populations (Graff 1953, Satchell 1967).

Upper temperature limits may be either physiologically determined by the
inability of gas exchange across the cuticle to keep pace with increasing metabolic
requirements, or related to breakdown of the collagen fibers in the body wall
(Rigby 1968). Lower limits are probably due to the freezing of fluids within the
body, resulting in cell rupture.

Several factors may confound observations of minimum, maximum, and

optimum temperature ranges in earthworms. Mangum (1978) suggested that there

may be
to differ
metabol
interact
(197m
increasir
synergist
Consum]
uith bot]
23°C ant
Li
eclosion 1
temPeratt
Vflloen e1
(N0rdstré
temperatu
Co
found that
15°C; me;
500 Eire
ofhatchun

28

may be quantitative differences in oxygen consumption between worms acclimated
to different temperatures because of qualitative differences in carbohydrate
metabolism as regulated by neuroendocrine hormones. Soil moisture also seems to
interact with temperature effects. Reinecke (1975) and Nordstrom and Rundgren
(1974) have observed that preferred or optimum temperatures increased with
increasing moisture content. Soil temperature and moisture were found to have a
synergistic effect up on litter consumption rates in immature L. terrestris.
Consumption, therefore assimilation and growth, increased roughly exponentially
with both temperature and moisture until the optimum was reached (approximately
23 °C and -9 kPa), above which it fell rapidly to zero (Daniel 1991).

Life history characters are strongly affected by temperature. Time from
eclosion to mature clitellate worm often decreases markedly with increasing
temperature, as does size at maturity (Frenot 1992, Viljoen and Reinecke 1992,
Viljoen et al. 1992). Growth rate and activity increase with temperature
(Nordstrom 1975). Cocoon production generally is highest in the upper
temperature range for the species (Viljoen and Reinecke 1992, Butt et a1. 1992).

Cocoon development is also affected by temperature. Butt et al. (1992)
found that the optimum incubation temperature for L. terrestris cocoons was
15 °C; mean incubation time at this temperature was 70 days, but it was 275 days at
5 °C. Eiseniafetida cocoons showed significant change in incubation time, number
of hatchlings per cocoon and percent hatchability with temperature (Figure 2).

Time to hatching decreased as a negative exponential with respect to temperature,

29

 

 

 

 

1 20 j
"7’;
90
fl 7 l
LU
E?
l— \,
z 60 4 \
£2 \\\\
f—
< L,
s \
L) 30—1 \\\
Z \}\\J
_L
O l l l m
10 15 20 25
TEMPERATURE (°C)
1 00 7 7 2-4
B o HATCHABILITY
El N 1cocoow
o\°
v 2
='. J O
m 50 O
< 0
it) \
'— Z
254 1-8

 

 

10 1E 20 25
TEMPERATURE (°C)

Figure 2. Changes in Eisenia fetida cocoon parameters with temperature. A: Incubation
tlme. Bars indicate range. B: Hatchability and number of hatchlings/ cocoon. Adapted
from Tsukamoto and Watanabe (1977).

 

and botl
or less li
A
species c
A. 001th
rubidus
can sure
About or
(Holmstr
E-
be Specie
would in.

31.1991)

Sc
Ca
gr mm] at
pOlllllatio
miCIObia]
OrganismE

30

and both hatchability and number of successfiil worms per cocoon decreased more
or less linearly as temperature increased (T sukamoto and Watanabe 197 7).

Although most adult worms cannot tolerate freezing, cocoons of some
species can survive fro st. Among earthworm species found in northern Michigan,
A. caliginosa (tuberculata, turgida, trapezoides), A. longa, and Dena’rodrilus
rubidus cocoons can survive mild frost, circa -1°C. A few of the latter two species
can survive a moderate freeze, ca. -5 °C; as can most cocoons of L. terrestris.
About one third of D. octaedra cocoons can survive a hard freeze at - 10°C
(Holmstrup et al. 1990).

Embryonic development proceeds even at low temperatures, but there may
be species- specific temperature thresholds below which hatching is inhibited. This
would insure that hatchlings find a favorable environment for growth (Holmstrup et

al. 1991).

Soil Properties
Organic Matter and Chemical Nutrient Composition

Carbon and nitrogen are the two nutrients most important to earthworm
growth and survival. Although availability of one or the other occasionally limits
populations, it is usually the C:N ratio which is most important. Animal and
microbial tissue have a C:N ratio of about 5; as this ratio increases in food sources,

organisms experience difficulty in extracting the nitrogen necessary for tissue

production. Nitrogen availability appears to be one of the most important factors

aliecting earthr
lowlLee 1983)
Bouché

and found that

remaining spec:
species \rith op
system showed
endogeés; epigl
undecomposed
examined whicl

optimum. abou

W

133%

Since th
and compositio
fare in a given i
relations and 0

Coarse.
drought) rarely
Conversely, soi
depauperate ea

(Lee 1935).

31

fl‘ecting earthworm distribution, especially in tropical soils where its content is
w (Lee 1983).
Bouché (1972) examined the C:N ratios in the food of 67 French species,
(1 found that optima for 49 species (“eubiotic” forms) were < 13, and the

maining species (“mesobiotic” forms) had food C:N optima 2 13, including two

 

ecies with optima > 17. Comparison of these groups with his guild classification
stem showed that almost all aneciques were eubiotics, as were most of the
dogeés; epigeés, adapted to living in substrates composed primarily of
decomposed plant litter, fell into the latter group. Of the species Bouché

amined which are found in upper Michigan, A. caliginosa has the lowest C:N

 

ptimum, about 11.7, and D. octaedra has the highest at 14.3.

exture, Porosity, Compaction, Water-holding Capacity, Clay Content and Other
hysical Factors

Since the soil is the medium in which earthworms live, its physical texture
rd composition are extremely important factors in how earthworm populations
re in a given location. Ease of movement, availability of suitable food, water
lations and 02/C02 tensions are all affected by the medium’s physical nature.

Coarse-textured soils, due to their abrasive nature and susceptibility to
ought, rarely contain substantial earthworm populations (Lee 1985).
inversely, soils with high clay content in regions of high rainfall also have

oauperate earthworm communities due to extended periods of oxygen deficit

:e 1985).

Soil tYP‘
we, English b
populations Illa
is no significant
earthworm mas
(Zajonc 1972. c
input into the Sr

Soil con‘
not only affects
permeability. F
showed the higl
lowest biomass
with differentia'

in lighter substr

H Alkalinit a
Soil pH i

Chloride conten

concentration (I

Ierrestris and L
thresholds, wlm

earthworms sho

Only 26 0f67 ta

32

Soil type is more closely correlated with earthworm abundance than is litter
:ype. English beechwoods on mull soils support not only larger earthworm
)opulations than beechwoods on mor soils, but more species as well, whereas there
s no significant correlation with litter type (Phillipson et al. 1978). Mean
:arthworm mass and total biomass per unit area, however, are tied to litter type
Zajonc 1972, cited in Phillipson et al. 1978), since leaf litter is a major organic
nput into the soil system and therefore the ultimate food of many earthworms.

Soil compaction is another important factor in earthworm distribution. It
,ot only affects the ease of burrowing, but also water-holding capacity and air
ermeability. Field observations of earthworm communities and populations
howed the highest activity and biomass in uncomp acted arable soils, whereas the
)west biomasses were recorded from wheel-rutted paths. Column experiments
rith differentially compacted soils showed significantly more and longer burrows

[lighter substrates (Stichtig and Larink 1992).

H, Alkalinity and Other Chemical Properties

Soil pH is not generally limiting except in soils with a pH below 4.0.
hloride content is probably a more significant factor, as is calcium ion
)ncentration (Lee 1985). Laverack ( 1961) demonstrated that A. longa, L. .
rrestrz's and L. rubellus would not burrow into soils with pH below their specific
resholds, which ranged from 4.6 to 3.8. Bouché’s (1972) study of French
.rthworms showed that the majority were found in soils with pH from 5.0 to 7.4;

Lly 26 of 67 taxa were found in soils with pH < 4.0, and four were found only in

 

 

 

 

 

 

 

 

 

 

 

soils with pH >
predominate at

Arailabi
(Lee 1985). C3

for proper func

Light

Earthwo
ultraviolet light
more highly pig

to light damage

Effects of Eart

Soil mac
and decomposit
microbial popul
Petersen and Li
have been founr
digeStion of p13:
capacity to dige
interactions bet.

detritus into suc

33

soils with pH > 6.6. Straminicolous species such as D. octaedra and L. rubellus
)redominate at lower pH levels (Nordstrom and Rundgren 1974).

Availability of Ca++ may also be very important to some endogeic species
Lee 1985 ). Calcium carbonate acts as a pH buffering agent, and is also necessary

’or prop er functioning of the digestive system, particularly the calciferous glands.

Light

Earthworms avoid bright light when possible. Short wavelengths,
.ltraviolet light in particular, damage the cuticle and may be lethal. Generally, the
lore highly pigmented species which live in or eat surface litter are less susceptible

3 light damage (Lee 1985).

lffects of Earthworms on Their Environment

Soil macrofauna are generally believed to indirectly affect litter turnover
ud decomposition via comminution of larger debris and stimulation of soil
ricrobial populations (Anderson 1988, Edwards and Fletcher 1988, Lee 1985,
etersen and Luxton 1982); however, moderately large populations of L. terrestris
we been found to affect carbon cycling more directly by assimilation and
.gestion of plant remains (Daniel 1991). Other anecic or epigeic species with the
rpacity to digest plant structural molecules may do the same. Symbiotic
teractions between earthworms and soil microorganisms comminute large

>tritus into successively smaller fragments, eventually incorporating them into

 

 

 

 

and rrercher 1‘

Litter i

Earthwt
derived 501“ 16
1985). Up 10 9
terrestrrs in a f
apple orchards
February (Raw
N into soil on p
earthworms (A-
increasing pastr

Leaf bur
microbial popu'
activity in turn

plants.

Soil Org
Endogei
tunneling. In la
and L. rubellus

estimated that u

34

water- stable aggregates and making their nutrients available to plants (Edwards

and Fletcher 1988).

Litter Decomposition and Turnover

Earthworms are an important factor in degrading and cycling organic matter
derived from leaf litter in North American floodplain forests (Knollenberg et al.
1985). Up to 93% of the annual litterfall in such forests was utilized by L.
terrestris in a four-week period in field microcosms. Populations of this species in
apple orchards buried 2-106 g-ha'l leaf litter between leaf fall and the end of
February (Raw 1962). Likewise, L. rubellus is important in incorporation of litter
N into soil on permanent pasture systems (Syers et a1. 1979). Presence of
earthworms (A. caligz‘nosa) also promotes incorporation of organic matter,
increasing pasture production (Stockdill 1982).

Leaf burial by earthworms is followed by rapid decay, proliferation of
microbial populations, and increased soil buffering (Hartenstein 1986). All this
activity in turn increases the rate at which labile nutrients are made available to

plants.

Soil Organic Matter Turnover and Nutrient Cycling

Endogeic and anecic earthworms consume a large amount of soil while
tunneling. In laboratory rearing studies, Martin (1982) found that A. trapezoides
and L. rubellus consumed 2 - 6 g dry soil g'1 live worm day"". Lavelle et a1. (1989)

estimated that up to 60% of the humic pool in the upper 10 cm of the soil passes

 

 

 

 

through earthv
1.2% organic r.
in the top 20 cr
assuming only
results in accel
Curry and Cott
organic matter
Edwards and L
Several
positively affec
about 5090 oft]
(Parmelee and l
Processes in the
Haimi and BOu<
forests increase
available KCi-e
nutrients availal
b

eech seedlings

nrcreasing Stem
abovegrolmd pc

Stickan 1991 )'

plant glOWth f0]

Pontosco[ex Cor

35

through earthworms every year in African tropical savannas with soils that average
1.2% organic‘matter. Hartenstein (1986) calculated that up to 4% of the organics
in the top 20 cm of soil in temperate areas are utilized by earthworms yearly,
assuming only 6 months activity. This prodigious feeding activity by earthworms
results in accelerated nutrient release and availability (Barley and Jennings 1959,
Curry and Cotton 1983, Vimmerstedt and Finney 1983); mineral cycling and
organic matter decomposition are thus enhanced by earthworms (Bouché 1972,
Edwards and Lofty 1972).

Several studies in both field and laboratory have shown that earthworms
positively affect N cycling and availability to higher plants. Earthworms processed
about 50% of the nitrogen inputs due to plant litter in a Georgia agroecosystem
(Parmelee and Crossley 1988), and enhanced biological activity and decomposition
processes in the humus layer in a coniferous forest soil (Haimi and Einbork 1992).
Haimi and Boucelham (1991) found that presence of L. rubellus in coniferous
forests increases N mineralization and nitrification rates, as well as increasing
available KCl—extractable N and P than did controls, making more of these
nutrients available to plants. Pot experiments with Octolasion lacteum (Orley) and
beech seedlings indicated that N availability was higher in worm-worked soils,
increasing stem production by shifting the transfer of C and N toward the
aboveground portion. It also increased the largezfine root ratio (Wolters and
Stickan 1991). Pashanasi and Lavelle (1992) demonstrated significant increases in
plant growth for two of three tropical fruit tree species when inoculated with

Pontoscolex corethurus (Muller). Enhanced availability of soluble N and P may

 

 

also be detrime
increased 501“
activity may 3C

percolation.

Effects

Large-Sr
or enhancemen
have a profoun
Zealand pasturr
water-holding (
1992). Organic
soil. This in tur
invertebrates at

Earthwo
maintenance of
result in mixing
Hole 1964, Hor
Darwin (1881)
surface of a pas
per year. Anotl‘.
0f the soil is by

many more largt

 

36

also be detrimental in the long run; R.W. Parmelee (pers. comm.) has found that
increased soluble N and P coupled with increased permeability of soils due to worm
activity may actually remove these nutrients from corn agroecosystems by

percolation.

Effects on Soil Texture, Porosity, and Water—holding Capacity

Large- scale introductions of earthworms into areas previously lacking them
or enhancement of dep auperate earthworm faunas have shown that earthworms can
have a profound effect on soil structure. Introduction of lumbricids into New
Zealand pastures was shown to increase porosity, friability, soil moisture, and
water-holding capacity, as well as reduce runoff (Stockdill 1982, Springett et al.
1992). Organic matter, lime, and agrochemicals are also mixed throughout the
soil. This in turn provides a more favorable environment for both other soil
invertebrates and microbial decomposer communities.

Earthworms have also been found to be important in the building and
maintenance of soil structure via their burrowing and casting activities, which
result in mixing the lower mineral layers with the organic surface layer (Nielsen and
Hole 1964, Hoogerkamp et a1 1983, Stewart and Scullion 1988, van Rhee 197 7).
Darwin (1881) estimated that surface-casting lumbricids may bury objects on the
surface of a pasture about 3 cm in 10 years, moving as much as 18 T. soil per acre
per year. Another important way in which earthworms affect the physical structure
of the soil is by formation of water-stable aggregates. Worm-worked soils contain

many more large water- stable aggregates than moist soils simply stirred with a

 

 

 

glass rod (Piea
to the producti
1977, van Rhet

Water rt
Burrowing acti
rates (Smettem
1992). and can
(1992) have de
soluble N03-N
seen after storm
drilosphere con

Earthwc
and carbon allo
increased incor
Density Ofthin
Worm Plots, am

(Van Rhee I977

SeVeral :
soil/litter Syster

e

 

37

glass rod (Piearce 1981). Soils with earthworms are more resistant to erosion due
to the production of these water-stable soil aggregates in earthworm casts (FAO
1977, van Rhee 1977).

Water relations of earthworm-amended soils are often dramatically changed.
Burrowing activity increases hydraulic conductivity, infiltration and percolation
rates (Smettem 1992, Ehlers 1975, German et a1. 1984, Lee 1985, Joschko et a1.
1992), and can have a significant effect on the quality of soil water. Edwards et al.
(1992) have demonstrated that presence of burrows increases the transport of
soluble NO3-N downward through the soil profile. Higher concentrations were
seen after storms which followed prolonged dry periods, suggesting that the
drilo sphere contributes to nitrate infiltration.

Earthworm-induced changes in soil structure also affect plant production
and carbon allocation. Introduction of earthworms into fiuit tree plantations
increased incorporation of organic matter, aggregate stability and air permeability.
Density of thin roots as well as the ratio between thinrthick roots increased in the
worm plots, and fruit production was 2.5% higher in worm plots than in controls

(van Rhee 1977).

Modelling of Earthworm Populations
Several attempts to model earthworm populations and their effects on the
soil/litter system have been made in the last 20 years. Some of the more notable

efforts are discussed below.

 

 

 

 

 

Reichle (1971}

Modelli
study; its aim v
by modelling tl
one of the com

rates of the ear

Bouché and K
llodel of Eart

The “EC
conceptual mot
casting and but
agroecosystem

This mo
agroecosystem
model, it serve
not an end in it

directly test the

LaVelle and M
l)ynamics M0!

COHSifucted Wi]

 

 

38

Reichle (1971) - Carbon Flux in a Deciduous Forest

Modelling of earthworm populations was not a primary concern of this
study; its aim was to describe the ecological energetics within a forest ecosystem
by modelling the flow of carbon through various comp artmentsf Earthworms were

one of the compartments; the model predicts ingestion, egestion, and respiration

 

rates of the earthworm community as a whole.

Bouché and Kretschmar (1977), Bouché (1980) -- R.E.A.L., a Descriptive

 

Model of Earthworm Population Dynamics in Agroecosystems
The “Ecological and agronomic role of Lumbricidae” is a compartmental

conceptual model diagramming carbon and nitrogen flow, microbial activity,

 

casting and burrowing activity, and attempts to explain the role of earthworms in
agroecosystems.

This model delineates the effects of lumbricid communities on
agroecosystems, by tracing nutrient flow through the system. As a descriptive
model, it serves as a good framework with which to direct further research, but is
not an end in itself. Being qualitative rather than quantitative, it cannot be used to

directly test the effects of changes in one compartment on another.

Lavelle and Meyer (1977, 1982) - Allez les Vers, a Complex Population
Dynamics Model of Millsonia anomala Omodeo, Based on Individuals
This population dynamics model of Millsom'a anomala seems to have been

constructed with extensibility to other species and environments in mind. It is

quite complex
burrowing beh
the substrate if
mortality are lr‘
increases or dt
but the enviror
(much drier) fr
satisfactorily. 4
within the bou‘

The pri:
even though se
eIIVirOnmental
abOVe model, \

iilnitations as t

Martin and L;
Vertical Distrl

DRILO'
drilosphere (thr
submodel deSm
conditions Sucl

as individual be

39

quite complex, taking into account weight classes within developmental stages,
burrowing behavior with respect to changing soil moisture profiles, food quality of
the substrate in various soil horizons and the litter layer, and reproduction and
mortality are keyed to the number of days within each period an individual
increases or decreases in mass. The model was tested with an independent data set,
but the environmental conditions in the second set were substantially different
(much drier) from those of the first set. Even though the model performed
satisfactorily, one should use validation data values which are found substantially
within the bounds of the set used to build the model.

The primary drawbacks of the model are (1) cocoon incubation time is fixed,
even though several studies have shown that it is often dependent up on
environmental factors, particularly temperature; and (2) this model, just like the
above model, works with discrete “individuals”, and is subject to the same

limitations as that of Mitchell (1983) discussed below.

Martin and Lavelle (1992) - an Elaboration of the 1982 Model, Taking
Vertical Distribution into Account

DRILOTROP (in FORTRAN) is a model to simulate the functioning of the
drilosphere (the soil surrounding earthworm burrows) in a tropical savanna. A
submodel describes the vertical movements of M. anomala due to environmental
conditions such as depth-specific temperature and moisture, and biotic factors such

as individual behavior and depth-specific organic content (food quality).

 

 

 

Mitchell (191
Population D
This is

of temperaturt
conversion for
from Dr. Mite?
be a tight mod
extensively rer
adaptable for r
for which it w;
Other p

it lumps a]
A large imr
aclitellate
it assumes
rates. In It.
f0od qualit

gTOW'th ratl

used to cre

ll] diViduals
eIlViIollmer
Standpoint,
many Weig]

40

Mitchell (1983) - WORM.FOR, a Model of Production, Growth and
Population Dynamics for Eiseniafetida in Sewage Sludge

This is a predictive model, the purpose of which is to investigate the effects
of temperature and food type on p0pulation dynamics, biomass change, and waste

conversion for E. fetida. I have obtained the program source code (in FORTRAN)

 

from Dr. Mitchell, and translated it to Turbo Pascal for examination. It seems to
be a tight model for E. fetida under controlled conditions, but it would have to be
extensively reworked for use under changing natural conditions. It may not be

adaptable for use with another earthworm species -- it is quite specific to the task

 

for which it was designed.

Other possible barriers to adaptation to other species and systems include:

 

0 It lumps all developmental stages, dividing individuals into mass classes alone.
A large immature very likely behaves differently from an adult of similar mass --
a clitellate allocates more of its energy to reproduction rather than growth.

a It assumes a maximum mass, using a logistic equation to determine growth
rates. In nature, environmental conditions such as temperature, moisture, and
food quality may in part determine maximum mass and change parameters in
growth rate equations, both of which are not considered in this model.

0 Although it is a predictive model, validation with data sets independent of those
used to create the model was not performed.

' Individuals are treated separately, even though the same growth equations and
environmental state transitions affect all animals equally. From a computing
standpoint, this approach uses extra computer memory and storage, especially if
many weight classes are used, and takes a lot of computer time.

 

Detaile
species that ar
animals are im
extensively. or
cocoon develc
reproduction.
the field; only
(Edwards and
multiple envir.
and availabilit;

Many 3
various biocid
these substanc
Perturbation p
Currents 0r fie

resplratory SUI

 

Chapter 2

OBJECTIVES AND RATIONALE

Introduction

Detailed dynamics of natural earthworm populations, particularly those of
species that are not economically important, are poorly understood, although these
animals are important members of the soil community. Even for species studied
extensively, only selected life history parameters have been observed, such as
cocoon development time (usually at a single temperature), mean time to first
reproduction, or fecundity. Little is known about life spans or life expectancies in
the field; only a few notes have been made of life spans in laboratory situations
(Edwards and Lofty 1972). Very little is known about interactions between
multiple environmental variables, such as temperature, moisture, and food quality
and availability.

Many studies have been done on the sensitivity of certain earthworms to
various biocides and heavy metals, and on their ability to bioaccumulate some of
these substances, thus passing them up the food chain. One environmental
perturbation previously left unexamined is their possible sensitivity to small electric
currents or fields within the soil. Earthworms must maintain a moist cuticle as a

respiratory surface, rendering their bodies conductive. Because their nerve fibers

41

 

 

 

are not inSlllat
induced in the
an electric out
toward the C at
electric curren
sampling deyir

The U.!
Michigan‘s UP
in frequency It
held contacts :
polarity of the
producing an a
so that the gro
which is rapidl

One of
was designed t
earthworms (S
this ten-year ii
to the antenna
was activated;
Population- or

field induced b

42

are not insulated by a myelin sheath, they may be susceptible to electric fields
induced in the soil. Edwards and Lofty (197 2) describe earthworms’ responses to
an electric current: in water, they become U-shaped, with both ends pointing
toward the cathode. Many earthworms come up to the soil surface when an

electric current is applied, and this has been used as the basis for earthworm

 

sampling devices (Rushton and Luff 1984).

The U. S. Navy’s Extremely Low Frequency (ELF) radio antenna in
Michigan’s Upper Peninsula generates an alternating electromagnetic field similar
in frequency to household line frequency (76 Hz nominal). Whenever a magnetic
field contacts a conductive medium, it induces a current in the conductor. As the

polarity of the magnetic field changes, the induced current also changes direction,

 

producing an alternating current. Soil, especially if moist, can conduct electricity,
so that the ground near the antenna is subjected to a weak alternating current,
which is rapidly dissipated with increasing distance from the source.

One of the elements of the ELF ecological monitoring program in Michigan
was designed to look for just such a localized EM effect on soil fauna, particularly
earthworms (Snider and Snider 1987, 1988; Snider 1994). The first five years of
this ten-year field study were used to obtain baseline data in a TEST site adjacent
to the antenna and a CONTROL site removed from its influence before the antenna
was activated; the second five years, the study areas were monitored for any
population- or community-level changes that may have occurred due to a weak EM

field induced by ELF antenna operation.

 

This st‘

( 1) To dev
lumbrir

12) To des'
differer
along I

13) To use
of indu
phaset
Result:

this thesis; C h

during the opt

ELF EM field

fate of the pol

Three f
Widely differe
antenna; APO
Lumbricus tn
the r-K cOntin

dporre
occasionally c

Individuals C0

43

Research Objectives
This study had three objectives:

( 1) To develop dynamic structured population models of three species of
lumbricids based on soil temperature and moisture regimes.

(2) To describe the life cycles of the three species and compare their success in
different environmental regimes, ranking their responses and adaptations
along the r-K continuum.

(3) To use this modelling approach to detect and describe any significant effects
of induced EM fields on earthworm populations during the operational
phase of the ELF project.

Results pertaining to the first two objectives are presented in Chapter 5 of
this thesis; Chapter 6 details the application of one of the models to ELF data
during the operational period to determine if and in which developmental stages the

ELF EM field had an effect, and if so, what consequences it may have had for the

fate of the population.

Species Studied

Three lumbricid species were chosen for examination because of their
widely different lifestyles and their high abundance in the study area near the ELF
antenna: Aporrectodea tuberculata (Eisen), Dendrobaena octaedra (Savigny), and
Lumbricus rubellus Hofimeister. They can be placed in three distinct sites along
the r-K continuum (MacArthur and Wilson 1967 ).

Aporrectodea tuberculata was common at the ELF TEST site and
occasionally collected at the CONTROL site. Twenty-five to 30% of the

individuals collected on any given date were large adults (either clitellate or

 

 

 

 

aclitellate) wi'
when compare
than those of '
population str
species tends

Dendn
at maturity. 1'
much less con
June and early
number of cor
and they were
adapted specj

Lumbr
CONTROL. '
expansion at (
Was more gket
becoming 11101
immfltures Wa
Was nearly as
structure tOge

continuum sor

44

aclitellate) with masses up to about 1.0 g, and cocoonzclitellate ratios were low
when compared to other species found in the ELF sites. Cocoons were also larger
than those of most other species, averaging about 20 mg. The seasonally stable
population structure, comparative size and frequency of cocoons suggests that this
species tends toward K-adaptation.

Dendrobaena octaedra is a small species, rarely weighing more than 0.15 g
at maturity. It was the most commonly found species at the CONTROL site, and
much less common at TEST. Small immatures dominated the population during
June and early July, shifting to large immatures and adults later in the summer. The
number of cocoons found per clitellate was much greater than in A. tuberculata,
and they were much smaller, averaging 3 to 4 mg. These factors indicated an r-
adapted species with high juvenile mortality compensated by high fecundity.

Lumbricus rubellus was common at the TEST site, but uncommon at
CONTROL. There were indications that it was in the early stages of population
expansion at CONTROL (R.M. Snider, pers. com.). The population structure
was more skewed toward immatures than in the other two species, with clitellates
becoming more common in the late summer and fall. No marked increase of
immatures was evident at any time, unlike D. octaedra. The cocoon:clitellate ratio
was nearly as high as in D. octaedra. The more seasonally stable population
structure together with high cocoon production suggests a position on the r-K

continuum somewhere between the other two species.

 

 

Both 5
and developm
but the combi
for many spec
between deve
lose mass and
population dy

Matrix
suited to this 1
growth and re
also sensitive
developmenta
in many biolop
behavior ofa 3
Change, the be
anew matrix
likely not an a
at the Same pl:
work on J a ck-

between geog]

45
Earthworm Population Modelling Approach

Both soil temperature and moisture affect the cocoon production, growth
and developmental rates of earthworms (Laverack 1963, Edwards and Lofty 1972),
but the combined effect of these environmental variables has not been quantified
for many species. These effects will differ not only between species, but also
between developmental stages within a population. Since earthworms can easily
lose mass and may regress developmentally when stressed, a predictive model of
population dynamics can quickly become complex.

Matrix projection models, such as those of Leslie ( 1945, 1948), are well
suited to this type of modelling, as the matrix can be so arranged as to allow both
growth and retrogression, and differential fecundity based on age or size. They are
also sensitive to small changes in the behavior of individual age, size or
deve10pmental stage classes. The major disadvantage of this type of model as used
in many biological or ecological studies is that it is static, and describes the
behavior of a given population under specific conditions. If those conditions
change, the behavior of the population will also change, requiring the generation of
a new matrix. This means that a matrix produced from one year’s data at one site is
likely not an accurate predictor of the behavior of a population elsewhere, or even
at the same place under different conditions. An example is Bierzychudek’s (1982)
work on J ack-in-the-pulpit: data collected during different seasons and within and
between geographically separated populations produced matrices that indicated

population increases in some instances, and declines in others.

 

One w
to choose. dej
its own problr
one encounte'
exists? A mo
size and deve
in a matrix. 5
change in pop

The st
Leslie matrix
This is not a ‘
transition pro
Probabilities

The ai
Meyer 1977 ;
fIamework C,
pOSSibly for r
the two mOde
having a Sepa
states of 3 Su
environlIlelltz
memberS Ofa

cooeons. All

46

One way around this problem is to produce a stack of matrices from which
to choose, depending upon the combination of conditions. This solution produces
its own problems. How does one choose between several valid matrices? What if
one encounters a particular combination of conditions for which no current matrix
exists? A more elegant, but complex, way is to calculate predicted mean changes in
size and developmental stage due to environmental influences, and place the results
in a matrix. Standard matrix algebra can then be used to calculate an expected
change in population size and structure, just as with the typical matrix model.

The structure of the model used in this work is more complex than a typical
Leslie matrix because changing environmental variables are taken into account.
This is not a “static” transition matrix -- a single matrix containing immutable
transition probabilities. Instead, it is a “dynamic” matrix, one whose transition
probabilities change with environmental conditions.

The aims of this model design are similar to those of Lavelle’s (Lavelle and
Meyer 197 7 and 1982, Lavelle et a1. 1989), where the production of a generic
framework can be utilized for a variety of species under different conditions,
possibly for monitoring effects of environmental changes. The difference between
the two models is one of approach -- probabilistic versus deterministic. Instead of
having a separate case for each individual, which winds its way through different
states of a suite of environmental and biotic variables, a set of equations based on
environmental conditions was used to determine probabilities of state change for all
members of a given group of similar size and/or developmental stage, including

cocoons. All that changes in consecutive periods is the number of individuals in

 

 

 

each class -- 1
separately for
and as a resul
species and or
be tested in t1
Advan

° The abilit
ages. devr

model as.

' The possi
existing p

‘ A way of
environm

The b:
histories of n
POPUiation g:
iw°01house.
aI1d forest m;
melilods hay.
these are dyn
occuflying th

may be unfar

47

each class -- parameters for each individual do not have to be changed and stored
separately for the next iteration. Therefore, this model is more compact and faster,
and as a result should be easier to test. It should also be more adaptable to other

species and conditions; in fact, it was designed with this feature in mind, and will

be tested in this work with multiple species.
Advantages of the transition matrix approach are:

0 The ability to easily change parameters for response of particular classes (sizes,
ages, developmental stages) by a small amount to test the robustness of the

model as a whole to small changes.

0 The possibility of testing the success of an introduction or the fate of an
existing population, given a regime of environmental conditions.

0 A way of pinpointing the classes that are most affected by a given
environmental change.

The basic method has been used in a variety of applications, from life
histories of individual species (Bierzychudek 1982, Crouse et al. 1987) and
population growth projections to prediction of multispecies interactions
(Woolhouse and Harmsen 1989), and also wildlife (Rosenberg and Doyle 1986)

and forest management (Pakkala and Kolstrom 1988). As a result, a number of
methods have been deve10ped to test the validity of any model produced. Because
these are dynamic matrix models, with equations rather than discrete numbers

occupying the cells, statistical methods used to test their validity and sensitivity

may be unfamiliar.

 

Model Const
A first
size. or develt
individuals int
within those S
Develc
embryonic sta
Immature wr
individuals ha
developed clit
Presence of a
WOrIn denotes
00nsidered wl
mirror size~sp
experiments.
Model
and treated ill
(I) l)etern
stage c
change
determ
Stage c
SDread
deviati
muhm;

PTObab

48

Model Construction
A first step to any transition matrix model is the appropriate choice of age,

size, or developmental stages. A dual system was chosen, first separating

individuals into groups by developmental stage, then into smaller size classes

within those stages.
Development was separated into four broad stages. The cocoon is the

embryonic stage of the earthworm enclosed in a mucopolysaccharide capsule.
Immature worms are those that have no external sexual characters. Aclitellate

individuals have obvious tubercula pubertatis and genital papillae, but no well-

develop ed clitellum; these are prereproductive or nonreproductive adults.

Presence of a glandular clitellum that is often a lighter color than the rest of the
worm denotes the reproductive, or clitellate condition. Only this last stage is
considered when calculating fecundity. Size classes within stages were chosen to

mirror size-specific growth rates observed during the field and laboratory

experiments.
Model construction was divided into three major sections, summarized here

and treated in detail in the next chapter:

(1 ) Determination of transition probabilities for each size class/developmental
stage combination using pairs of equations that define the mean size/stage
change and the spread of values about the mean, and an equation that
determined the probability of survivorship. It was assumed that size or
stage change within a population emulated a normal distribution, so the

spread was taken to be the standard deviation. The mean and standard
deviation in each case were estimated via bootstrapping (incubator runs) or
multiple regression of actual data (field microcosms). Size class transition
probabilities were apportioned into three possibilities using the calculated
size and spread: the current size class, and the size classes immediately

 

 

above
detern

12) Calcul
fertilit
rate fo

1” Comb
wager
dheht
probal

Ponm
popuh

Soumesand

I)ata 1
sources. ()n<
Conditions, a
exllOSed to nt
pandecmw

Variables

lncuh
RBpht
C(auditionS at
individuals, \
individual m;

49

above and below. Developmental stage change probabilities were
determined in a similar manner.

(2) Calculation of fecundity equations (cocoons produced per clitellate),
fertility (percentage of viable cocoons), survivorship, and developmental

rate for cocoons.

( 3) Combination of matrices that modelled growth within developmental stages,
stage change, and survivorship, then addition of fecundity estimates for
clitellates to produce a single master matrix containing the transition
probabilities for that particular set of environmental conditions.
Postmultiplying by a population vector would then produce a projected
population structure, just as one would use a "static" transition matrix.

Sources and Use of Data Collected

Data used in constructing and testing the models were collected from three
sources. One was carried out in incubators with controlled environmental
conditions, a second was a “natural experiment” in which captive populations were
exposed to near-natural conditions that were closely monitored, and the third was a

periodic census of natural populations and monitoring of natural environmental

variables.

Incubator Rearing under Controlled Conditions
Replicate subpopulations were reared in incubators under constant
conditions at several levels of both temperature and soil moisture. A few
individuals, widely spaced in size, constituted each subpopulation, allowing
individual masses and developmental stages to be tracked. See Chapter 3 for

details of the methods used in processing and analysis.

 

 

 

 

 

 

Field 1‘

Replic:
conditions wit
microcosms it
intended to be
individuals in
facilitate trac

and 4.

Periot
Biwee
taken of natu
data were col

earlier.

Build
Data ‘
[0 develop [y
and tested ag
Similar. Aft
SOUrces We”
A Sim

Possible dug

50

Field Microcosm Rearing under Near-natural Conditions

Replicate populations reared in microcosms under semi-natural field
conditions with closely monitored temperature and moisture regimes. Since these
microcosms were substantially larger than the incubator microcosms and were
intended to be maintained for a longer time with more worms per population,
individuals in these microcosms were permanently marked using tattoos to
facilitate tracking of individuals. Pertinent techniques are outlined in Chapters 3

and 4.

Periodic Censuses of Natural Populations

Biweekly censuses (May through October) spanning a 10-year period were
taken of natural populations in the field, with associated environmental data. These
data were collected as part of the ELF ecological monitoring project discussed.

earlier.

Building and Validating the Models, and Testing for ELF Effects

Data from the incubator and field microcosm experiments were employed
to develop the models. Separate models were constructed from the two sources,
and tested against each other to determine whether they were sufficiently
similar. After similarity between models was established, data from the two
sources were combined, and a composite model for each species was built.

A similar strategy for validating the models using pre-ELF data was not

possible due to the nature of the field data. It was a point-sampling population

 

census rather
method was e
vector four w
throughout th

Testin,
tuberculara p
process. by C1
determined bj
year. and mm
allowed isola

overall succe

51

census rather than an extended record of individuals. As a result, a different
method was employed, that of using a single date’s population vector to project a
vector four weeks later, then comparing the modelled with the actual vectors
throughout the season.

Testing for potential effects of ELF electromagnetic fields on the A.
tuberculata population was accomplished in the same manner as the validation
process, by comparing projected population vectors against actual vectors
determined by periodic censuses. These comparisons were analyzed by month,
year, and entire ELF operational period, 1989 through 1993. The model structure
allowed isolation of key life cycle stages to determine their importance in the

overall success of the population.

 

 

ELF Field Si

The si‘
in Snider and
approximatel
CONTROL sf
Chosen for th.
lOCiited in nor
basswood. an
SlliCebush (L,
coIlrrnon shru
iiOor. The alt
Sites tended S.
occasional 00
TEST and Cc

chapter) 11 Se (1

Chapter 3

METHODS

Site Descriptions

ELF Field Sites

The sites used in the ELF ecological monitoring project are described fully
in Snider and Snider (1987). Two sites, a TEST site (T.44N, R.29W, sec.25)
approximately 80 m from the north-south overhead element of the antenna, and a
CONTROL site (T.44N, R.30W, sec. 11) 11.5 km from the same element, were
chosen for their similarities in soil type, forest cover, and elevation. Both were
located in northern deciduous forests with approximately 80% sugar maple, 10%
basswood, and 10% other deciduous trees composing the canopy and subcanopy.
Spicebush (Lindera benzoin L.) and leatherwood (Dirca palustris L.) were
common shrubs, and various grasses and spring flora sparsely covered the forest
floor. The altitude of both sites was approximately 420 m. The A horizon in both
sites tended strongly toward mull and was developed on sandy glacial till with
occasional cobbles and small boulders. Table 3 shows the physical makeup of the
TEST and CONTROL soils compared with the prepared soil (described later in this

chapter) used in both field and laboratory microcosm studies.

52

 

 

 

 

 

 

Table 3. Phy
used in micrt

Parametr
% Organic
°b Sand
0'0 Silt
0", Clay
Texture
Horizon de
pH

\

M

Awe

Control site.
S‘UdY- Mear
throughout 1
through earl
260C in Jul)
range from .

Each
betWeen the
site comm“

te“literature

(1987),

53

 

Table 3. Physical parameters of soil from Test and Control sites, and prepared soil
used in microcosm studies.

 

 

 

TEST Site CONTROL Site
A B A B
Parameter Horizon Horizon Horizon Horizon Microcosm

% Organic 9.6 2.7 9.3 2.0 5.7
% Sand 59.7 59.8 58.6 58.7 65.3
% Silt 23.3 22.6 24.9 23.2 19.2
%Clay 17.0 17.6 16.4 18.9 15.5
Texture sandy sandy sandy sandy sandy

loam loam loam loam loam
Horizon depth 8 - 15 cm > 75 cm 5 - 15 cm > 55 cm ---
pH 5.9 5.9 5.8 5.8 6.2

 

 

 

 

Sources for TEST and CONTROL site data: Snider and Snider (1987).

A weather station in Iron Mountain, approximately 30 km south of the
Control site, provided mean annual weather data for the 30 years preceding the
study. Mean annual precipitation is 768 mm, more or less evenly distributed
throughout the year. Snowfall and snow cover generally occur from late October
through early May. The mean annual air temperature is 54°C, with a mean high of
26°C in July and a mean low of -15 °C in January. Average daily temperatures
range from -9°C in January to 19°C in July.

Each site contained 20 sampling quadrats, 10 m square, with 1 m aisles
between them. Besides the sampling quadrats, one quadrat near the center of each.
site contained the equipment for monitoring soil temperature at several depths, air
temperature, and humidity. Maps of the sites are shown in Snider and Snider

(1987)

 

Field Micro

An ar
for placemen
was clear of

reasonably h

Earthworm

Earth
sampling deg
1986). The :
humus layer
iloriZOn). Sa
backfilled wi
and eanhWO.
Snider (1984
formalin, U1
and 00coons
electronic ba
regressinns \

presen/ed m:

54

Field Microcosm Site

An area near the northwest corner of the ELF CONTROL site was selected
for placement of field microcosms. Although completely shaded by canopy trees, it
was clear of brush, herbaceous ground cover, and large boulders, making it

reasonably homogeneous throughout.

Field Population Sampling Methods

Earthworm Censuses

Earthworms were collected at two-week intervals using a stratified random
sampling design consisting often 25 X 25 cm samples per date (12 in 1985 and
1986). The samples consisted of five subsamples: the litter layer (O-horizon), the
humus layer (A-horizon), and three successive 10 cm samples of subsoil (B-
horizon). Samples were removed to the field lab for processing, and the holes were
backfilled with similar soil from outside the site borders. Samples were processed
and earthworms and cocoons retrieved using the protocol outlined by Walther and
Snider (1984). Earthworms were killed with alcohol and preserved in 10%
formalin. Up on identification and determination of developmental stage, worms
and cocoons were weighed to the nearest 0.1 mg using a Mettler AE-200
electronic balance. Since formalin dissolves out some lighter fats and oils,
regressions were derived (R.M. Snider, pers. com.) and used to convert

preserved mass into live mass before worms were assigned a size class:

A. tuberculata: FW= —0.7186 + 1.0214 ><PW
L. rubellus: FW= 1.5609 + 1.024XPW

 

 

where F W =
was calculatt

live mass.

Environmer
Soil t
late October
1991-92 and
at midnight.
each samplin
averaged to
beginning, It
eStimate of r
[10f taken an
5°11 moisture

first Spring S

Dend
hand. sorting
State ForeSt

55

where F W: fresh (live) weight and PW: preserved weight in mg. No regression
was calculated for D. octaedra, and preserved mass was presumed to be equal to

live mass.

Environmental Data Collection for Field Populations

Soil temperatures were monitored at 2-hr intervals from early May through
late October throughout the study using Omnidata #222 Datapods. During the
1991-92 and 1992-93 winter periods, the datapods recorded soil temperature daily
at midnight. Soil moisture was determined gravimetrically fiom samples taken on
each sampling day. Temperatures throughout each four-week period were
averaged to obtain a mean temperature for that period. Moisture levels at the
beginning, middle and end of each four-week period were averaged to obtain an
estimate of mean soil moisture over the period. Since winter soil moistures were
not taken and the soil surface was frozen and under snowpack during these periods,
soil moisture during the winter was presumed to be the mean of the last fall and the

first spring soil moistures for modelling purposes.

Collection of Earthworms for Rearing Experiments
Dendrobaena octaedra were collected near the ELF CONTROL site by
hand- sorting moist litter in leaf-filled depressions. An area in the Copper Country
State Forest (T.41N, R.29W, sec.22), ofiMerriman Road and approximately 15 km

NW of Iron Mountain, was used as a collection site for A. tuberculata and L.

 

 

rubellus. Col
they might ha

Aporrr
traps." in an
numbers in th
disturbed by 1
approximatel
from the vari
replaced in o:
well-rotted h
for two to Si)
through the c

times by reap

Soil and '
Soil fi

ELF CONTF
Weeks befon
all collected
prepartition l
Preln

itseh Was illg

reSuiting in 1

56

rubellus. Collecting these species near the ELF TEST site was not prudent, since
they might have been affected by the proximity of the ELF antenna.

Aporrectodea tuberculata and L. rubellus were collected using “worm
traps.” In an early collecting trip, it was noted that earthworms were found in high
numbers in the soil directly below deer droppings, and in soil that had been
disturbed by previous collections. Traps were constructed by digging pits
approximately 40 cm square and 30 cm deep in natural depressions, keeping soil
from the various horizons separate. Roots were sorted out, and the soil was
replaced in original horizon order. Before replacing the leaf litter, about 3 cm of
well-rotted horse manure was spread over the filled hole, and was left undisturbed
for two to six weeks between collections. Earthworms were collected by sorting
through the disturbed soil horizon by horizon. The traps could be reused several

times by reapplying manure.

Soil and Litter Preparation for Incubator and Field Microcosm Rearings
Soil for both incubator and field rearing experiments was collected near the
ELF CONTROL site. Soil and leaves for microcosms were collected two to three
weeks before each sampling date, and soil and leaves for incubator rearings were
all collected in early May 1993. In all other respects, the methods of soil
preparation for these two portions of the study were identical.
Preliminary trials suggested that the structure of dried and sieved topsoil by
itselfwas highly altered leaving little pore space for aeration and percolation,

resulting in high earthworm mortality. Subsequent trials showed that mixing a

 

quantity of sa
aeration suffi
water relatior

Soil w
material frorr
quantity of s:
after collectir
sheets in the
removed. Rt
10psoil. The
for two to 111
When comp]
screen (1.25
then passed 1
thiOUgh, AS
prepared soi

TOpsi
toliSOil and (
With approx-
in a tray belt
VOIume, was

After deSiCc

ihOTOUghly ‘

57

quantity of sandy subsoil equal to 25% by volume improved water infiltration and
aeration sufficiently to reduce mortality and more closely simulate natural soil
water relations.

Soil was collected by first scraping off the leaf litter and coarse organic
material from the soil surface, after which squares of A horizon were removed. A
quantity of sandy subsoil was also taken and kept separate. On the first sunny day
after collection, the top soil squares were broken up and spread thinly on plastic
sheets in the sun. Any earthworms and cocoons found in the process were
removed. Roots were sorted out and placed in a corner of the sheet to dry with the
topsoil. The subsoil was spread out similarly. Both fractions were left to sun-dry
for two to three days to kill and desiccate any earthworms or cocoons remaining.
When completely dry, the top soil and subsoil were sieved through a hardware cloth
screen (1.25 cm mesh) to remove rocks and small roots. Top soil and subsoil were
then passed through a 3 mm mesh sieve, using the dried root mass to rub them
through. As a result, a quantity of dried fine root material found its way into the
prepared soil, adding structure.

Top soil and subsoil were mixed in the final sieving step: three measures of
t0p soil and one measure of subsoil were shaken through a window screen sieve
with approximately 1 mm mesh to separate the smaller fraction, which was caught
in. a tray below and saved. The larger fraction, usually 15 to 20% of the total
volume, was dried overnight in a 105 °C oven to desiccate any remaining cocoons.
After desiccated cocoons were removed, this oven-dried fraction was mixed

thoroughly with the finer fraction.

The le
floor by rakir
those that wu
suspended in
week of dryi
sealed plasti
:Obigofl
numbered pl
turning the 1

water.

Physical C1

Seve
Prepared so
milde betwr
determining
Characteris‘
imPOrtant i
also detern

San
°°mnositin

580% org;

58

The leaves used in both experimental settings were removed from the forest
floor by raking, taking care to remove only clean, nearly intact leaves and leave
those that were well decomposed or partially buried in the soil below. Leaves were
suspended in nets of 1.27 cm plastic mesh and allowed to air dry. After at least one
week of drying, leaves destined for use in incubator rearings were stored in loosely
sealed plastic bags. The bucket leaves were treated in the following manner: 15.0
i 0.05 g of leaves were placed into each of fifty plastic bags with 10 ml water and a
numbered plastic identification tag, and sealed. The leaves were moistened by
turning the bags several times over the next 24 to 48 hours to redistribute the

water.

Physical Characteristics of Prepared Soil

Several procedures were used to characterize the physical properties of the
prepared soil used in the field and incubator microcosms so comparisons could be
made between this and other studies. Composition analysis consisted of
determining the proportion of organic matter, sand silt, and clay. Water-holding
characteristics were examined by determining its field capacity; specific gravity,
important in the relation between gravimetric and volumetric soil moisture, was
also determined.

Samples were taken of three lots of prepared soil to determine its
composition. An average of the lots (Table 4) showed that the soil contained

5.80% organic matter, determined by combustion at 600°C. The inorganicfi'action

 

 

 

 

 

 

_fi—

Table 4. Per
three lots of

contained 6f
method outl

Field
procedure: '
Oi'ernight, t
40 cm long,
then droppe
Similarly to
10 the top 0
dripping f“
bag to redu
gravity for
the center r
deSiCCatOI,

three tllbes

59

 

Table 4. Percent organics and size distribution of inorganic fraction by mass in
three lots of experimental soil mixture.

 

 

 

% Organic Inorganic Fraction
Lot # Matter % Sand % Silt % Clay
1 5.82 64.0 20.3 15.7
2 5.76 66.2 19.0 14.8
3 5.82 65.7 18.3 16.0
Mean 5.80 65.3 19.2 15.5

 

 

 

 

contained 65.3% sand, 19.2% silt, and 15.5% clay, determined using a hydrometer
method outlined by Bouyoucos (1927).
Field capacity of the prepared soil was measured via the following

procedure: The soil was hand mixed with about 20% water by mass, allowed to sit

 

overnight, then mixed again before placing it in a plastic tube 5 cm in diameter by
40 cm long, with filter paper fastened to the bottom. The tube was filled with soil,
then dropped three times from a height of 10 cm onto a hard surface to compact it
similarly to the treatment of soil in the field microcosms. Water was added slowly
to the top of the tube until all the soil was thoroughly saturated and water was
dripping from the bottom. The top of the tube was loosely covered with a plastic
bag to reduce evap orative drying, and water was allowed to drain from the soil by
gravity for approximately 48 hours. Three 40 g samples of soil were taken from
the center of each tube, weighed, dried at 105 °C for 24 hours, cooled in a
desiccator, and re-weighed to measure water loss. A one-way ANOVA with the

three tubes as treatments and the samples within each tube as replicates showed no

significant di
capacity was
apparatus us

Speci
of air-dry so
ensure pack:
with water. '

mean of the

A me
earthworms
basis in the
A proce

The act]
and pig]

‘ Amflm
animal.
Anesthesia
Several
SOIUtion, c]

at 10W 00m

coIlcentrat?

60

significant difference within or between the tubes at the a==0.05 level. Mean field
capacity was 28.98% (SD=0.89, n=9) water by mass. Figure 3 illustrates the
apparatus used.

Specific gravity of the soil was determined by weighing ten random samples
of air-dry soil level full in 100-ml crucibles, tapped on the counter several times to
ensure packing. After removing the soil, the same crucibles were weighed level full
with water, the ratio of soil mass to water mass being the specific gravity. The

mean of the ten samples was 1.089 with a standard deviation of 0.027.

Earthworm Tattooing Procedure
A method was developed that allowed essentially permanent marking of
earthworms so that growth and development could be observed on an individual
basis in the field microcosms. The technique consists of three parts:
0 A procedure to anesthetize worms that does not adversely affect them;

0 The actual tattooing process, using a technique that does not damage the worm
and pigments that are nontoxic and stay in place; and

a A method to easily observe tattoos without overly stressing or damaging the
animal.
Anesthesia
Several anesthetic agents commonly used on invertebrates, such as LiCl
solution, chloral hydrate, and carbonated water, were tested. All were ineffective

at low concentrations or detrimental, often lethal, to the worms at higher

concentrations. Carbonated water was the mo st effective, but mortality was still

 

Figlue 3. D
SOll uged ill ti

61

 

 

 

 

 

 

 

 

 

 

I

.7, .V/zuuuyuuuzvzuyézz

 

 

e field capacity of the prepared

ents. See text for description.

rim

e

microcosm exp

in the incu

Figure 3. Diagram of the tube apparatus used to dete
bator and field

soil used

unacceptably
solution of N

Worm
toweling for
pile of bakin;
prostomium.
anesthetic re
was taken fr
anesthetized
Earthworms
minutes. allr
Procedure. <
then transfg

Overnight b.

Tattooing ]

A va
Coloring, It
The tattoo
in a matter
cases, Tart
the tip and

Clifllp e d Wl

62

. unacceptably high. Working on a hypothesis that pH was the detrimental factor, a

i solution of NaHCO, (baking soda) was tested and found both safe and efi‘ective.
Worms whose guts had been voided by holding them on moist pap er

. toweling for 24 hrs were placed in a Petri dish partly filled with water and a small

* pile of baking soda until the worm was unresponsive to prodding of its

; pro stomium. It was then immediately blotted dry and readied for marking. If the
anesthetic required more than three minutes to take effect, additional baking soda
was taken from the pile and mixed into solution. Twenty to 30 worms could be
anesthetized before the water required changing due to mucus buildup.
Earthworms treated in this manner were immobilized for approximately five
minutes, allowing them to be tattooed if one worked quickly. After the tattooing
procedure, earthworms were placed in clean water until they revived. They were

then transferred to plastic containers with moist pap er toweling to recuperate

overnight before being returned to moist soil.

Tattooing Procedure

A variety of dyes and pigments were tested for efiicacy, including food
coloring, India and other drawing inks, and commercially available tattoo pigments.
The tattoo pigments were found to work most satisfactorily; dyes and stains faded
in a matter of days to weeks, and drawing ink pigments proved lethal in many
cases. Tattooing was performed using a Minutien Nadel bent approximately 30 ° at
the tip and mounted in a wooden dowel using a short piece of aluminum tubing,

crimp ed with pliers to hold the pin (Figure 4). This arrangement also allowed the

 

 

Figure 4. Di
tube, crimper
wooden hand

llllllll

Figure 5. A
tattoos. T .
Catherine N

63

A B C

Figure 4. Diagram of tattooing tool. A -- Minutien Nadel with bent tip; B -- aluminum
tube, crimped to hold needle in place (can be opened to allow tip replacement); C —-
wooden handle. Total length of tool is approximately 12 cm.

 

Figure 5. Anterior 2/3 of sexually mature (clitellate) A. tuberculata, showing placement of
tattoos. T -- tattoos; C -- clitellum; P -- prostomium. Earthworm illustration by
Catherine Nerbonne.

 

needle to be
through a (ll
amounts of
taken not to
worm. Alth
affected seg
posterior to
tattooing. '
circulatory
buildup ma
pigment gr:

Fou
along the d
four diflere

COllltl be re

Viewing 0t

To 1
(Figure 6)
fmm a Wat
the Subject
and develo

rllbber hOS

64

needle to be replaced if necessary. The integument was pierced several times
through a drop of pigment placed on the worm‘s integument, burying minute
amounts of pigment between the integument and outer muscle layer. Care was
taken not to penetrate the coelom, as pigment in the coelom was found to kill the
worm. Although necrop sies were not performed to determine the cause of death,
affected segments filled with fluid and increased greatly in size, while segments
posterior to the affected segments wasted and became necrotic within 48 hours of
tattooing. This suggested that a buildup of fluid probably pinched off the
circulatory system, and possibly the gut as well, at the point of injury. Fluid
buildup may have been due to blockage of nephrostomes or dorsal pores by
pigment granules.

Four tattooed dots, three segments apart, were placed behind the clitellum
along the dorsolateral row of setae (Figure 5). By using four sequential dots of
four different colors (white, red, blue, and green), 256 unique marking patterns

could be recognized.

Viewing of Tattoos
To observe marked earthworms, a mouth-op erated aspirator apparatus

(Figure 6) adapted from Thielemann (1986) was used to suction an earthworm

from a water-filled dish into a glass tube of a diameter appmpriate to immobilize

, the subject. The tube allowed thorough examination of the worm for its markings

; and deveIOpmental stage under a dissecting scope without damage. While the

 

rubber hose above the examination tube was pinched, the worm could be held in the

 

 

 

 

Figure 6. A
u%tAr
Stopper; D ..

 

 

 

 

 

C

Figure 6. Aspirator apparatus used for viewing earthworms (adapted fromThielemann
1986). A -- Glass tube for examining earthworms; B -- rubber tubmg; C -- Jar with rubber
StOpper; D -- mouthpiece.

 

tube by hyd

hose washe

Microcosm

Micr
moisture re
(Figure 7).
cross-strip:
of NITEX’E
to tack it it
and replace
escape of e
Oil, and lht
Permanent
removal an

A s‘
Passage of
50 the Cab]
and the Co

66

tube by hydrostatic suction; when the hose was released, the extra water in the

hose washed the worm out without injury.

Field Microcosm Rearings

Microcosm Construction

Microcosms for rearing earthworms under semi-natural temperature and
moisture regimes were constructed from 5-quart polyethylene ice cream pails
(Figure 7 ). Four quadrants were cut from the bottom of each bucket, leaving intact
cross- strips 1 cm wide to support the weight of the soil it would contain. A circle
of NITEX® netting (80 mesh) was fastened inside the bottom, using a soldering gun
to tack it in place and latex caulk to seal the edges. Circles were cut out of the lids
and replaced with netting that would breathe and admit rainwater while preventing
escape of earthworms. The bottom 3 cm of an additional set of buckets was cut
off, and the wire handles were removed. These buckets served as sleeves placed
permanently in the ground to receive the microcosms, simplifying their periodic
removal and replacement.

A small hole was pierced at one edge of the netting on each lid to allow
passage of the moisture probe cable, and was reinforced with stiff plastic squares
so the cables could move freely without escape of worms. Since both the probe

1

: and the connector plug on the other end of the cable were larger than the hole, the

probe became part of the lid assembly.

 

 

 

 

 

Figure 7.
Screened 1
buttom; C
removal fr

67

 

Figure 7. Diagram of bucket microcosms used in the field portion of the study. A —-
Screened lid, shown here without the TDR moisture probe; B -- bucket with screened
bottom; C -- bucket with bottom removed, used as a sleeve for easy placement and
removal from the ground.

 

 

 

Field Place

Micr
in four para
alternately (
array to sim
soil surface
contained r
fourth was
chemistry a
project. Fi

Eac'
bottom anc
screen and
and the em
movement
t0p 0f the ;
Weather.

A 5
main set 0:
One Specie
These wer
Stage Were

respeCthe

68

Field Placement of Microcosms

Microcosms containing prepared soil, leaves, and earthworms were placed
in four parallel rows often holes each, 0.5 m apart. Microcosms were placed
alternately 0.5 and 1.0 m apart along each row, leaving wider paths through the
array to simplify their tending and retrieval. Holes were deep enough so that the
soil surface inside and outside each microcosm was at the same height. Three rows
contained replicate populations of each of the three worm species studied, and the
fourth was a set of buckets with no worms, used as a control group for soil
chemistry and physical property studies which were not completed during this
project. Figure 8 shows the layout of the microcosms within this grid.

Each hole was lined with a sleeve, and loose sandy subsoil was placed in the
bottom and smoothed flat to provide adequate hydraulic contact between the
screen and the native subsoil below. A garden stake was driven next to each hole,
and the end of the moisture probe cable was rubber-banded to it to decrease probe
movement within the microcosm. A 100 ml plastic sample jar was inverted over the
top of the stake and the connection plug to provide some protection from the
weather.

A separate set of 12 microcosms, four for each species, was placed near the
main set of field microcosms for field-rearing cocoons. All cocoons gathered from
one species’ microcosms each month were placed in a separate cocoon microcosm.
These were examined monthly, and the number of cocoons in each developmental
stage were counted. Hatchlings of each species were distributed back into their

respective earthworm microcosms as described elsewhere in the text.

 

OCT
CON
RUB
TUB

Figure 8. I
011 center, v
buckets, C(
rubellus bu.

69

 

 

1 23 5 67 9 10
OCT. 0. O O. O O
CON. 0. O O. O O
RUB. O. O O. O O
TUB. O. O O. O O

 

Figure 8. Layout of the bucket microcosm site. Buckets in each group are 0.5 m apart
on center, with 1.0 m aisles between groups. Codes for treatments: OCT -- D. octaedra
buckets, CON -- control buckets with no worms (not part of this project), RUB -- L.
rubellus buckets, TUB -- A. tuberculata buckets.

Temperatu:
Arnb
study. using
bucket arraj
loggers bec
buckets we
random for
horizon 5 c
was there It
more than ;
immediatel
during the
equal] to th
Soi'
reflectome
V°lumetrie
beginning
adehuately
1991) Flt
recording
WaVefmm
i11this stul

printer (1:.-

70

Temperature and Moisture Monitoring

Ambient soil temperature was monitored at 2-hour intervals throughout the
study, using the data logger for the ELF Control site, about 35 m distant from the
bucket array. During the second year of microcosm experiments, extra data
loggers became available to test whether soil temperatures inside and outside the
buckets were equal. Three field microcosms, one of each species, were chosen at
random for temperature probes. Two additional probes were placed in the A
horizon 5 cm deep at both ends of the bucket array to monitor ambient soil. Never
was there more than :t O. 5 °C difference between individual buckets, nor was there
more than i 0. 5 °C difference between bucket and ambient soil temperature, either
immediately outside the microcosm array or at the ELF monitoring site 35 m away,
during the three months these temperatures were monitored. This difference was
equal to the precision of the temperature probes and data loggers used.

Soil moisture in the microcosms was measured using time- domain
reflectometry (TDR). TDR has been used as a nondestructive method of measuring
volumetric soil water content for about 15 years (Topp et a1. 1980), but is just
beginning to gain general acceptance because of the experience necessary to
adequately interpret the waveforms generated by the apparatus (Catriona et al.
1991). Field measurements have usually been made by producing a paper tape
recording of the oscilloscope waveform on the TDR meter. Interpretation of the
waveforms invariably led to problems with human error. The improvement utilized
in this study employs a laptop computer linked directly to the meter in place of the

printer (Figure 9). The waveform is fed into the computer, where the trace is

 

 

Figure 9.
5011 moistu
TDR meter

71

 

 

 

 

 

 

 

 

 

 

Figure 9. Time domain reflectometry (TDR) apparatus used in this study. A -- coaxial
soil moisture probe; B -- Tektronix 1502C TDR meter; C -- laptop computer connected to
TDR meter via a serial cable.

 

smoothed a
program str
Developme
computer a
moisture c:
100 wavefi

approxima

Microeosr
Mir
November
Sampling \
sampling (
Continued
On
approxim‘
e(illilibrat
filling, tn.
moist soil
soil_ A la
reintrOdu
01

72

smoothed and mathematically analyzed, providing consistent interpretation. The
program stores the result in a *.WKl-formatted spreadsheet file (© Lotus
Development Corp.) Both the TDR meter (mounted on a pack frame) and the
computer are battery-p owered, so that a nearly instantaneous measurement of soil
moisture can be obtained using portable apparatus. The TDR meter can test 75 to
100 waveforms on a single battery recharge, and the laptop computer can run

approximately two hours on one battery.

Microcosm sampling

Microcosms with earthworms were first placed in the field in early
November 1991. The first regular sampling date was in early May 1992, and
sampling was repeated every fourth week until late October, for a total of seven
sampling dates the first season. The second season began in mid-May 1993, and
continued until early October, for a total of six monthly samples the second year.

One day before bucket filling, prepared soil was rehydrated to
approximately 20% gravimetric moisture, mixed thoroughly and allowed to
equilibrate overnight. Just before
filling, the soil was again thoroughly mixed. The buckets were loaded with 2 l of
moist soil and dropped three times from a height of about 10 cm to compact the
soil. A layer of moistened leaves was placed on the soil after earthworm
reintroduction.

On each sampling date, the microcosms were removed to a field laboratory,

the leaves were hand-sorted, and the soil first hand— sorted then water-sieved to

 

 

 

retrieve all
(1984). Af
weighed ( g‘
rdoadedtv
she. Large
process. N
thepopulal
7. or 8 thrc
uereaddet
they vvere l
(tephcate
neretreah
Earthwom
removed fi

toreducet

Illcubatm

Let
rOtatable (
listed in T
(22.5% m

fit tests (It

73

retrieve all worms and cocoons, using techniques identical to Walther and Snider
(1984). After all earthworms were identified by their tattoos, examined and
weighed (gut-full mass), they were returned to their microcosms, which had been
reloaded with moistened soil and leaves, and the buckets taken back to the field
site. Large worms were replaced if they had died or were injured in the retrieval
process. Newly hatched worms from cocoon rearing microcosms were added to
the populations of one-third of the buckets, either buckets 1 through 3, 4 through
7, or 8 through 10 depending on the month. By rotating the times when hatchlings
were added to each set of buckets, distinct size cohorts could be maintained until
they were large enough to be removed and marked individually. Multiple
(“replicate”) cohorts allowed calculation of the variance between them. Cohorts
were treated as sets of individuals of identical size in the model generation phase.
Earthworms were out of soil less than one hour, and the buckets were usually
removed from the field in early morning and returned the evening of the same day

to reduce disturbance effects.

Incubator Rearings
Incubator Experimental Design
Levels of temperature and gravimetric soil moisture were arranged in a
rotatable central composite design (Gill 1978), with levels at each design point
listed in Table 5. Four additional points were placed at the center of the design
(22.5% moisture, 10°C) to simulate an orthogonal design and allow goodness-of-

fit tests (Myers 1971). The ranges of each variable were chosen to parallel the

 

”’—

Table 5. Pa
experiment:

\-

Each level

Commonly
collected a

Eac
for each 01
50i1 0f the
EM cont:
Octaedra)
Containers
during the
within eac
experimen

Aft

Weeks f0r

74

 

Table 5. Pairs of temp erature and moisture levels used in the incubator rearing
experiments.

 

 

Temperature Moisture
(°C) (% gravimetric)

3 22.5

5 17.2

5 17.2
10 15

10 * 22.5 *
10 30

15 17.2
15 17.2
17 22.5

 

 

Each level = four replicates except *, which had five sets of four replicates.

commonly observed ranges of these two parameters, based on environmental data

collected as part of the ELF project.

 

Each design point was represented by four l-quart polyethylene containers
for each of the three earthworm species. They were loaded half full of prepared
soil of the proper moisture level, and ten to twelve moistened sugar maple leaves.
Each container was stocked with 6-8 (A. tuberculata and L. rubellus) or 8-10 (D.
octaedra) individuals of different sizes and developmental stages. Additional
containers with spare earthworms were kept at each level to replace those that died
during the experiment. By monitoring a few individuals with widely spaced masses
within each container, individual worms could be tracked through the entire
experiment with little chance of confusion.

After a two-week acclimation period, containers were sampled every two

weeks for ten weeks. On each sampling date, the gut-full mass of each worm was

 

measured tc
developmer
httle damag
containers ‘
containers,
developmer
and monito
weekly as c
Ana
biweekly 5:
periods use
converted
comm. ).
Occ
IWO-week
decompos
eases Were
have been
highmort:
Sets, anal)

rePlaceme

75

measured to the nearest 0.1 mg with a Mettler AE-200 electronic balance and its
developmental stage determined. After hand— sorting to remove worms with as
little damage as possible, the soil was water-sieved to retrieve cocoons. The
containers were washed and reloaded, the live worms placed back in the same
containers, and dead worms replaced by earthworms of similar mass and
developmental stage. Cocoons were placed in Petri dishes filled with moist soil
and monitored every second week for development, and up to two or three times
weekly as cocoons neared hatching.

Analysis was performed on mass and stage changes over three pairs of
biweekly sampling periods: periods 1-3, 2-4, and 3-5, corresponding to the 4-week
periods used in the field microcosm rearing experiment. The actual masses were
converted to discrete size classes, minimizing autocorrelation effects (I. Gill, pers.
comm.)

Occasionally, all worms in a container would be found dead at the end of a
two-week interval. Cause of death could not be determined, because earthworms
decompose very rapidly and would have to be found shortly after death. These
cases were restricted to replicates at the two highest temperature levels, and may
have been a result of buildup of metabolic wastes, resulting in poisoning. Due to
high‘mortality and the sensitivity of response surface designs to unbalanced data
sets, analysis was performed on bootstrapped means derived by resampling with

replacement of the values obtained at each design point.

 

 

 

Response 5
Res;
of multiple
application
because ecr
restrictions
problem (C
experiment
where thes
Purpose of
an optimur
augmentati
pomts repr
RSM prodt
behavior 0
An
design, or
were (let/e]
mmafim
points) at 1
rotation ()1

denote the

76

Response Surface Methodology and Bootstrapping Techniques Employed

Response surface methodologies (RSM) were developed to examine effects
of multiple factors in situations where fully factorial designs were impractical. The
application of response surface methodologies in ecology is quite limited, partly
because ecologists are unfamiliar with the technique, and partly because of
restrictions placed on the experimental design that make it inappropriate to the
problem (Clancy and King 1993). The necessity of predetermined levels of
experimental variables, makes it impossible to use RSM in many field situations
where these levels cannot be precisely controlled. RSM is useful when the
purpose of the experiment is to find the levels of two or more factors that produce
an optimum in the response variable (Gill 197 8:271). It consists of either
augmentation of a general factorial design or a simple geometric pattern of design
points representing specific combinations of the treatment variables. Secondarily,
RSM produces a linear or quadratic regression that may be used to predict the
behavior of the response variable in the region surrounding the optimum.

An important class of response surface designs is the central composite
design, or CCD. Although not limited to second-order response surfaces, CCDs
were developed with second-order polynomials in mind. The basic CCD begins '
with a factorial design in an orthogonal array, adding a design point (often multiple
points) at the center of the array, and also points placed along all of the axes of
rotation of the original factorial design. The design points are usually coded to

denote the direction and distance from the center point in Cartesian coordinates.

 

 

 

 

 

 

 

 

 

 

The
is often con
the distance
rotatable. t
for the incu
fiddbound
nearly unil
approximat
Gill 1978).

The
as a fully f
second-on
disadvanta
factorial d
earthworn
when exar
temperatu
emPloyed.
%mmab
times and

estimated
p0Dulatio

heint in ll

77

The center of the CCD might not be close to the optimum being sought, so it
is often constructed so the precision of the estimated optimum is only affected by
the distance from the center point, not its direction. Such a design is termed
rotatable. One of the simplest of these is illustrated in Figure 10, the design used
for the incubator rearings. Multiple points are placed in the center, producing a
field bounded by the outer ring of points in which the precision of the estimate is
nearly uniform, and is unaffected by distance from the design center. It then
approximates an orthogonal design, allowing goodness-of-fit tests (Myers 1971,
Gill 1978).

The advantage of this design is that it accomplishes roughly the same thing
as a fully factorial 5 X 5 design with 13 points instead of the 25 necessary for a full
second-order factorial, realizing a great saving in time and effort. The
disadvantage of RSM is that, due to the reduction in number of points from a fully
factorial design, a balanced design is essential. High mortality in all three
earthworm species at the two highest temperatures severely unbalanced the design
when examining growth rates and deve10pmental stage change with respect to
temperature and soil moisture. To correct for this, bootstrapping (Stine 1990) was
employed, in which the available data were randomly sampled with replacement to
obtain a balanced data set with which to run analyses. By resampling multiple
times and determining a new mean of a fixed-size set each time, a variance of
estimated means could be generated about the true population mean. This
population mean was then used as the response variable at the appropriate design

point in the RSM, and a second-order multiple regression was calculated to model

 

30

MOISTURE (%)
N
01

N
o

15

Figlue 10. ;
factor level i
g‘aVimetric 1

 

 

78

 

 

0, «IE
30 -- ( - )
(-1, 1) (1, 1)
E 25 --
g (' fir 0) (0! 0) ( E! 0)
E (5 points)
0 ..
E 20
(-1i-1) (1! '1)
(0. 47)
15 J. 1L 4 a
0 5 10 15 20
TEMPERATURE (°C)

Figure 10. Design of incubator rearing experiments. Numbers in parentheses are the
factor level codings used in the rotatable central composite design. Moisture is
gravimetric moisture.

 

the behavior 0
population me

A pro g
It calculated 1
based on the 1
program also

° No first-o
or interac
all six var
class (S).
they were
eliminate

‘ Terms we
u=0.05 lt
facto pm

° The least
criterion
tmmdn

The pro(
Significa‘
Table R .

values, a
from the

An e
been Publisl

Boo
number fro

1990); ther

79

the behavior of each size class or developmental stage in one of the incubator
population models.

A program written in Turbo Pascal accomplished the resampling efficiently.
It calculated multiple regression, storing the coefficients and their significances
based on the resampled data set in a spreadsheet file for later analysis. The
program also allowed Stepwise elimination of variables, following a few basic rules:

0 No first-order variables were removed before first eliminating any higher order
or interaction terms including that variable. For instance, the first step involved
all six variables: temperature (T), moisture (M), T2, M2, T XM, and initial size
class (S). In this case, T and M would not be considered for exclusion, because
they were included in higher-order terms, but T2, M2, TXM, and S could be
eliminated on the first step.

' Terms were eliminated only if they were found to be nonsignificant at the
a=0.05 level in at least 200 of the 1000 bootstrapped data sets, resulting in a de

facto power of at least 0.8 for each variable.

0 The least significant term was considered for elimination first; if the first
criterion was not met by this term, the choice moved to the next least significant

term if it satisfied the previous criterion.

‘ The process was repeated until the R2 of the equation began to decrease
significantly, or the critical value for R2 was not met at the or = 0.05 level.
Table R in Rohlf and Sokal (1995: 125) was used as the source for critical
values, although critical values for five and six variables had to be extrapolated

from the table.

An excellent review of resampling methods for ecological applications has
been published by Crowley (1992). The concept for the bootstrapping program

used here came indirectly from ideas presented in Crowley’s work.

Bootstrapped significance values are sensitive to deviations in sample

number from that actually observed (or expected) during data collection (Stine

1990); therefore, a different number of individuals per design point were used for

W
and stages us

_Schi
D. oct

each species
these levels
aPproximatt
Observed) a
Classes had
expected in
aclitellate 5
against pre

Appendix 1

80

 

Table 6. Incubator experiment demographic breakdowns for all earthworm species
and stages used to calculate multiple regressions.

 

Total Worms N used

 

Species Stage / Size Class Observed in model
D. octaedra all immatures 354 40
immature classes 3, 4 236 20
all aclitellates 196 20
all clitellates 216 25
L. rubellus all immatures 629 70
immature classes 4, 5 185 20
all aclitellates 85 10
all clitellates 125 15
A. tuberculata all immatures 632 70
immature classes 5, 6 261 30
all aclitellates 247 25
all clitellates 141 15

 

each species and developmental stage to approximate the N expected at each of
these levels. Table 6 lists the Ns used for each of these. The N used is
approximately the total divided by nine (the number of experimental levels
observed) and rounded to the nearest five individuals. The smallest immature size
classes had no possibility of stage change, so smaller numbers approximating the
expected number of large immatures were used to calculate the immature-to-
aclitellate stage change regression. Results for the equations were also checked
against predefined upper and lower bounds, which are listed in the tables in

Appendix A with the coefficients for each equation.

 

The fit
size, or stage
subtle difiere'
decrease the ‘
error. Size c.
instantaneou
width of eacl
ideal conditi
from size cla
developmen‘
immature to

After
tracked fror

could be me

Where ”1‘ :
k are Specie
hatching ]
under half 1
glOWth ran

based on n:

81

Model Development

The first step in developing any matrix model is deciding how many age,
size, or stage classes to use; too few classes, and the model becomes insensitive to
subtle differences between developmental stages or sizes; too many classes
decrease the precision and make the model unwieldy and prone to multiplication of
error. Size classes need not all be the same width, if their width reflects the
instantaneous growth rate of that size class during a specified time period. The
width of each size class should be chosen so that the organism, when subjected to
ideal conditions, will not grow entirely through a class during a single period, i.e.,
from size class 3 to size class 5. Similarly, if the life cycle is split into
deve10pmental stages, it should not be able to skip stages, for instance from
immature to clitellate, without spending at least one time period as an aclitellate.

After examining eight to ten individuals of each species that could be
tracked from small immature through reproducing adult, growth in all three Species

could be modelled using a von Bertalanffy growth curve (Poole 197 4):

m,=M.....><r1—(b><e "X" “or

where mt = the predicted mass at time t, Mm, = the maximum mass attainable, b and
k are species-specific growth coefficients, and Mo = the time elapsed since
hatching. This produces an asymmetrical S-shaped curve whose inflexion point is
under half the maximum mass. Using these calculated growth curves, the total
growth range of each species was divided into six to eight discrete size classes

based on mass, which cut across all developmental stages. Table 7 shows the

W

Dendrobae
Size 1
code N2 m:
0

4

20 (
57

5
7
66

b) (4)

1
2
3
4
5
6

matrix t

1 Developm
C -- clitell

3 N = noon

m

maximum or:
end of a give
tuberculata
least 5% 0f(

Stagt
Size COde f0
aClitellate 5
have a Code

Cocr

0=t
1=t

:t
3=r
4=r

82

 

Table 7. Maximum mass in each size class for the three earthworm species studied.

 

Dendrobaena octaedra Lumbricus rubellus Aporrectodea tuberculata
Size Max. Devel. Size Max. Devel. Size Max. Devel.
code N2 mass (g) stagesl code N2 mass (g) stagesI code N 2 mass (g) stages1

 

0 --- --- coc --- --- coc --- --- coc
4 0.01 I 3 0.05 I 8 0.0313 1
20 0.025 I 2 0.1 l 13 0.0625 I
57 0.06 I,A 0.2 I 31 0.125 I

49 0.5 I,A,C
so 0.75 A,C
94 >o.75 C

35 0.25 I
41 0.5 LA
38 0.8 I,A,C
80 1.1 A,C
121 > 1.1 C
matrix order = 11 matrix order = 10 matrix order = 13

Developmental stages: I -- immature, A -- aclitellate or postclitellate,
C -- clitellate, coc -- cocoon.

2 N = individuals in each size class, irrespective of developmental stage, used to
generate the VBGF.

137 0.14 A,C
66 >o.14 C

Gméth—IO
00

1
2
3
4 135 0.1 I,A,C
5
6

OOQGMAUJNl-‘o

 

 

 

maximum mass in each size class for the three species studied. Size classes at each
end of a given developmental stage were combined (for instance, a few A.
tuberculata aclitellates were actually size class 4) until each size class comprised at
least 5% of the total number of individuals within that developmental stage.

Stage designations were included in the coding scheme by adding 10 to the
size code for aclitellates and 20 to the size code for clitellates; thus a mid-sized
aclitellate D. octaedra would be coded as a 14, and the largest clitellates would
have a code of 26.

Cocoon development was tracked using a scoring system:

0 = cocoons showing no sign of development
1 = those with a recognizable embryo
2 2 embryos exhibiting a functioning vascular system and pharyngeal hearts

3 = mature embryos with pigmentation on at least the first ten segments
4 = hatchlings

 

Cocoon deve
approach to t
cocoons sinc

as enough frt

Determinati

Durin
° living or
t increasin

' entering
(develop

The first of r
survivorship
divide them
Same class 0

All a:
normal distr
calculated u
assunlptiona
(mass at lit:
to the Kolm
combilltttlm

exception: 1

83

Cocoon developmental rates were modelled with these scores using a degree-day
approach to estimate length of time to development. Moisture was not a factor for
cocoons since cocoon culture trials showed that they continued to develop as long

as enough free capillary water was present for efl‘ective gas exchange.

Determination of Fate Probabilities for Inclusion in Matrices
During each time period, a given worm has the possibility of:
0 living or dying (measured as survivorship);
0 increasing, decreasing or staying in the same size (growth / retrogression); and

0 entering the next developmental stage, regressing, or staying at the same stage
(developmental stage change).

The first of these is straightforward, and can be represented as a probability of
survivorship during the period. The latter two, however, require more effort to
divide them into the three possible fates (growth, retrogression, or remaining in the
same class or stage), and assign probabilities to those fates.

An assumption was made that in a large population these fates follow a
normal distribution around some mean size or stage change, which can be
calculated using levels of environmental variables as coefficients. To test this
assumption, normality of growth difierences between adjacent monthly samples
(mass at time 2 minus mass at time 1) was tested using the Lilliefors modification
to the Kolmogorov-Smirnov goodness-of-fit test (Sokal and Rohlf 1995). A
combination of both field and incubator microcosm samples were used, with one

exception: the winter field periods (dates 1-2 and 7-8) were excluded because they

 

 

 

 

 

Table 8. Lill
monthly srze

Siz
_CE

1
2
3
4
5
6
7
8

represented
probabilitie
was groupe
different ra‘
less than 0,
conform to
goOdness-c
is a contim
Values in ti,

Bot
deserjbe th
Possible fa
f“notion W

leided lllt

84

 

Table 8. Lilliefors probabilities for Kolmogorov- Smirnov goodness-of-fit tests on
monthly size increments for three earthworm species.

 

 

 

D. octaedra L. rubellus A. tuberculata

Size Proba- Proba- Proba-
Class N bility N bility N bility
1 39 0.0221 120 0.0809 31 0.3807
2 97 0.4324 160 0.3982 92 0.1136
3 296 0.3015 186 0.1035 253 0.6259
4 342 0.1434 288 0.1202 103 0.9702
5 306 0.1575 193 0.3501 118 0.0654
6 227 0.0151 168 0.7606 115 0.0024
7 --- --- --- --- 126 0.4930
8 --- —-- --- --- 127 0.9720

 

 

 

represented growth over approximate six-month periods. Table 8 shows
probabilities of significant departure from normality. The dataset for each species
was grouped by initial size class, because earthworms of different sizes grow at
different rates. Although some size classes departed from normality (those with
less than 0.05 probability), size increments were generally found to roughly
conform to a normal distribution, and the assumption was accepted. A similar
goodness-of-fit test could not be performed on developmental stage change; mass
is a continuous variable, whereas stage change is not: it has only three possible
values in this growth model.

Both the mean and standard deviation must be derived from the data to
describe the size increment distribution before it can be divided among the three
possible fates. The standard deviation can be used to generate a cumulative normal
function with an offset equal to the mean change, and the resulting curve can be

divided into three regions representing the three fates (Figure 11). In this example,

 

 

 

o 9
Fun on

PROBABILITY
c

Figure 11.
Labelled art
and B deno.
resptthlVely

85

1" SHRINKAGE NO (GROWTH

 

 

 

 

, CHANGEt
0.8 --
t /A . a. e .A .
g 0.6 -- i
m
at.
O 0.4 _.. B
0-2 " Mean change (0.5)
0 If + a; + . fl 4 A 4. —i
.2 -1 0 1 2 3

Z-SCORE

Figure 11. Hypothetical cumulative normal curve adjusted right to a mean change of 0.5.
Labelled are the Z-scores which represent growth, no change in size, and shrinkage. A
and B denote the probabilities associated with shrinkage and [no change + shrinkage],
respectively.

 

 

 

the shrinkag

probability c

was generat
were summt
function is l

logistic funt

where p = t

score Z. T?
A St

calculated

for each de

Where y :
Size/ Stage
the matrix
three Sepa
third Sum

to prOdUCt

matrix, a r

86

the shrinkage (B) and growth (1.0-A) probabilities are both about 0.3, and the
probability of no change in size (A-B) is approximately 0.4.

To derive the cumulative normal function, a standard normal curve
was generated to six standard deviations on either side of 0.0, and the increments
were summed every 0.01 SD to obtain a cumulative normal curve. Since this
function is based on powers of e and is symmetrical about its inflexion point, a

logistic function was used to model it:

1
1+ e {-1.701 x Z)

 

p:

where p = the probability that the measurement is equal to or less than a given 2-
score Z. This function has an R2 = 0.9999, and an F(2,1193) = 10,328,626.

A second-order multiple regression in two variables, plus size, was
calculated for size change and spread, stage change and spread, and survivorship

for each deve10pmental stage. The form of the full equation was
y=a +b-T+c'M+a('T2 +e-M2 +f°T°M+g°S

where y = the response, T = soil temperature, M = soil moisture, and S = the
size/ stage class. Once the probabilities were calculated, they could be loaded into
the matrix in their respective cells. The final matrix is actually the combination of
three separate matrices: one describing size change, another stage change, and a

third survivorship. The size and stage change were multiplied element-by-element

 

to produce a combined matrix; this matrix was postmultiplied with the survivorship

matrix, a matrix with the size- and stage- sp ecific survivorships placed along the

 

diagonal. Fi
having been
illustration (
the size-stag
represents t?
The block it
immatures.
changes bet
The Gs alor
numbers in
must sum u
nonnegativ
below dent

negative (1:

Val
Slluetures
from Pre-I
within ind-
proleCtlon

pOPUlatim

87

diagonal. Finally, the size- specific fecundities were placed along the top row after
having been multiplied by the calculated cocoon fertility. Figure 12 is an
illustration of an 11 X 11 D. octaedra matrix. The top row of numbers represents
the size-stage codings at the beginning of a period, and the left column of numbers
represents the possible fates of each class, denoted by letters within the matrix.
The block letters A, B, D, and F respectively indicate the position of cocoons,
immatures, aclitellates and clitellates. The Roman letters C and E represent stage
changes between immature and aclitellate, and aclitellate and clitellate respectively.
The Gs along the top row represent size- specific fecundities. These are the only
numbers in the matrix which may be > 1.0; with them removed, individual columns
must sum up to 51.0. Blank cells are always zero; cells with letters in them are
nonnegative. Cells on the long diagonal represent no size or stage change; those
below denote growth or positive development, and those above shrinkage or

negative development.

Model Testing and Validation
Validation was performed by comparing monthly ELF project population
structures to those predicted by the models. Population vectors were constructed
from pre-ELF (1984-1988) populations by pooling all population data for a species
within individual sampling dates. These data were used as starting points for
projection, and the projected vectors were compared by ANOVA with observed

population vectors taken from the same site four weeks later. Ten monthly sets

 

 

 

 

 

 

 

 

 

 

 

Figllre 12.

88

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

o 1 2 3 4131415242526
0A GGGj
1ABB

2 BBB

3 BBBCC

4 BBCC

13 CCDD

14 CCDDDEE
15 DDEE
24 EEFF
25 EEFFF
25 FF

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Figure 12. Representation of the D. octaedra matrix. Details in text.

 

 

(dates 1-3, 2
pre-operatio
for the final ‘
with serial or
individuals s
total popula

temperature

Rout
SS’Stat, Inc.
with Quattr
were produ
data collect
Internation
PTOgrams, ;

Sev
in the Turb
Lotus *_ W.
use thut
Spreadshe6

prolect. N

89

(dates 1-3, 2-4, 3-5, . . . , 10-12) were obtained for each year throughout the ELF
pre-operational phase except 1987, when earthworms were sampled only monthly
for the final two months. Although the dates overlapped, there was no problem
with serial correlation because sampling was not performed using the same
individuals sequentially; each vector was a temporally distinct subsample of the
total population at the site. The only overlap involved temporal patterns of

temperature and moisture.

Data Collection and Analysis, Model Building,
and Other Computer Programs Used

Routine statistical analysis was performed with SYSTAT for Windows (©
Systat, Inc. 1990). Database, spreadsheet and graphing fianctions were performed
with Quattro Pro 6.0 (© Novell, Inc. 1994). Other graphics in this manuscript
were produced with Presentations 3.0 (© Novell, Inc. 1994). Many programs for
data collection and model building were written in Turbo Pascal 6.0 (© Borland
International 1992). These included TDR and multiple regression bootstrapping
programs, and programs used to build, run and test the population models.

Several routines and program units released to the public domain were used
in the Turbo Pascal programs. One was a unit that simplified reading and writing
Lotus *. WK 1 spreadsheet files, written By Dan Glanz. This unit made it possible to
use output from custom Turbo Pascal programs directly in statistics and
spreadsheet applications. It is found in nearly all the programs written for this

project. Numerical techniques for multiple regression and matrix manipulation

 

 

 

were taken f
(Savitzky an

for greater a

90

 

were taken from Press et al. (1986). A numerical method for quadratic smoothing
(Savitzky and Golay 1964) was used in the TDR program to smooth the TDR trace

for greater accuracy.

 

 

 

Tim

The 3

and after dr
incubator ar
it is destruc
best left un
(3) results ;
drying. TE
Same volur
due to dim
necded to i
(2) the difl

the volum‘

are to be u

TD

f0r monitc

TDR Volu

Chapter 4

VALIDATION OF SPECIALIZED TECHNIQUES

Time Domain Reflectometry vs. Gravimetric Moisture Methods

The gravimetric soil moisture method, which involves weighing soil before
and after drying, is one of the most commonly used methods, and was used for the
incubator and field portions of this project. The problems with this method are (1)
it is destructive, and cannot be used for monitoring changes in experimental units
best left undisturbed; (2) time and space are required for drying the samples; and
(3) results and length of drying time fluctuate with humidity, especially when air-
drying. TDR, on the other hand, is nondestructive, can be used to monitor the
same volume of soil periodically, is nearly instantaneous and free from variation
due to differential drying. Its weaknesses are (l) considerable proficiency is
needed to interpret the waveform traces, unless they are mathematically analyzed;
(2) the difference in expense between the necessary apparatus; and (3) it measures
the volumetric water content rather than water content by mass. If these methods
are to be used together or compared, a conversion formula must be used.

TDR and gravimetric methods were compared in the field microcosms used
for monitoring earthworm growth. During the first five collection dates in 1992,

TDR volumetric moistures were taken in each bucket immediately before removal

91

 

 

 

and transport
microcosm w
oven-dried a‘
desiccator ar
over a wide r

gravimetric ‘

where G = g
equation wa
F-ratio was
periodically
units used ir

individual p

0111}
tllberculatc
lllOCeClure3
became lar

Apc
SPeCies, bc
tattooing.

92

and transport to the field lab. Approximately 100 ml of soil taken from each
microcosm was weighed to the nearest 0. 01 g using a Mettler AE35 balance and
oven-dried at 105 °C for 24 hr. Samples were cooled to room temperature in a
desiccator and reweighed to determine water loss, resulting in 200 comparisons
over a wide range of moistures. The conversion formula from volumetric to
gravimetric water content was

G = 18.95 - 0.2148>< V + 0.01844 x V2
where G = gravimetric and V = volumetric moisture, respectively. The R2 for the
equation was 0.823, its standard error was 2.55 for 200 samples, and the ANOVA
F-ratio was 463.6 (p = 0.00000). This equation was used to convert the
periodically-monitored moisture data in the buckets to the gravimetric moisture
units used in the remainder of the study. Figure 13 shows a graph of the 200

individual points and the regression line.

Testing the Tattooing Technique

 

 

Only the larger worms (above approximately 80 mg) of L. rubellus and A.
tuberculata were tattooed. Smaller individuals were too easily damaged by the
procedure, and were analyzed as cohorts of similar—sized individuals until they
became large enough to tattoo.

Aporrectodea tuberculata accepted tattoos the mo st readily of the two
species, both in terms of longevity of the marks and of survival immediately after
tattooing. One hundred eighty-one individuals of this species were marked over

the course of the project; 131 (72%) survived the tattooing procedure and were

erlv n”- 1

M
3.: MEDFQOE Diem—5.5.59.0

Figure 13
filled With 1

 

 

93

 

 

 

40
§
33
D 30
l-
£9
‘2’
9- 20 ~- I?
m r
I—
I.”
E 10 --
E
0
0 T T +— r 46 Ar fii
0 10 20 30 40

VOLUMETRIC MOISTURE (%)

Figure 13. Regression of gravimetric on volumetric soil moisture in field microcosms
filled with prepared soil.

 

 

 

 

 

placed in fie
longer. All
small but sit

observed in

where M=ir
0ft]
various tim
mean lerrgt'
for the rem
worms rerr
group to b-
time of tht
= 0.022) tl
than the In
between tl
Other factt
Tat

brown pig
SPeeies th
Own pigm
(74%) sur

SurViVorg

94

placed in field microcosms. Of the survivors, 112 (85%) lived one month or
longer. All individuals dying during the first month initially weighed < 0.6 g. A
small but significant linear increase (11 =19, R2 =0.319, F(1.17) =9.45, p=0.007) was

observed in survival with increasing initial mass (Figure 14):
Survival = 0.799 + 0.232 X M,

where M=individua1 worm mass.

Of the 112 earthworms tracked at least one month, 82 were removed at
various times during the study because two or more of the four dots were lost. The
mean length of time that at least three dots remained visible (the "duration time")
for the removed worms was 279 days (Figure 15). At the end of the study, 20
worms remained alive with readable tattoos; seven of these were among the first
group to be marked, having retained their marks for 698 days. The mean duration
time of these 20 individuals was 527 days, significantly longer (t = 7.54, df=100, p
= 0.022) than worms that were removed; all of them had been in microcosms longer
than the mean duration time of the group that had lost its tattoos. Mean masses
between the two groups was not significantly different at the a = 0.05 level, and no
other factor that may have been the cause of this difference has been found.

Tattoos on L. rubellus, whose integument is iridescent and contains reddish-
brown pigments, were not as easy to see. The marking duration was shorter in this
species than in A. tuberculata because faded marks were masked by the worm’s
own pigments. One hundred ninety-three L. rubellus were tattooed; of these, 142
(74%) survived long enough to be placed in microcosms; 107 (7 5%) of the

survivors remained in the microcosms at least one month. A strong tendency

 

 

SURVIVAL

SURVIVAL

Figure 14
following .
SUn/ival ra

95

P
on

P
a:

L. rubellus

SU RVIVAL
P
h

 

 

 

l I I I I I 1 4]
l I 1 r ——r ‘I v

0 0.2 0.4 0.6 0.8 1
INITIAL MASS (g)

 

.°
co

P
as

A. tuberculata

SURVIVAL
o
In.

P
N

 

o
t
r

0.2 0.4 0.6 0.8 1
INITIAL MASS (g)

 

‘_ L 1 IL 4
v p I

O

Figure 14. Survival rate vs. initial mass of tattooed L. rubellus and A. tuberculata
following their first month after introduction to the field microcosms. Points are actual
survival rate; lines are regression lines. See text for the equations.

 

 

20 y
15 ._

10 ‘-

 

30 a—

20 ~~

10 -—

 

Figure 15.
Worms wer
‘0 ealculat.
The marke

96

+231 L. rubellus

 

 

.,
200 300 400 500 600 700
DURATION (DAYS)

 

 

  

 

 

3° " 279
l’ A. tu berculata

25 L,
20 --
15 --
1O ,,
5 l l]
0 L__l __ 7' _

0 1 00 200 300 400 500 600 700

DURATION (DAYS)
I Removed due to faded marklngs
Marked and active at end of trial

Figure 15. Marking duration distribution of worms of the two species studied. Removed
worms were those removed due to loss of markings during the experiment, and were used
to calculate the mean marking duration time (numbers next to arrows at top of graphs)
The marked and active worms retained their marks throughout the experiment.

toward dec
R2 =0.914,

negative ex

where M =
observed it

Of 1
the study (1
time. as de
active at ti
A. Iubercu
group and

Tat
earthworn'
Approxim.
Procedure
more, also
Species) 31
WithOut u]
mg was w
low. SUrv

wfiw‘l

97

toward decreased survival as a function of mass was noted in this species (11 =15,
R2 =0.914, F(2,12) =1015.0, p=0) (Figure 14). The relationship was modelled with a
negative exponential:

Survival = 1.0 — [1.168 X e('5'241 Xm],

where M = individual earthworm mass. No mortality of tattooed individuals was

 

observed in L. rubellus individuals more than 0.5 g.

Of the 107 individuals tracked at least one month, 36 were removed during
the study due to disappearance of two or more tattoos. These had a mean duration
time, as defined above, of 231 days (Figure 15). Thirty-six individuals remained
active at the end of the field study, with a mean duration time of 176 days. Unlike
A. tuberculata, no significant difference between duration times of the removed
group and those remaining alive was found at the p = 0.05 level.

Tattooing was an efi‘ective technique for marking and tracking individual
earthworms, and might also be used with other soft-bodied invertebrates.
Approximately 85% of aclitellate and clitellate worms survived the marking
procedure, depending on size and species, and marks remained visible for a year or
more, also depending on species. One drawback, however, was that certain
species, such as D. octaedra, and small immatures were too fragile to be marked
without unacceptably high mortality rates. Survival of earthworms smaller than 80
mg was well below 50% for L. rubellus; survival for D. octaedra of any size was
low. Survival rates for A. tuberculata, however, were relatively high, even for

small immatures.

 

 

DEVELO

Indi
(Poole 197
bounds. It
inflexion p
and catabr
of their en

for mainte

Where mt,
length of
because t
dimen Si0i
this stud:
each Spei

 

Chapter 5
DEVELOPMENT AND VALIDATION OF EXPERIMENTALLY DERIVED

POPULATION MODELS

 

General Growth Pattern

Individuals of all three species fit a you Bertalanfl’y growth function (VB GF)

 

(Poole 1974). Like the logistic, it increases monotonically but has upper and lower
bounds. It is unlike the logistic curve in that it need not be symmetrical about its

inflexion point. This model balances anabolism (tissue synthesis and development)

 

and catabolism (tissue breakdown), reflecting the tendency for animals to put most
of their energy into growth early in life, switching gradually to using their energy
for maintenance and reproduction later. The form of this equation is

m, =Mmax X [l — b X e Jim-t")?
where m,, the mass at a given time is a function of the maximum mass Mm, and the
length of time the animal has been growing, t-to. The entire equation is cubed
because the original model derived by von Bertalanfi‘y described growth in one
dimension -- length -- rather than three dimensions -- mass, which is the object in
this study. Table 9 lists equation parameters for a composite of all members of
each species tracked for at least three months (four consecutive mass

measurements within one season).

98

__—-————

Table 9. Pa
species fror

 

Species

D. octaed.
L. rubella
A. tuberc.

“l Number
‘3’ Number

 

Pop
with (1) th
greater the
calculated
all curves '
data forth
these long
3-month (
the first re
startin g W
mass the}
known to
the analys
lived indi‘
A. tuberci

Species, i1

99

 

Table 9. Parameters of the von Bertalanffy growth function for three earthworm
species from experimental rearings in field microcosms.

 

 

N N
Species Mm, b k points 0’ worms (2) R2
D. octaedra 0.1192 2.4705 0.4067 426 61 0.5174
L. rubellus 1.0988 1.8655 0.2793 106 18 0.8142
A. tuberculata 1.2695 1.3702 0.1844 305 40 0.8309

 

(1) Number of points used in regression.
(2) Numb er of individual earthworms used to build the comLosite.

Population composites were constructed by first selecting the individuals
with (1) the longest continuous record that (2) began with masses that were no
greater than 10% of the maximum mass recorded for that species. VBGFs were
calculated for each worm fitting these criteria, and the dates were adjusted so that
all curves passed through the mean cocoon mass as determined from the pre-ELF
data for that species. A preliminary VBGF was calculated for the composite of
these long-lived individuals, and the remainder of individuals that had a continuous
3-month (four consecutive mass measurements) record were added in turn, placing
the first recorded mass on the preliminary composite curve by time adjustment,
starting with the longest-lived worms; if records began below 10% of the maximum
mass, they were time-adjusted in a manner similar to the initial set. Individuals
known to have been damaged during retrieval or examination were eliminated from
the analysis. Figures l6, l7, and 18 show curves for (A) a representative long-
lived individual and (B) the population composite for D. octaedra, L. rubellus, and
A. tuberculata, respectively. Since the VBGF fits the growth patterns of all three

species, it was used to divide each developmental stage into size classes.

_‘l

 

 

 

O

MASS (g)

MASS (9)

Figure 11
field micr
Paramete

 

 

 

 

 

 

 

 

100
0.2 - A. Worm #292
0.15 - A .
§ ° , ‘
5?, 0.1: v i
g -l AClitellate
0 0 5 v OAclitellate
- Vlmmature
o . ..... ...fifi.f.f..f.fi
0 5 10 15 20
MONTH
0.2 - B. Population
0.154 _I'__._ ._,}--._-'-. - - __ ' _ . _
1..“ -"-, .: .' - -
§ :3 ‘:_" ' ;"_-_' -3 .
g 0-1 ‘ '_ = ' “III-5?"- _"'__ . -.
0.05 -
o .Tfi.fi-f...fi. . fl

 

 

 

Figure 16. Von Bertalanify growth curves for (A) a representative individual, and (B) the
field microcosm population of D. octaedra. Construction details in text; VBGF
parameters for (B) are found in Table 9.

MASS (g)

MASS (9)

Figure 11
field micr
Walnete;

 

 

100

 

 

 

 

 

 

 

 

0.2 .1 A. Worm #292
0.15 - . A
A . A
3 .
g 0.1 - v
é ACIitellate
v OAclitellate
0.05 -
vlmmature
0 —'— —fi —T l’ fi f ' ‘1' T fl
0 5 10 15 20
MONTH
0.2 ~ B. Population
0.15 - ,3-'-._‘-.'- - - _ . .
g ;:::_'_;_;4_ ..
(I) "'-.-'-'--' -
o 1...fffi...+ffi.fififr..fiere..fi.-....fi
0 5 10 15 20 25 30 35

MONTH

Figure 16. Von Bertalanfiy growth curves for (A) a representative individual, and (B) the
field microcosm population of D. octaedra. Construction details in text; VBGF
parameters for (B) are found in Table 9.

 

 

 

MASS (g)
c: o

('5

NIASS (9)

field micr
for 0’) ar

 

 

 

 

 

 

 

 

101
A. Worm WBRR
0.8 -
A A
A ‘ A
ACIitellate
oAclitellate
vlmmature
15 20
1-5 l B. Population
3 - - .
1 _
g, l
(D .
(é) .
0.5 ,
i
0 .fiffrfi. .. ,fifi..TT.. .j.r.fijfi...1r..fi
0 5 10 15 20 25 30 35

MONTH

Figure 17 . Von Bertalanfiy growth curves for (A) a representative individual and (B) the
field microcosm population of L. rubellus. Construction details in text; VBGF parameters
for (b) are found in Table 9.

MASS (g)

MASS (9)

Figure 1
field mic:
Paramete

102

A. Worm WGGW

 

 

 

1.5 -
A A
A A A A
1 - ‘
g .
U)
m .
g 0.5 - ACIitellate
OAciitellate
" vlmmature
O .f.f...%.fi....7.fi.f.
0 5 10 15 20
MONTH
2 j B. Population

 

MASS (g)

 

 

Figure 18. Von Bertalanfiy growth curves for (A) a representative individual and (B) the
field microcosm population of A. tuberculaz‘a. Construction details in text; VBGF

parameters for (B) are found in Table 9.

Sinc
temperature
complete fr
order. To t
microcosm
25 in Appe
experiment
outlined in
experimen'
incubator t
the microc

Inc
to Project
data (11 = 4
individual
and develt

The ANO'

Where N 2
Class C of
because [i

interac‘tio

103
Comparison of Incubator and Field Microcosm Models
Since microcosm data on immature through clitellate worms covered
temperatures from about 9 °C to 17°C while the incubator data were more
complete from 3 °C to 10°C, a combined model incorporating both data sets was in
order. To this end, it was necessary to ascertain whether the incubator and field
microcosm models were statistically similar in their predictions. Tables 23 through
25 in Appendix A show the regression coefiicients calculated for the incubator
experiment from each of the three species, using the bootstrap regression technique
outlined in Chapter 3. The regression coefficients for the field microcosm
experiments are found in Tables 26 through 28 in Appendix A. Unlike the
incubator trials, these were calculated directly from all the raw data collected from
the microcosms without the use of bootstrapping techniques.
Incubator and field microcosm models were compared by using each model
to project the next month’s population vector from pre-ELF natural population
data (11 = 48 monthly sets), then performing an ANOVA on the results, using
individual class densities as the dependent variable and developmental class (size
and developmental stage) and model origin (incubator or microcosm) as factors.
The ANOVA took the form
N = M + M ><C + e
where N = the class density or projected number of individuals in develoPmental
class C of model M, and e = the error term. This statistical model was chosen

because M should be sensitive to growth rate differences between models and the

interaction M X C shows difi‘erences between predicted population structures of the

 

two models
differences
artificially
As structui
are statisti
results for
Sin
and data v
incubator
generated
microcosr
and worm
cocoons e
transition
enumerat
C1
transitior
149 Mb
tril’hyllu
SlZe 0ftl
Present 1
SlZeS t0 1

through

104

two models. The single term C was not included because one would expect marked
differences between difierent sizes and stages, and the associated error would
artificially deflate the F-ratios and increase the type 2 error for the pertinent tests.
As structured, a nonsignificant F -ratio for both tests indicates that the two models
are statistically similar and their data can be combined. Table 10 gives the ANOVA
results for all three species.

Since the tests for all three comparisons (species) had nonsignificant results
and data were collected similarly for both sets (incubator and field microcosm), the
incubator data files were appended to the microcosm files and new models were
generated in the same manner as the microcosm models. The N for field
microcosm, incubator, and combined models is listed in Table 11. Since cocoons
and worms were analyzed separately using different techniques, the total number of
cocoons examined in each phase of the study is listed, whereas the number of
transition pairs (state at beginning of month and fate at end of month) is
enumerated for worms.

Compared with other studies using transition matrices, the number of
transition pairs per model cell is quite high. Bierzychudek (1982) used 104 and
149 individuals, respectively, to construct her two 7 X7 matrices of Arisaema
triphyllum L. demography; hers is one of only a few studies that even mentions the
size of the data set. The high numbers, however, are ofiset by the fact that the
present models are built using multiple regression equations, requiring large sample
sizes to be good predictors. Terms for the combined models are given in Tables 29

through 31 in Appendix A.

 

 

———-——_

Table 10. .
models for

_————

Source

D. octaedt
MODEL

MODELXl
ERROR

L. rubella
MODEL

MODELx
ERROR

A. tuberct'
MODEL

MODEL x
ERROR

\

TM
model typ
by collect

 

777 'l

105

 

Table 10. AN OVA results from comparison of incubator and field microcosm
models for all three earthworm species. n.s. = not significant at 0.05 level.

 

 

 

Signif-
Source SS DF MS F P icance
D. octaedra :
MODEL 443.93 1 443.93 0.0516 0.8203 n.s. '
MODEL x CLASS 1.8226><104 12 l.5189><103 0.1767 .9992 n.s.
ERROR 4.0673><105 1234 8.5982><103
L. rubellus
MODEL 255.31 1 255.31 0.7746 0.3790 n.s.
MODEL X CLASS 2022.6 12 168.55 0.5114 0.9085 n.s.
ERROR 4.0673><105 1234 329.60
A. tuberculata
MODEL 200.82 1 200.82 1.0505 0.3056 n.s.
MODEL x CLASS 2.7876><104 14 199.11 1.0416 0.4082 n.s.
ERROR 2.7222><105 1424 191.16

 

 

Table 11. Number of cocoons and worm transitions used in model construction, by
model type and developmental stage for the entire study. A more detailed summary
by collection date and transition type may be found in Appendix C.

 

 

 

 

SPECIES] TOTAL
MODEL COCOONS IMMATms ACLITELLATES CLITELLATES WORMS
D. octaedra
Incubator 490 377 209 215 801
Microcosm 1209 860 202 421 1483
Combined 1699 1237 411 636 2284
L. rubellus
Incubator 129 605 85 125 815
Microcosm 484 546 130 181 857
Combined 613 1151 215 306 1672
A. tuberculata
Incubator 66 632 247 141 1020
Microcosm 383 910 282 269 1461
Combined 1043 1542 529 410 2481

 

 

 

 

A co
actual value
forcing the :
relationship
without the
and a small
the observe

Regr
(worms and
octaedra, L
respectivelj
ELF projec
several rear
with an ext
lower right

lOial p0pu]

 

 

 

106

Comparison of Composite Models with Pre—ELF Subset
A common method for validation of complex models involves regressing
actual values on model predictions without a constant term (Poole 1974). By

forcing the regression line through zero, one obtains both a slope, indicative of the

 

relationship between the observed and predicted, and a standard error of the slope,
without the confounding variables introduced by a constant term. A slope near 1.0
and a small standard error suggests that the model behaves in a manner similar to
the observed data.

Regressions without constant terms of the total projected populations
(worms and cocoons) on the total observed populations after one month for D.
octaedra, L. rubellus and A. tuberculata are shown in Figure 19, 20, and 21
respectively. Observed populations and environmental data were taken from the
ELF project pre-ELF data set (1984-1988), with an n = 48. This set covered
several reasonably normal years (1984-1986), a warm, wet year (1987 ) and a year

with an extended drought in midsummer (1988). The abnormally low points in the

 

lower right portion of Figure 19 (D. octaedra) are from the summer of 1988. The
total population regression for each species was:
D. octaedra: Observed = 1.0196 x Model, Std. error = 0.0539
L. rubellus: Observed = 1.0724 >< Model, Std. error = 0.0545
A. tuberculata: Observed = 1.0220 >< Model, Std. error = 0.0344
where Model = the total population predicted by the model and Observed = the

total observed field population. In each case, the model prediction fell within the

or = 0.05 limits of significance.

OBSERVED

Figure 19.
actual ObSe‘
preoperatio
W°u1dindr
dashed line

107

 

1000 ,

 

// /

800 " I/ I

I. // /

600 ul- // //
I , ’

I / ,
/
I / /

OBSERVED

 

 

 

 

 

 

/ I
I.- / '
I // I
/
/
///I
//
O I A 1 J 1 4 l I 4
T I I ’—r ——l f I

o 200 400 600 800 1000
MODEL 1

Figure 19. Comparison of total size of modelled populations (worms and cocoons) to
actual observed populations of D. octaedra at the CONTROL site during the
preoperational phase of the ELF project. The solid line is a least- squares regression line
with slope = 1.0196 and r2 = 0.884; intersection of this line with the upper right corner
would indicate a 1:1 correspondence between modelled and observed p0pulations. The
dashed lines bound the 95% confidence interval about the slope of the regression line.

 

OBSERVED

Figure 20.
DOPUlation
PIOject. sf
SiS‘Iificanc

108

 

 

300

200 --

100 ‘-

OBSERVED

 

 

 

 

 

 

300

 

Figure 20. Comparison of total size of modelled populations to actual observed
populations of L. rubellus at the TEST site during the preoperational phase of the ELF
project. Slope = 1.0724 and r2 = 0.892 for the regression line. See Figure 19 for line and
significance details.

 

OBSERVED

Population:
Project. 31
Significance

 

109

 

400

300 -

200 -

OBSERVED

100 --

 

 

 

 

Figure 21. Comparison of total size of modelled populations to actual observed
populations of A. tuberculata at the TEST site during the preoperational phase of the ELF
project. Slope = 1.0221 and r2 = 0.490 for the regression line. See Figure 19 for line and
significance details.

 

 

As fir
developmen‘
There was it
survivorship
immature st
variables. 1
classes 1-2
individuals
because cla
while class.

Tab
worms, T.
soil moistu
and stage (
determinin
iIlfluencei
to weakly
Were Stror
ahIlost en:
mOde, ant

mOSt clite

110

The D. octaedra Model

As first conceived, the model divided the population first into
developmental stages, then assigned individuals to size classes within the stages.
There was to be one set of equations for size change, stage change and
survivorship for each stage. Examination of data showed that, especially in the
immature stage, the size classes did not respond similarly to the environmental
variables. Immature size classes were therefore divided into two subsets (size
classes 1-2 and 3-4) to better reflect the nature of the environmental effects up on
individuals of different sizes. The division was placed between classes 2 and 3
because classes 3 and 4 were also associated with change to the aclitellate stage,
while classes 1 and 2 were not.

Table 12 shows the effects of temperature and moisture upon each group of
worms. Temperature and initial size were important in growth of immatures, and
soil moisture levels were moderately to weakly associated with their survivorship
and stage change. Temperature and initial size also played critical roles in
determining growth and stage change of aclitellates, with moisture being a minor
influence in stage change only. Clitellate growth and survivorship were moderately
to weakly influenced by temperature and moisture, but growth and stage change
were strongly affected by initial size class. These last two effects combine to
almost ensure that once a D. octaedra becomes clitellate it remains in reproductive
mode, and the weak positive effect of temperature on survivorship indicates that

mo st clitellate worms will probably not survive through a long winter. Strongly

 

 

 

m
sizes and dev
moderate (M
regression si
indicated by

Immaturt
Classes l-I

Immatur.
Classes 3-.

Aclitellate

\

Clitellate
\

positive ter
replenish c
Bec
Variable te
Parameter
intrinsic g.
and moistr
the “Mn
at Several

map of th

 

111

 

Table 12. Effects of environmental variables on D. octaedra individuals of various
sizes and developmental stages. Effects are classified as strong (S), with a ps0.01,
moderate (M), 0.01<ps0.05, or weak (W), p>0.05 determined from multiple
regression significance values. The direction of change, positive or negative, is
indicated by a + or -.

 

 

 

 

 

Temp Term)2 Moist Moist2 Size

Growth M+ W-
21:13:31: Survivorship M+ S+

Stage Change no stage change

Growth S+ M- M-
3232:ng Survivorship regression nonsignificant; constant used

Sge Change M+ W+ W- S+

Growth S+ S- S-
Aclitellates Survivorship regression nonsignificant; constant used

Stage Chang W- W+ W- S+

Growth M+ M- W- S-
Clitellates Survivorship W+

Stage Change S+

 

 

Temp = soil temperature, Moist = soil moisture, Size = beginning size class.

 

positive temperature-dependent growth and stage change of aclitellates serve to
replenish clitellate stocks early in the warm season.

Because changes in population structure and size depend on a regime of
variable temperatures and soil moistures, it is impractical to determine population
parameters without first knowing the environmental regime for a given site, and the
intrinsic growth rate of the population at any practical combination of temp erature
and moisture. In matrix algebra, the growth increment, e', can be approximated by
the dominant eigenvalue, 1,, of the matrix (Caswell 1989). By determining the 1.1
at several levels of both soil temperature and moisture, one can construct a contour

map of the 71.1 surface. It is then possible to plot a trajectory on the contour map

 

correspondir
determine if
the year as d
the map witl
spent in this
This
11 in a 7X7
5% from 10
weighted le
for D. octa.
moistures,
22B), and g
moisture. '
microbes; :
Wetter) in ;
individuals
Several lay
Subject to

SuSceptibl.

Un

112

 

corresponding to the temperature and moisture regime of the area in question to
determine if a modelled population is viable at that site. If a substantial portion of
the year as determined by the temp erature-moisture trajectory is spent in an area of
the map with A, > 1.0, the population will remain stable or grow; if little time is
spent in this area, it will decline.

This was accomplished for each of the earthworm species by determining the
A, in a 7X7 grid of temperature (every 5°C from 0°C to 30°C) and moisture (every

5% from 10% to 40%). Contours were plotted using SYSTAT with distance-

 

weighted least- squares interpolation. Figure 22A shows the A, response surface
for D. octaedra graphically. Increasing temperature above 15 °C at moderate

moistures, or 20 ° C at low or very high moistures results in population growth (Fig.

 

22B), and growth rates rise rapidly to a maximum of >1.1 near 20°C and 30%
moisture. This makes sense because D. octaedra consumes litter conditioned by
microbes; an environment more conducive to microbial growth (warmer and
wetter) in and on leaves would also provide more food for this species. Quiescent
individuals were observed early in the spring on the surface of the A horizon under
several layers of leaf litter. If this is their mode of winter diap ause, they would be
subject to wide swings in temperature and moisture during the winter, and more

susceptible to a high mortality rate.

The L. rubellus Model
Unlike D. octaedra, L. rubellus immatures did not have to be divided into

multiple groups. The number of transitions used to generate this model was

Figure 22‘
OCtaedra f
illdiGates t]

 

   

 

 

30

 

 

 

 

 

 

MOISTURE (%)

 

 

 

5 1?) 13 ac
TEMPERA TURE ( °C)

Figure 22. A: 3-dimensional response surface plot of the population growth rate for D.
pctuedra from modelled data. B: Contour plot of the same surface. The shaded portion
indicates the temperature and moisture conditions under which population growth occurs.

—_———'—

Table 13. E1
sizes and de‘

Immature

Aclitellatc

Clitellate

 

substantial
regression
tuberculat

Ter
stage char
size was a
strongly a
relationsh
determjni
SllTViVOrs
aClliellatt

and mucl

114

 

Table 13. Effects of environmental variables on L. rubellus individuals of various
sizes and developmental stages. See Table 12 for symbol and caption explanations.

 

 

 

 

 

Temp Temp2 Moist Moist2 Size
Growth S+ W- W+ S-
Immatures Survivorship W- S+
Stage Change W- M- W-
Growth S+ S-
Aclitellates Survivorship W+
Stage Change W+ M+ M- S+
Growth S-
Clitellates Survivorship S+
Stage Change M+ M+ M-

 

substantially less than either of the other two; the result is evident in weaker
regression coefficients (Table 13) as compared with either D. octaedra or A.
tuberculata.

Temperature figured prominently in determining growth, survivorship and
stage change of L. rubellus immatures, while moisture was less important. Initial
size was also important in growth and survivorship. Aclitellate growth was
strongly affected by temperature factors, whereas moisture was more important in
relationship between temperature and stage change in clitellates, this stage is
determining stage change. Increase in size was weakly associated with higher
survivorship, and strongly related to a higher probability of stage change in
aclitellates. Because of the positive linear common in mid-to-late summer samples,

and much less so in the early spring or late fall. Clitellate survivorship, as in the

 

 

other stages
strongly wit

The '
populations
but fare ver
octaedra d:
it can also i
likely deter

individuals

Lik
to the env:
Significant
field censr
This may
immature

Th
D. octaec
Change fi
behaved.
Second 1)

115

other stages, increases with individual size, but the tendency to grow decreases
strongly with size, placing a de facto maximum size on individuals.

The response surface and contour plot (Figure 23) show that L. rubellus
populations do well at combinations of moderate temperature and high moisture,
but fare very poorly in low moisture-high temperature situations. Just as D.
octaedra depends on conditioned leaf litter for food, so does L. rubellus, although
it can also consume soil organic matter. The survival of the smallest size classes
likely determines the fate of the entire population; if conditions are favorable when

individuals are small, juvenile mortality is lower and the population will grow.

The A. tuberculata Model

Like D. octaedra, the immature size classes did not show similar responses
to the environmental variables across the range of sizes (Figure 24). The model
significantly underestimated the number of class 1 individuals actually collected in
field censuses, and overestimated the class 6 population by a substantial margin.
This may be expected, as the range in mass, from class 1 hatchling to largest class 6
immature, spans nearly two orders of magnitude.

The first attempt at dividing the stage was similar to the course taken in the
D. octaedra model: separate the larger classes which have the capacity for stage
change from the smaller classes which do not. After separation, classes 5 and 6
behaved well, but class 1 was still very diflerent from classes 2 through 4; in the
second phase, class 1 was isolated, and the modified model became substantially

more accurate.

 

Figure 2:
r “bellus 1
i1ldicates

 

116

 

 

 

MOISTURE (%)

 

 

 

 

5 i0 7 5 2o
TEMPERA TURE ( ”0)

Figure 23. A: 3-dimensional response surface plot of the population growth rate for L.
rubellus from modelled data. B: Contour plot of the same surface. The shaded portion
indicates the temperature and moisture conditions under which population growth occurs.

 

 

 

SLOPE

0.5

Wm 24
inmature
for all imr
Populatio
I110del. P

 

 

117

2 ‘ } Legend:
SLOPE-> +}:1: 95% C. I.

1.5 — 5 _-

 

y———_—E__-_———

SLOPE
T
+
|
|
I
I
I
I
I
|
I
|
I
I
I
I
I
I

0.5 —- +

 

 

0 I L l l l V l l
ALL 1 2 3 4 5 6
IMMA TURE
CLASSES CLASS

Figure 24. Slopes of observed vs. modelled A. tuberculata populations for each of the
immature size classes individually, from data initially modelled with one set of equations
for all immature size classes. Slope > 1.0 indicates that the model underestimated the true
population; slope < 1.0 shows that the observed population was overestimated by the
model. Pooled slope of all six classes is shown for comparison.

 

Sine
model was ‘
data were 5
could be us
could servr
equation w
earthworm
differentia
them to ar

1 model in
model wa:
observed
model are
Ta
Ofthe eqr
SUl’Vlvors
the imma
classes; i
Silrvivorl
Suwivor:
I

Terrlpera

118

Since the class 1 equations still underestimated the population and this
model was to be used for further experimentation, the pre-ELF field population
data were split into two subsets, 1984-85 and 1986-88, so that the first portion
could be used to fine-tune the model by adjusting the equations and the second
could serve as a validation subset. Only the a, or constant, term of the survivorship
equation was adjusted because it was thought that the method used to remove
earthworms from the buckets (water-sieving of a volume of soil) may have
difi‘erentially injured the smallest worms more than larger individuals, subjecting
them to artificially high mortality. After adjustment of this term to bring the class
1 model into agreement with actual numbers using the 1984-85 subset, the entire
model was applied to the 1986-88 subset for validation. Regression statistics of
observed numbers on modelled populations for each stage/ class combination in the
model are shown in Table 14.

Table 15 shows the level of significance of environmental variables in each
of the equation sets. Temperature was the only significant factor in the growth and
survivorship of the smallest immatures, and continued to be important throughout
the immature stage. Moisture became important in the intermediate immature size
classes; interestingly, it had a negative influence on mid-sized immature
survivorship. Size was also important in survivorship: an increase in size increased
survivorship.

Aclitellate growth increased, but survivorship decreased with moisture.

Temperature was important in survivorship, but the only significant factor in stage

_————'—-

Table 14. E
population:
adjust the c
the adjustn

 

Stag
Cocor

Class
Classes
Classe:
Aclitel
Clitell

Total pop

m
various si

\

\

11nmatr
Class

\

Immat
Classes

\

lInmat
Classes

\

Atlitell

\

Clitell:

\

119

 

Table 14. Slopes and confidence intervals of regressions of observed on modelled
populations of A. tuberculata after adjustment. The 1984-85 subset was used to
adjust the exp erimentally- derived model; the 1986-88 subset was used to validate

the adjustments.

 

 

 

1984-1985 subset (n=20) 1986-1988 subset (n=28)
Stage Slope 95% C.I. Slope 95% C.I.

Cocoons 0.9991 0.8232 - 1.1749 1.0104 0.8282 - 1.1926
Class 1 1.1172 0.9122 - 1.3223 0.8540 0.7067 - 1.0014
Classes 2-4 1.0237 0.8756 - 1.1719 0.9524 0.8407 - 1.0641
Classes 5-6 0.9528 0.8009 - 1.1047 0.9269 0.7816 - 1.0721
Aclitellates 1.0857 0.8832 - 1.2882 1.0960 0.9231 - 1.2689
Clitellates 1.1167 0.9407 - 1.2927 0.9317 0.7285 - 1.1349
Total population 1.0285 0.9208 - 1.1362 1.0181 0.9225 - 1.1137

 

 

 

 

Table 15. Effects of environmental variables on A. tuberculata individuals of
various sizes and developmental stages. See Table 12 for explanations.

 

 

 

 

 

 

 

 

Temp Temp2 Moist Moist2 Size
Growth M+ M-
Inéraizulre Survivorship W+ M-
Stage Change no stage change
Grth S+ W+ W-
EEEZSEIL: Survivorship S- S+
Stage Change no stage change
Immature Growth . M+ W- . . S-
Classes 5_ 6 Survrvorshrp Regressron nonsrgnrficant; constant used
Stage Change S+
Growth M+ M-
Aclitellates Survivorship W+ S- S- S+
Stage Change S+
Growth W+ S- S+ S- S-
Clitellates Survivorship W- W- S+
Stage Change W+ S- S+ S-

 

change wa:
important :

The
map, in co
tended to ?
decline sh.
not nearly
“plateau”
response :
under fav

less stablI

OI

used to i]
Organism

to answe

Temper:

 

120

change was size. On the other hand, the entire suite of variables seemed to be

important for growth, survivorship and stage change of clitellate A. tuberculata. —

 

The 2.1 response surface and contour map are presented in Figure 25. This
map, in contrast with that of D. octaedra, shows that A. tuberculata populations
tended to increase at relatively low soil temperatures and moderate moistures, and
decline sharply as temperatures rise. The point of maximum population growth is
not nearly so pronounced, either; the area of positive growth is a nearly flat, wide
“plateau” rather than a steep “mountain” as is the case with the D. octaedra
response surface. This is evidence of D. octaedra ’s ability to rapidly colonize
under favorable conditions, and A. tuberculata ’s tendency to maintain a more or

less stable population over more variable conditions.

 

Life Cycle Inferences and Comparisons Using Models
Once accurate population models are generated and tested, they may be
used to infer and examine many aspects of the life cycle and life history of an
organism, such as survivorship and longevity, and phenological questions about the
timing of development and reproduction. This portion of the chapter will attempt

to answer selected questions pertaining to life histories of the three species

Temperature-related cocoon development

Cocoon development times were modelled using a degree-day approach:

N: DD
(TM-T...)

tabercul‘
indicates

 

121

 

 

 

 

N
01

MOISTURE (%)
N
o

 

 

 

 

5 To is 20
TEMPERATURE (°C)

Figure 25. A: 3-dimensional response surface plot of the population growth rate for A.
tuberculata from modelled data. B: Contour plot of the same surface. The shaded area
Indicates the temperature and moisture conditions under which population growth occurs.

 

 

where N = .
constant fo
temperatur
developme
The
(Figure 26
significant
developm
developm
020°C. i

deveIOpm

where mt
the deno-
related t.

C
Octaedrl
some we
through
Were f0.
HolmStl

lultlbric

122

where N = calendar days needed to complete development, DD = the degree-day
constant for the species (a unit of energy accumulation over time), Tm = mean
temperature over the time measured, and T a = the base temperature below which
development ceases.

The cocoons of the three species developed at somewhat different rates
(Figure 26), although 95% confidence bands about the regression line indicate no
significant difl‘erence. Dendrobaena octaedra cocoons were found to continue
deve10pment at < -1°C (Table 16), whereas cocoons of L. rubellus cease
development below 014°C, and A. tuberculata cocoons will not develop below
020°C. Each point used in these regressions represents the mean daily

developmental rate of a cocoon cohort between two sampling dates:

mean score 2 - mean score 1
date 2 - date 1

 

where mean score x is the mean of the cocoon developmental scores at time x, and
the denominator is the number of days between two consecutive observation dates,
related to the mean soil temperature between the two dates.

Cocoons were also deposited at different levels in the soil. Most D.
octaedra cocoons were found on or very near the surface of the A-horizon, and
some were found in the leaf litter. Lumbricus rubellus cocoons were found
throughout the A-horizon and into the mineral soil. The cocoons of A. tuberculata
were found deeper in the mineral soil than those of either of the other species.
Holmstrup et a1. (1990) investigated the frost tolerance of cocoons of several

lumbricid species, among them D. octaedra and A. caliginosa [tuberculata?].

 

$5.3 sz2d04m>mo $.83 sz§u04m>m0

Figure 2
ten11)erat
A, B and
between

123

 

 

 

 

 

   

 

 

 

 

1000 l 1000 ..
T"; 800 I E 800 I-
In. as
3 3
E 600 I- E 600 -(
[U [U
a a.
O 400 ~- 0 400 ..
in a
a a
D 200 Q 200 .
0 0 If + t F:
0 5 10 15 20 0 5 10 15 20
TEMPERATURE (° C) TEMPERATURE (° C)
1000 ~- 400 y D
W", \— __________________
A ’0? ‘ ‘\ ‘ ' D. octaedra
3%: 800 l 3300 " \‘\ \\\ ""L. rubellus
1" ‘\ ——A. htberculata
E 600 .. 5 ‘
2 2 200 r
a. n.
E 400 -- g
> > 100 ..
g 200 ‘3
0 1 I o ’17 —r 1* 1
O 5 10 15 20 O 5 10 15 20
TEMPERATURE (° C) TEMPERATURE (°C)

Figure 26. Combined field microcosm and incubator cocoon development times vs.
temperature. A: D. octaedra; B: L. rubellus; C: A. tuberculata. Dashed curvesin graphs
A, B and C are 95% confidence intervals about the regression lines. D: comparison
between the three regressions; dashed horizontal line is 365 days.

 

__—_’——

Table 16. 1
derived fro:

 

D. octaedr
L. rubellu

A. tubercr

 

Dendroba
reduced h
-5°C. Sir
the soil, t
temperatI
temperat'
temperat
depositir
develom

factors (

develop]

these Sp

1

the coct

two/1.;

124

 

Table 16. Parameters for cocoon development equations for three lumbricids,
derived from combined incubator and field microcosm data

 

 

95% Conf. Interval
Ta Degree— N Regression (lower to upper)
(°C) days points R2 T, Deg. Days
D. octaedra -1.11 1854 30 0.838 -l.48 to -0.74 1874 to 1836
L. rubellus 0.14 2417 35 0.701 -0.84 to 1.11 2762 to 2148

A. tuberculata 0.20 1520 30 0.814 -0.13 to 0.53 1607 to 1443

 

Dendrobaena octaedra cocoons tolerated soil temperatures to -10°C, although at
reduced hatchability; Aporrectodea cocoons did not survive at temperatures below
-5 °C. Since the cocoons of the first species are deposited on or near the surface of
the soil, they must be able to withstand colder and possibly more variable
temperatures than those of species that deposit their cocoons in lower strata where
temperature effects are ameliorated. The series of minimum developmental
temperatures (Ta in Table 16) corresponds with the mean depth of cocoon
deposition, and to some extent the degree-day accumulation necessary for
development from zygote to hatched worm (Table 16), suggesting that all of these
factors (cocoon deposition depth, frost tolerance, developmental rate and minimum
development temperature) vary in concert in the evolution of life history traits in
these species.

Both incubator and field microcosm studies indicated that survivorship in
the cocoon stage was very high, close to 1.0. Three cocoons, one L. rubellus and

two A. tuberculata, were parasitized by an unidentified nematoceran fly. All other

 

—_———-—_

Table 17. 1
after cocor
and incuba

Species
_flr

D. octaer
nncro
incu
con

L. rubelI
micro
incu

cor

A. tuber
micrr
incr

co:

 

cocoons
rates, ho-
cocoon t
dlStlllCllt
deve10p1
study, p
DGlOngI
The fert
increaSe

daily ter

125

 

Table 17. Regressions of proportion of fertile cocoons on temperature shortly
after cocoon deposition in field microcosms, incubators, and combined microcosms
and incubators for three lumbricid species.

 

 

Species and N Std. ANOVA
Source Constant Slope points R2 Error 4)
D. octaedra
microcosms 0.776 0.0116 9 0.556 0.0485 0.021
incubators 0.125 0.0533 5 0.860 0.1001 0.014
combined 0.453 0.0325 14 0.402 0.1702 0.008
L. rubellus
microcosms 0.811 0.0086 8 0.321 0.0586 0.143
incubators 0.389 0.0364 5 0.794 0.0853 0.027
combined 0.587 0.0241 13 0.440 0.1143 0.008
A. tuberculata
microcosms 0.912 0.0224 11 0.035 0.0517 0.581
incubators 0.648 0.0106 5 0.607 0.0374 0.075
combined 0.766 0.0104 16 0.152 0.0934 0.076

 

 

cocoons that showed early evidence of development eventually hatched. Fertility
rates, however, were less than 1.0, and varied with temperature (in this study, any
cocoon that never visibly showed development was considered infertile; no
distinction was made between truly infertile cocoons and those that aborted early in
development). Table 17 shows fertility regression parameters for all phases of the
study. Infertile cocoons carried through with the rest of the cohort to which they
belonged appeared to be live, newly-deposited cocoons even after several months.
The fertility of field microcosm-raised D. octaedra and L. rubellus cocoons
increased linearly with temperature at the time of deposition (Figure 27). Mean

daily temperatures during the time of year when cocoons were deposited ranged

126

D. octaedra

 

 

 

 

 

 

0.8 .
2 0.6 y
E .
Lu 0.4 -
LL A
0.2- a e
0 S . . L
0 5 10 15 20
TEMPERATURE(° C)
L. rubellus
E ' l ,-, MICROCOSM
{E 0.4 1. ,. ......
LL INCUBATOR
0_2 A +
O COMBINED
o 5 10 15 20

TEMPERATURE? e)

A. tuberculaz‘a

 

 

 

o S - "r S

o 5 10 15 20
TEMPERATURE? C)

Figure 27 . Fertility rates of the cocoons of three lumbricids with respect to temperature
at time of deposition in field microcosms, incubator rearings, and combined microcosm
and incubator studies.

from about
cocoon fer
between 8

Apt
temp eratu
in the moc
in soil for
“new” (II(
have beer

soil for St

Phenolo;

A
mathema
day mon
Levels 0
part, we
obtain.“
at the B
through
data log
deriVed

April 2.

127

from about 10 to 20°C for both species. In this temperature range, D. octaedra
cocoon fertility was between 75 and 100%, and L. rubellus cocoon fertility was
between 83 and 100%.

Aporrectodea tuberculata cocoon fertility was not significantly affected by
temperature (Table 17); therefore, the mean (0.8814, or 88.14% fertility) was used
in the model. Because about 12% of cocoons did not develop, and they can remain
in soil for several months and appear to be fresh, some of the cocoons identified as
“new” (no apparent development) in this study and the ELF project data may not
have been new at all, but actually infertile or aborted cocoons that had been in the

soil for some time.

Phenology of Earthworms after Hatching

All analyses of survivorship and phenology in this section are based on the
mathematical models generated during this project. A “typical” year of thirteen 28-
day months was used, starting with May 1 as the first day of month 1 (Table 18).
Levels of soil temperature and moisture for months 1 through 6, and 7 and 13 in
part, were generated from the mean temperatures and moistures for each period
obtained by data loggers (temperature) and from A horizon soil samples (moisture)
at the ELF Control site from 1984 through 1993. Winter temperatures (months 8
through 12, and 7 and 13 in part) were taken from daily temperatures measured via
data logger during the winters of 1991-92 and 1992-93. Winter moistures were
derived from TDR readings measured through the snow on November 10, 1991 and

April 20, 1992. Winter moisture levels were assumed to be constant (the mean of

 

Table 18.
day month
models.

the abov
were ava

and moi:

Class 1 i
models.
used to
year dej

in Serie:

 

 

 

128

 

Table 18. Temperature and A horizon soil moisture means for each of thirteen 28-
day months in a typical year, employed for phenological analysis using earthworm
models.

 

 

Mean Mean
Month Start Date Temperature Moisture
1 May 1 9.5 33
2 May 29 12.5 30
3 June 26 15.0 24
4 July 24 16.0 25
5 August 21 15.0 25
6 September 18 11.0 30
7 October 16 8.0 25
8 November 13 5.0 25
9 December 11 1.0 25
10 January 8 0.0 25
11 February 5 0.0 25
12 March 5 1.0 30
13 April 2 3.0 33

 

the above two readings) under snowpack for modelling purposes, since no data
were available other than the begin-end points stated. Table 19 lists temperatures

and moistures for the months of a typical year.

Effects of hatching time on survival and development of worms

Cohort analyses of populations of earthworms hatching and appearing as
class 1 immatures at different times were examined using the derived population
models. A population vector consisting solely of 10,000 class 1 individuals was
used to seed the three models, starting at each of Months 1 through 6 of the typical
year defined above. Each month’s resulting vectors were passed to the next month

in series until Month 1 at the beginning of the third year, resulting in a period of 26

months (tv
months f0]
this time f
in the mot
productio
De
cohort dr
and total
3, with th
with pre-
were 0an
mid May
June. Tl
Summer
conditio-
Possibly
ShOWed
decreas.
these en
early in

develop

cohort

129

months (two fiill years) for the cohort starting in Month 1, and a period of 21
months for the cohort started at Month 6. The population vector was recorded at
this time for comparison to the other cohorts of the same species, and placed back
in the model until all had died to estimate maximum lifespan and cocoon
production. Table 19 lists the results of this set of model runs for all three species.

Dendrobaena octaedra cohort analysis shows that performance of the
cohort dr0ps significantly in terms of total population remaining at end of year 2
and total cohort cocoon production if cocoons hatch after the beginning of Month
3, with the highest numbers if cocoons hatch at Month 2. This timing coincides
with pre-ELF data from the CONTROL site, where old cocoons, ready to hatch,
were only found in quantity during the first and second sampling dates in early and
mid May, and many small immatures were collected from mid May through mid
June. The high number of small immatures hatching during spring and early
summer can not only take advantage of the ample food resources of newly
conditioned leaves from the previous autumn’s litterfall to grow rapidly, but may
possibly decrease the risk of predation by sheer numbers. Cohort analysis also
showed that maximum lifespan of a class 1 immature appearing after Month 4 was
decreased by more than 1 full year over those appearing earlier in the year. Under
these environmental conditions, it may be very advantageous for cocoons to hatch
early in the year, allowing time for small worms to grow and gain energy for
development and reproduction.

Lumbricus rubellus populations at the end of year two decrease with later

cohort start times, but not as dramatically as those of D. octaedra. Cocoon

 

 

Table 19.
times (Mo
lumbricid

Start 1
month in
Dendrobc

l

omgwtv

Lumbric

130

 

Table 19. Comparison of cohorts of 10,000 class 1 individuals started at different
tMes (Month 1 = May, Month 6 = late September to mid October) for three

lumbricid species at the end of Year 2.

 

Population structure at end of Year 2

 

 

Total Maximum
Start End Im- Aclitel- Clitel- Total Cocoon sum cocoons lifespan

month month ature late late Worms to date for cohort (months)
Dendrobaena octaedra

1 26 0 0 19 19 1705 1907 40

2 25 0 0 25 25 1665 1918 40

3 24 0 0 21 21 1240 1445 38

4 23 0 0 14 14 759 890 37

5 22 0 O 3 3 135 143 25

6 21 0 3 2 5 23 26 24
Lumbricus rubellus

1 26 6 2 7 15 729 844 42

2 25 8 2 4 14 501 571 37

3 24 9 2 4 15 240 346 42

4 23 10 1 3 14 48 106 32

5 22 9 1 2 12 18 29 25

6 21 8 1 1 10 2 4 24
Aporrectodea tuberculata

1 26 143 25 6 174 816 1157 88

2 25 126 22 5 153 667 994 87

3 24 28 4 1 33 84 136 71

4 23 20 2 0 22 26 40 70

5 22 30 4 1 35 44 96 69

6 21 37 5 1 43 57 109 68

 

End month = the number of months from start to end of Year 2.

 

productio
cohort to
to less thz
fairly higl
and decre
that this s
of its app
Al
in that th
maximur
Total p0
Point, to
LifeSpar
data fro-
period, ‘
densitie
lifeSpan
Studied
IIlortali'
stages,
"ubellu

throng]

131

production, however, does decrease significantly, from a maximum for the Month 1
cohort to 67% of the maximum for the Month 2 and 40% for the Month 3 cohorts,
to less than 1% of the maximum production in Month 6. Cocoons were found in
fairly high but variable numbers throughout the first three months of ELF sampling,
and decreased thereafter, also supporting model output. Cohort analysis showed
that this species, like D. octaedra, exhibited a decreased lifespan -- 1 to 1.5 years
of its approximate 3-year maximum -- if the young appear in Months 4 through 6.
Aporrectodea tuberculata shows a pattern similar to the other two species
in that the population remaining at the end of year 2, cocoon production, and
maximum lifespan are all highest in cohorts starting during the first two months.
Total populations at end of year 2 and cocoon production drop precipitously at this
point, to roughly 20% and 10%, respectively, of the values for earlier cohorts.
Lifespan also decreases, from roughly 7 years to a little over five years. Pre-ELF
data from TEST indicate that old cocoons are found throughout the sampling
period, but their frequency decreases somewhat after mid-July. Class 1 immature
densities are quite variable from year to year, but difi‘erences between date-Sp ecific
lifespan as indicated by the models varies significantly between the three species
studied. Dendrobaena octaedra loses 90% of its initial population to juvenile
mortality by month 4 of its 40-month lifespan, with noticeable mortality of all
stages, particularly clitellates, during the late winter months (10- 13). Lumbricus
rubellus loses individuals more slowly to juvenile mortality, with 10% surviving

through month 6 of its 41-month lifespan. It does, however, experience greater

 

 

 

mean den
dates, wh
of old cor
hatches tl
the perio
iteroparo
decrease

seems 10

P

B
producti
modelle.
at the be
records
Append

(
Octaedr
28, 29 2
Ofthe sf
(Deeve:
semflog

The dif

 

132

mean densities are not significant, with the exception of the mid- and late October
dates, which show a small decrease in small worms. The supply

of old cocoons and the presence of hatchlings seems to indicate that this species
hatches throughout the warm season, rather than concentrating most of the hatch in
the period most conducive for population increase. The longer lifespan and

iterop arous reproductive pattern of this species (discussed in the next section) may
decrease the selective pressure for concentrating the hatch into a shorter period, as

seems to be the case for D. octaedra and L. rubellus.

Phenological and life history comparisons between species

Because all three species studied showed maximum lifespan and cocoon
production for cohorts begun early in the warm season, comparisons between the
modelled populations were made using cohorts of 10,000 class 1 immatures started
at the beginning of Month 1 (May 1) of a typical year (Table 18). Complete
records of population developmental stage structure on a monthly basis are listed in
Appendix B.

Graph A of Figures 28, 29 and 30 shows survivorship curves for D.
octaedra, L. rubellus and A. tuberculata cohorts respectively. Graph B of Figures
28, 29 and 30 shows the phenology of aclitellates, clitellates and cocoons in each
of the species. All three of the survivorship curves resemble a Type II curve
(Deevey 1947 ) superficially, the more or less constant negative slope of the
semilo g graph indicating a constant mortality rate throughout much of the lifespan.

The differences between them lie primarily in either end of the curve. Maximum

 

 

1.00

C

4|.

0
n.

4|

nU.
0

_IWKO>_>N_3

ant.
0
m

 

0.01

f.\ (A.

KWMEDZ

Stages, ]

133

A. Survivorship

 

1.0000 ,

 

 

0.1000 53

 

 

0.0100

SURVIVORSHIP

‘1!

Jll
'rI‘l

 

 

 

0.0001

 

l

500 -

I

400 -

300 -

1

NUMBER

200 ~

I

100 -

 

 

 

YEAR

Aclitellates -—-— Clitellates ----- New Cocoons

Figure 28. (A) Total cohort survivorship and (B) phenology of selected D. octaedra
stages, based on a modelled cohort of individuals started May 1.

n__Iw~u.O\/_>N_DW

mmmEDZ

Fume

stages,

 

 

 

 

 

 

 

 

 

 

 

134
A Stwivorship

0- 5’
E .
(0
tr
0 -.
2
>
0:
3
(I) .1-

__ B. Phenology
tr
UJ
CD
2 ._
D
Z

 

 

Figure 29. (A) Total cohort survivorship and (B) phenology of selected L. rubellus
stages, based on a modelled cohort of 10,000 individuals started May 1.

 

Figure
stages,

 

135

A Suvivorship

 

 

 

 

 

 

 

_—
d
_—
_-
.1—
J

‘-

 

 

Figure 30. (A) Total cohort survivorship and (B) phenology of selected A. tuberculata
stages, based on a modelled cohort of 10,000 individuals started May 1.

 

mortality

worms (1

A
months.
with 10°/
asenher
between
graph is
constant
demonst
mortalit

(
showed
thatabc
reminis

Showed

both D.
“MW
of the 1
Winter,
cliteua

c0mm?

 

 

136

mortality during the winter months, particularly among aclitellate and clitellate
worms (Figure 29B and Appendix C).

Aporrectodea tuberculata had a much longer lifespan, a maximum of 88
months. Juveniles tended to survive a longer proportion of the lifespan as well,
with 10% of the individuals surviving 17 months or longer, more than twice as long
as either of the other species. This is evident in the part of each of the curves
between the top two “decade” lines. The slope of this portion of the A. tuberculata
graph is nearly the same as that of the next two decades, indicating a nearly
constant mortality through most of its lifespan. The other two species
demonstrated markedly steep er slopes in the first decade, indicating heavy juvenile
mortality.

On the opposite end of the survivorship curve, A. tuberculata (Figure 30A)
showed a significant negative change in slope between years 6 and 7, indicating
that about 0.1% of all individuals approached some maximum physiological age,
reminiscent of Deevey’s (1947) Type I survivorship graph. The other two species
showed a nearly constant mortality rate until the end of the lifespan.

The timing of reproduction was also different among the three species. In
both D. octaedra and L. rubellus, a significant number of clitellates began to
produce cocoons during the first year, with the peak falling in month 7, at the end
of the warm season. In D. octaedra, about 70% of clitellates died during the
winter, and were supplemented by overwintering aclitellates for a second major
clitellate peak in month 5 of the second year (Figure 28B). The first peak was

comprised of clitellates with a mean size/ stage class of 24.7 1; the second peak

 

!

 

mean 01:
gained 6
occurre
tempera
per clitt
about t1
Clitellat
as the f
About

year) b1
generat
Preject

fOund ‘

 

137

 

Table 20. Modelled maximum cocoons deposited per clitellate in any given month
of each year for three modelled lumbricid populations.

 

 

D. octaedra L. rubellus A. tuberculata
Maximum Month of Maximum Month of Maximum Month of
Year Cocoons Occurrence Cocoons Occurrence Cocoons Occurrence
l 144 6 48 6 --- ---
2 466 4 104 4 85 7,8
3 63 4 l7 1 80 6
4 10 2 39 6
5 15 7
6 6 6,7
7 --- ---

 

 

 

 

--- = Presence of clitellates, but no cocoon production.

 

mean class size was 24.95, indicating that the first-year reproducers may not have

gained enough energy for an intensive reproductive effort. The first peak also

occurred later in the year, when fecundity was depressed due to falling

temperatures. Another indication was the maximum number of cocoons deposited

per clitellate in any given month each year (Table 20). Although there are only

about two-thirds as many clitellates the second year, the combination of larger

clitellates and more favorable conditions produce more than twice as many cocoons

as the first year, approximately 68% of the total cocoon production (Appendix B).

About 11% of the total cocoon production by this cohort is deposited in the third

year, because clitellate densities rapidly decline. Peaks attributable to separate

generations are indistinguishable in the natural population censused during the ELF

project, but the greatest number of clitellates collected during an average year are

found during the late summer and early fall, just as the model predicts. Figure 28B

 

shows th
into the s
to overw
revert to
aclitellat
years, in
trend lat

decreas:

of the ii
are not
5 of the

Winter,

2, arou
in D. 0.
year 2 :
Chtella
as man
Second
by this
numbe

Worms

 

138

shows that about 75% of the clitellates from the first year do not survive the winter
into the second year. Very few D. octaedra clitellates revert to the aclitellate state
to overwinter (Table 21); a clitellate is about 14 times more likely to die than to
revert to the aclitellate condition. Figure 28B also clearly shows a decrease in
aclitellates coinciding with an increase in clitellate numbers during the first two
years, indicating that aclitellates become clitellates, but there is no reversal of this
trend later each year; aclitellate numbers remain stable or decrease as clitellates
decrease.

Given cocoon development times, D. octaedra cocoons deposited at the end
of the first year will likely hatch about Month 3 of the second year when conditions
are not as favorable for juvenile survival, whereas cocoons produced during Month
5 of the second year will have a chance to develop significantly before onset of
winter, hatching in Month 1 of the following year.

Like D. octaedra, L . rubellus also exhibits a second clitellate peak in year
2, around Month 3 (Figure 29B), although the peak is not nearly so pronounced as
in D. octaedra. The mean size class of clitellates at the year 1 peak is 24.85; in
year 2 it is 25.17. Again, the earlier timing of reproduction coupled with larger
clitellates that produce more cocoons over a longer period, allows only one-third
as many clitellates to produce nearly four times as many total cocoons during the
second year, compared to the first. Only about 14% of the total cocoons produced
by this cohort are deposited in subsequent years, due to rapidly declining clitellate
numbers; however, large clitellate size may partially compensate and allow a few

worms to produce a high number of cocoons (Table 19). As is the case with D.

 

Table 21
experien
lumbrici
datasets

ORIG
STA

D. octt

 

139

 

Table 21. Summary of proportions of each earthworm developmental stage
experiencing stage change or mortality during a sampling period, for three
lumbricid species. Proportions are calculated using both incubator and microcosm
datasets irrespective of temp erature and moisture.

 

 

 

 

ORIGINAL FATES
STATE IMM ACL CLI DEAD TOTAL
D. octaedra
IlVllVI 0.5835 0.0769 0 0.3396 1.0000
ACL 0.0047 0.7307 0.1635 0.1011 1.0000
CLI 0 0.0095 0.8578 0.1327 1.0000
L. rubellus
IMM 0.6138 0.0754 0.0008 0.3100 1.0000
ACL 0.0511 0.5069 0.3255 0.1165 1.0000
CLI 0 0.0649 0.7075 0.2276 1.0000
A. tuberculata
IMM 0.7201 0.0485 0 0.2314 1.0000
ACL 0.0508 0.6278 0.2068 0.1146 1.0000
CLI 0 0.1936 0.6962 0.1102 1.0000

 

IMM = immature; ACL = aclitellate; CLI = clitellate.

 

octaedrt
season \

reprodu

until de
die than
clitellat

during 1

sub stan
particuf
the cap
breedin
another
maXimr
It then
increas
number
Species
sPeciet
Clitella
eXplan

is the i

 

140

octaedra, many of the cocoons from the first year will hatch later in the second
season when juvenile mortality is high, conferring a selective advantage to delayed
reproduction under this temperature and moisture regime.

As is the case with D. octaedra, L. rubellus also tends to remain clitellate
until death, but not as markedly; L. rubellus is only about 3.5 times more likely to
die than to revert to aclitellate. Chris Klok (pers. comm.) has also observed that
clitellates of this species are more likely to move downward in the soil profile
during cold or dry conditions while retaining the clitellum.

Both of the above species exhibit basically a two-year life cycle, although
substantial differences exist between their survivorships and phenologies,
particularly in the areas of juvenile mortality and cocoon production. Both have
the capacity to live longer than two years and produce a few cocoons for another
breeding season, so generations overlap somewhat and cohorts spill into one
another. Aporrectodea tuberculata is very different. This species lives for a
maximum of nearly seven years, and begins cocoon production in its second year.
It then goes through a breeding cycle every year, the number of clitellates
increasing throughout the summer and peaking in Months 7-8, after which clitellate
numbers decrease rapidly to Month 13 (Figure 30B). Just as in the other two
species, total cocoon production in the second reproductive year (year 3 in this
species) increases by 26% over the first year, although there are only 79% as many
clitellates. Since the clitellate peak occurs at the same time each year, the only
explanation for the higher number of cocoons during the second reproductive year

is the increase in mean size class, from 27.12 in year 2 to 27.58 in the third year,

 

 

 

 

 

with lat}
increase
remainii

1
years 2
are dep
partly d
depend
develo;
the mo.
effecth
reprod‘
reprod
negatit

that se

increa:
Specie
Possib
memb.
a Sing]
reprot

also d

 

 

141

 

with larger individuals producing cocoons at a faster rate. Clitellate mean size
increases every year except the last, reaching 27.75 in the penultimate year and
remaining at that level.

Ninety percent of the total cocoons produced by this cohort are deposited in

years 2 through 4: 31%, 39%, and 20%, respectively (Table 19). Most cocoons

 

are deposited late in the warm season and during the early winter, and overwinter
partly developed. Since development ceases at temperatures near 0°C, they
depend on the comparatively low number of degree—days necessary for
development (Table 16) to allow them to hatch early in the warm season. Although
the modelled maximum lifespan of A. tuberculata is nearly seven years, the
effective length of the life cycle is approximately four years, with three
reproductive seasons and considerable overlap of successive cohorts. There is no
reproduction during the last year (Table 20); this, coupled with the increasingly
negative slope of the survivorship curve during this time (Figure 30A) indicates
that seven years is probably the maximum physiological age of this species.

Unlike the preceding two species, A. tuberculata aclitellate numbers 1
increase in winter, just as clitellate numbers are declining. This indicates that the
species uses a different overwintering strategy -- it reverts to an aclitellate state,
possibly to conserve energy during the cold season. Once they become clitellate,
members of the other two species remain so; three-quarters of them only reproduce
a single season. Aporrectodea tuberculata, however, is truly iteroparous,
reproducing up to seasons. Because this species reproduces for several years, it

also demonstrates a propensity for switching between aclitellate and clitellate

stages; 1
clitellate
changin;
I
remain s
an itero
octaedr
times bt
several
this spe
Populat
habitat

in numl

142

stages; uner the other two species, it is nearly twice as likely to revert from
clitellate to aclitellate as it is to die. Proportions between staying clitellate and
changing to aclitellate indicate that it remains clitellate for about 3.5 months.
Iterop arity in this species increases the probability that a population will
remain stable, because an extended period of suboptimal conditions will not immct
an iterop arous species as severely as essentially semelparous species like D.
octaedra or L. rubellus. It would also lessen the efi‘ect of suboptimal hatching
times because individuals hatching at these times are still likely to reproduce for
several seasons. Iteroparity is further evidence of K-adaptation in A. tuberculata;
this species is more effective at surviving periodically stressfiil conditions.
Populations of the other two species are probably better able to colonize new
habitat rapidly and proliferate under favorable conditions, but are likely to decline

in numbers when conditions are unfavorable.

 

U5

extreme
strengtl
site, by
(1984-
was abt
previ01
test the
Popula
tuberca

Predict

difiere
Operat
Popuh
signifi

densit

Chapter 6

USING THE A. tuberculata MODEL TO TEST FOR ELF EFFECTS

The ELF project was designed to detect soil ecological effects of low-level
extremely low frequency electromagnetic fields in soil (7 6 Hz nominal, field
strength in soil of 53.9i6.6 mV-m") (Snider and Snider 1994) near the ELF TEST
site, by comparing it with a similar CONTROL site in a before-after preoperational
(1984-1988) and an operational (1989-1993) 2><2 design. Because A. tuberculata
was abundant at the ELF TEST site, and the model developed for this species in the
previous chapter was a good predictor of its p0pu1ation dynamics, it was used to
test the hypothesis that there was no efl‘ect of the ELF EM field on earthworm
populations. If the EM field induced in the TEST site soil were to affect the A.
tuberculata population, a deviation of the actual field population from that
predicted by the model after antenna activation would be expected.

The two-sample t- statistic (Sokal and Rohlf 1995) was employed to test for
differences between field populations and model predictions during pre-ELF and
operational periods. A test of the regression slopes of observed on projected
populations before (11 = 48) and after (11 = 50) antenna activation showed no
significant difference (I = 0.0133, 94 (If) between total predicted population

densities during the pre—ELF and operational periods. When each stage was

143

 

—————I—

Table 22
operatic
separate

examir

becam

decrea
smalle
deviat
that tl

Cococ

conce
fecun

AAhhc

144

 

Table 22. t-tests of model prediction vs. field observations between pre-ELF and
operational periods, for the entire population and for each developmental stage

separately.

 

 

 

Pre—ELF Operational
Stage/Class Mean Std. Error Mean Std. Error t-value
Cocoon 1.0056 0.0620 0.8821 0.1093 1.1206
Class 1 1.0277 0.0633 0.8817 0.0664 1.5915
Class 2-4 0.9746 0.0428 1.0316 0.0431 1.0056
Class 5-6 .9387 0.0502 0.9675 0.0543 0.3895
Aclitellate 1.0927 0.0633 1.0819 0.0637 0.1203
Clitellate 1.0109 0.0681 0.6477 0.0691 3.7436***
Total 1.0221 0.0344 1.0214 0.0400 0.0133
P0pulation

 

 

 

 

N = 98, df= 94 for all tests. ***: Significant at at = 0.001.

examined individually, however, significant deviations from model predictions
became evident. The statistics for these comparisons are summarized in Table 22.

During the ELF operational period, clitellates exhibited a highly significant
decrease (or = 0.001) compared with pre-ELF model predictions. The cocoons and
smallest immatures also showed small, but not statistically significant, negative
deviations from model predictions for the operational period. It should be noted
that the cocoon values in Table 22 were calculated from the total number of
cocoons, not just those newly deposited.

One might wonder whether this effect occurred in all years, or if it was
concentrated in one or two years. Figure 31 shows the clitellate, total cocoon and
fecundity slopes and their associated 95% confidence intervals for each year.

Although there was a significant deviation below 1.0 in the 1986 clitellate slope,

 

Figur.
Year 1‘.
DaShe
exactl

 

 

145

A. CLITELLATES

SLqPE
i/
T—l—
l
|
|
I

 

 

B. ALL COCOONS

2.- Legend:

mtww

84 85 86 87 88 89 90 91 92 93

 

SLOPE
r

 

 

 

 

 

 

 

C. FECUNDITY
6..
5_
4..
m -_
83— l[ +
2—
1—
0 1 T T I 1 1 u l 1

Figure 31. Observed vs. modelled population slopes and 95% confidence intervals by
year for (A) clitellate numbers, (B) total cocoon numbers, and © clitellate fecundity.
Dashed lines in (A) and (B) indicate a slope of 1.0, where observed and modelled agree
exactly.

the pre—i
period t
1993) sf-
possible
electric
Anothe
ones he

of the \

seven t
develo
at abor
reprod
and p0
inform

two to

throup
lumpe
of tha
collec
5 in e

The h

 

146

the pre-ELF clitellate slopes were generally quite close to 1.0. Yearly operational
period values, on the other hand, were all below 1.0, with two of them (1989 and
1993) significant at the a = 0.05 level and one (1990) significant at or = 0.01. One
possible hypothesis is that the glandular clitellum makes them more susceptible to
electric current because it is more conductive than the remainder of the integument.
Another is that large worms are more susceptible to electric current than smaller
ones because there can be greater difference in electric potential between the ends
of the worm if it is longer, depending on how it is oriented with respect to the field.

Aporrectodea tuberculaz‘a is a long-lived species, with a possible life span of
seven to eight years (Satchell 1967, referring to A. caliginosa). The model
developed in Chapter 5 corroborates this estimate, and sets the maximum lifespan
at about seven years. Ifthey mature in two years, each worm may be able to
reproduce for up to five years. This results in considerable overlap of generations,
and population-level changes could be slow. It would be interesting and
informative to return to the TEST site after several years to sample for a season or
two to determine if the trend continues.

The total cocoon slopes (Figure 31B) did not vary significantly from 1.0
throughout the study; however, the 95% confidence interval about the 1989 slope
jumped to roughly twice the average of previous years, due to two sampling dates
of that year: one (date 3 in early June) where the actual number of cocoons
collected in the field far exceeded model predictions, and the following month (date
5 in early July; monthly periods overlap by 2 weeks) when the reverse occurred.

The high number of cocoons collected on Date 3 was due to a single sample in

 

which 31
explana'
year of

l
predict:
increasl
usual v
which t
number
individ
produc
fecund

(1)

:‘Qh

_ OD

mm

#nmms‘d‘omg

 

147

which several times the mean number of cocoons was found; a reasonable

explanation of this phenomenon has not been found. After the initial operational

year of 1989, confidence intervals returned to a more usual width.

F ecundity (Figure 31C) again showed no significant differences between

predicted and actual values after antenna activation; however, there was a marked
increase in both the slope and confidence interval in 1989, with a return to more
usual values afterward. The jump was due to two consecutive dates (3 and 4) in
which the fecundity was three to four times the normal level, again due to the large
number of cocoons found in a single sample. This is generally the period when
individuals which have overwintered as aclitellates become clitellate and begin to

produce cocoons. An explanation should be made here about the high slopes in the

fecundity plot. There are two possible reasons for this:

(1) Observed values are based on the number of “new” cocoons (those without a
develop ed embryo) found in the field. Data from the field microcosm and
incubator trials used to construct the models show that infertile cocoons or
those with embryos that died early in development can remain in the soil and
appear viable to external examination for several months. These same data
show that approximately 12% of all cocoons deposited are infertile. This
alone may inflate the number of observed new cocoons by 50% or more.

The method of dividing cocoons into categories used in this study varied

(2)
somewhat from that used in the ELF monitoring project. In the latter,
” “intermediate”, and “old”, the new class

cocoons were divided into “new ,
being cocoons which did not display any development, as well as those which

had a small embryo which was not yet wormlike. They were also observed
after being preserved in formalin, which may have rendered the yolk more

opaque, obscuring small embryos. This study divided cocoons into four
” “embryo”, “hearts”, and “old”, where the first stage was

classes: “new ,
undeveloped and the second contained a recognizable embryo, viewed while
alive. This again could substantially inflate the number of observed cocoons

assumed to be newly deposited.

 

A.
structu
have be
comcm
ngnfic
the m0
operafi
antenn
Iuberc
thm,e
shorte:
clitella
indivit

period

148

Although these tests do not prove conclusively that difi‘erences in population
structure were attributable to an ELF EM field effect, they show that there may
have been an effect due to a time-dependent factor not included in the model which
coincided with activation of the ELF antenna. The number of clitellates was
significantly lower, and the number of cocoons was only somewhat lower than what
the model predicted; as a result, the fecundity of A. tuberculata was higher in the
operational period than in the pre-ELF period, at least for the first year alter
antenna activation. It is interesting to note that the total population of A.
tuberculata was not significantly different in the operational period, suggesting
that, even though there were fewer clitellates (or adults remained clitellate a
shorter time), they produced enough cocoons to make up the difference. Since
clitellates make up only a small fraction of the total population, the loss of a few

individuals did not significantly afl‘ect population levels as a whole during the

period of study.

 

 

(ZDeIui
exanni
natura
follow
antenr
reared
rearet
natura
Uppet
above
Popul
aPPTC

OCtae

tatto.

Penn

 

Chapter 7

SUMMARY AND CONCLUSIONS

Aspects of the life cycle and life history of three lumbricid species
(Dendrobaena octaedra, Lumbricus rubellus, and Aporrectodea tuberculata) were
examined using data collected from three sources: ( 1) biweekly partial censuses of
natural earthworm communities five years before (1984- 1988) and five years
following (1989-1993) activation of an extremely low frequency (ELF) radio
antenna in the vicinity of the sampling sites, (2) replicate marked populations
reared in field microcosms under near-natural conditions, and (3) earthworms
reared in incubators controlled for constant temperature and moisture. Sites for
natural population censuses were located in mixed deciduous forest in Michigan’s
Upper Peninsula. The TEST site was situated approximately 100 m from an
aboveground element of the US. Navy’s ELF antenna, and contained substantial
populations of L. rubellus and A. tuberculata; the CONTROL site was
approximately 11.5 km from the antenna, and contained a high population of D.
octaedra. The field microcosm rearings were performed near the CONTROL site.

In order to obtain sequential records of individuals in the field microcosms, a
tattooing technique was developed and shown to be an effective method of

permanently marking earthworms. Modifications were also made to the time-

149

 

domain
to allox
modifit
gravim
conver
D
were 0
were v
period
betwee
predic'
aspect
based

day 1111

develt
derive
reasoi

Prima.

embq

150

domain reflectometry (TDR) method of nondestructive soil moisture measurement
to allow continuous monitoring of soil moisture within the microcosms; this
modified procedure was validated and compared to a more commonly used
gravimetric method, and a second-order regression equation was developed to
convert values from one to the other.

Dynamic matrix population models driven by soil moisture and temperature
were constructed for each species using the microcosm and incubator data, and
were validated and tested with census information from the five-year pre-ELF
period. Once validated, the A. tuberculata model was used to examine dilferences
between population behavior during the ELF operational period and model
predictions. All three models were utilized to delineate the life cycle and various
aspects of the life history of the three earthworm species in northern Michigan,

based on mean temperature and moisture data over a “typical” year of thirteen 28-

day months, starting on May 1.

Summary of Modelling Techniques and Approach
Cocoon development was modelled using the degree-day approach. Rate of
develoment of cocoons was shown to be directly related to temperature, and
derived degree-day equations for each species described developmental rate
reasonably well. No significant difference between the three species was seen,
primarily because of the degree of developmental variation within each species.
A substantial proportion of cocoons deposited were either infertile, or the

embryos did not develop. Cocoon fertility rate was found to be significantly

pmMm
positive
tubercu

11
inthev
condni
laterin
inpoor
Mnow
Month

.A
tended
gradua
ICpIOd
Separa
rate p(
000001

Stages

Separa
1110 de]

agreet

151

positively correlated with temperature in D. octaedra and L. rubellus, and
positively, although not significantly, correlated with temperature in A.
tuberculata.

Hatchlings of all three species have the highest survival rate if they hatch early
in the warm season (May and June), when there is abundant food in the form of
conditioned and decomposing leaf litter and soil moisture is high. Those hatching
later in the season enter their first winter smaller, less able to burrow, and possibly
in poorer condition to handle the cold stress. Because of this, life cycle and life
history inferences, based on a “typical year” of thirteen 28-day periods, used
Month 1 (May 1) as the starting point for hatchling cohorts.

After hatching, earthworm growth irrespective of temperature and moisture
tended to follow the von Bertalanffy growth equation. Use of assimilated energy
gradually switched from growth in young individuals to maintenance and
reproduction in older, larger earthworms. This growth behavior was utilized to
separate populations of each of the three species into size classes related to growth
rate potential. Populations were also separated by developmental stage into
cocoon, immature, aclitellate (nonreproductive adult) and clitellate (reproductive)
stages.

The models derived for each species treated each size class and stage
separately with respect to soil temperature and moisture. Comparisons of expected
model predictions with observed censuses of natural populations showed close

agreement between observed and expected.

 

l
maturi
micro]
“r-ada
comp:
of 4.5
and p1
positi'
about
accor

woult

medit

C0001

152
Life Cycles and Life Histories of Individual Species
Dena’robaena octaea’ra (S avigny) is a small worm, rarely over 0.15 g at
maturity. It is epigeic and straminicolous, and consumes leaf litter conditioned by

microbes. Satchell’s (1980) classification scheme would place D. octaea’ra in the

 

“r-adapted” category, and the model produced in Chapter 5 confirms it. It is small,
comparatively short-lived, produces many small cocoons (maximum monthly rate
of 4. 5 cocoons per clitellate in its second season), has high mortality early in life,
and probably is only sexually mature for a single summer. Cocoon fertility shows a
positive linear relationship with temperature; cocoons deposited at 5 °C are fertile
about 60% of the time, and cocoons deposited above 17°C are always fertile,

according to the model. A generalized life cycle for this species in upper Michigan

would be:

0 Most cocoons are deposited in mid to late summer. They go through much of
their development during summer and fall, completing it slowly over the winter.
Most cocoons hatch during the month after snowmelt the following spring.

0 Hatchlings grow very quickly but experience high mortality, 90% having died
by the fourth month after hatching. Some reach the clitellate stage the first
year. Of these, approximately 7 5% do not survive the winter. Those that do,
begin producing cocoons shortly after becoming active once again. Those that
overwinter as immatures or aclitellates rapidly grow and become clitellate,
leaving very few one-year-old nonclitellate individuals by early July of the

second year.

a Only about 15% of the worms in their second year survive to reproduce in the
third, and these die over winter. The model and its supportive data give no
indication that clitellates revert to a nonclitellate state when stressed by cold.

L. rubellus Hoffmeister digs shallow, horizontal temporary burrows, is

medium- sized (1 g or less as an adult), produces a moderate to high number of

cocoons (maximum monthly rate of 3.40 cocoons per clitellate at the beginning of

 

 

imthnd
increas
condfin
connnj
confine
chteflat
reprod‘
rdafio:
the tin
for Up

° C0

Prodt
Dercl

T-ada

 

153

its third year) , is straminicolous as an immature and moves more into the soil as it
increases in size, consuming a combination of raw humus and leaf litter it has
conditioned on the surface by burying it with castings (R.W. Parmelee, pers.
com.) This species is in roughly the same place as D. octaedra in the r-K
continuum, growing to maturity just as rapidly but producing fewer cocoons per
clitellate, except at the very end of its life when it seems to spend all of its energy
reproducing. Cocoon fertility rates in this species also show a positive linear
relationship with temperature; cocoons deposited at 5 °C are fertile about 7 0% of
the time; those deposited above 17°C are always fertile. A generalized life cycle

for Upper Michigan follows:

Cocoons are deposited throughout the warm months. Those that are deposited
early and hatch late in the season have a limited chance of survival, as winter
mortality of small worms is high. Those that overwinter as cocoons and hatch
the next spring have a better chance of success, but mortality of spring
hatchlings is still approximately 90% over a period of six months.

Hatchlings grow rapidly; 17% reach the clitellate state by the end of the first
warm season. Of these, about 80% die during the winter, but by the middle of
the warm season of the second year, 42% of all worms from this cohort are

clitellate.

0 Only about 15% of the clitellates alive at the beginning of their second year
survive to the third, but all third-year members of the population are clitellate.
Only a rare individual survives its third winter, but if it does, it produces
cocoons until it dies. As in D. octaedra, the model shows that clitellates retain

the clitellum; only occasionally do they revert to the aclitellate state.
Aporrectodea tuberculata (Eisen) tends more toward being K—adapted. It
produces a few large cocoons per season (maximum monthly rate of 2.75 cocoons

per clitellate in its fourth season), has significantly lower juvenile mortality than the

r-adapted D. octaedra and L. rubellus, finally reaching 10% survival in the fourth

 

 

month t
rarely, :
mineral
for sew
betwee

0000011

0 Cor
Nor
dev
sno
war

° The
few
Clltt
109

' Rep
yea
rep:
m0]
cyc
con

stat
the
larg
con

eac
cyc
i110]
PTO
ach'
pro

the

 

154

month (August) of the second year. It is endogeic, comes to the surface only
rarely, and constructs semipermanent horizontal burrows extending well into the
mineral soil. Adults over 1 g are regularly found, and may be reproductively active
for several seasons. Unlike the preceding two species, no significant relationship
between temperature and cocoon fertility was found; mean fertility of deposited
cocoons was about 88%. A generalized life cycle derived from the model follows:

0 Cocoons are deposited mainly in the last half of the warm season and into
November. They undergo rapid development when the soil is warm, but
development is arrested for three to four months during the winter. After

snowmelt, cocoons rapidly complete their development and hatch early in the
warm season.

0 The first year is spent in slow growth relative to the other two species. Very
few (about 1%) become aclitellate the first year, and there are almost no
clitellates. Juvenile mortality is low compared to the other two species, with
10% of all members of this cohort living until the middle of the second year.

0 Reproduction commences about late July of the warm season during the second
year, with approximately 12% of the total second-year population in
reproductive condition at the peak in mid-September to mid-November. Adult
mortality is low and relatively constant throughout the remainder of the life
cycle. The third year sees 21% of the total population in reproductive
condition, which rises to roughly 25% in the following years.

0 During the warm season, the probability for clitellates to revert to the aclitellate
state is about 40%, depending on temperature, moisture and size according to
the model outlined in Appendix 1. At the same time during the warmest month,
large aclitellates will become clitellate about 60% of the time. It can be
concluded from the summer clitellatezaclitellate stage change probabilities that
A. tuberculata undergoes cyclic reproduction, probably twice or three times
each season, between which it becomes aclitellate to gain energy for the next
cycle. During the winter, the probability for clitellate reversion to aclitellate
increases to approximately 90% and the aclitellate-to-clitellate stage change
probability decreases to about 10%, indicating most mature worms would be
aclitellate . This species also loses reproductive capacity over the winter,
probably because it is costly to retain the clitellum during this time. Maximum
lifespan in a northern deciduous forest is about seven years, with about 0.1% of
the population living to what appears to be a maximum physiological age.

 

 

10 c
rapidly, p
characte:
other hat
reprodut
maximu:
mortalit

the othe

Si
dueto
clitella'
predict
years,
someu
net efl
0f the
affect
Whole
future

earth

 

155

In conclusion, the first two species were similar in that they developed
rapidly, produced many cocoons in a short time, and had a relatively short lifespan
characterized by high mortality early in life. Aporrectodea tuberculata, on the
other hand, developed more slowly, taking two years to reach maturity. Its
reproductive period was extended over several years, and it tended to live to a
maximum physiological age. Absolute juvenile mortality, as well as juvenile

mortality in pr0portion to lifespan, was substantially lower in A. tuberculata than

the other two species.

Effects of ELF Exposure on A. tuberculata

Significant changes in the life cycle and population structure of A. tuberculata
due to exposure to ELF electromagnetic fields were shown to be restricted to
clitellates. A lower proportion of clitellates to total adults was observed than was
predicted by the model (significant at p=0.05 in three of the five ELF operational
years, but remaining lower than predicted throughout), but these clitellates had a
somewhat higher than predicted (nonsignificant at p=0.05) fecundity, balancing the
net effects on the population. It must be concluded that, even though the operation
of the ELF antenna does change the population structure, it does not significantly
aflect the reproductive capacity or the intrinsic growth rate of the population as a
whole. It would be informative to return to the TEST site at some time in the
future to sample for a couple years to test this possibility, censusing the

earthworms to determine if population-level changes have occurred, or if the

 

 

depressi

version 1

Se
models.
mrhMC
develop
cocoon
toleran
hicoco
respon

inform

somer
and La
intoth
intolo
eXperi
QUaht
eXperf

each]

156

depression of clitellate numbers seen in the ELF operational period is an extended

version of the depression of observed vs. expected clitellate numbers seen in 1986.

Directions of Future Research
Several areas can be studied more intensively to increase the accuracy of these
models. The first would be studies to determine a minimum temperature threshold
for hatching. It would then be feasible to subdivide the cocoon stage into new,

developing, and fully-developed cocoons, using the last group as a repository for

 

cocoons that complete development during the winter. More study on cocoon frost
tolerance (sensu Holmstrup et al. 1991) would also allow this factor to be included

in cocoon mortality calculations. Testing the frost tolerance of clitellates and their

 

response to cooling temperatures in the laboratory would also yield usefiil
information for incorporation into the model.

Another factor that would make the model more general is the addition of
some measure of food availability and assimilation over each time period. Martin
and Lavelle (1992) came to similar conclusions, incorporating burrowing behavior
into their model, together with depth- specific soil organic content: worms forced
into lower strata due to drought or heat did not grow as fast, because they
experienced lower food quality in lower soil horizons. Determination of food
quality was attempted as part of this study, but the prop er chemical tests of the
experimental soil and litter (total and soluble N and P, and total C) before and after

each month were prohibitively expensive.

 

Th
earthwo
resp ons
trajecto
graph.
growth
this waj
experie
raised E
grower
estimar

mainta

157

The general form of the model was designed to be extensible to other
earthworm species, given the proper data to construct it. The contour plots of
response surfaces presented in Chapter 5 could be particularly useful, because a
trajectory of mean monthly temperatures and moistures can be plotted on the
graph. If a substantial part of the year is spent in the area where the intrinsic
growth rate, r, is greater than 1.0, the population will remain stable or increase. In
this way, the success or failure of an introduction or a natural population

experiencing environmental stress can be predicted. Applied to a commercially

 

raised species such as Eiseniafetida, application of the model would enable worm
growers to maintain optimum conditions for population growth, allowing an

estimate of the number that can be culled from the population while still

 

maintaining viability.

 

h
tocak:
Inputa

‘

rune:

K

COCOC
develo;

_\

fecundi
fertility
sun/ivor

IMMATI
QFOWth
Sprea
stage
Sprea
Survivor

ACLITE
9th
8Ma
Stage
Sprea
sUfVivor

CLITEL

9'°Wth
Sprea

stage

Sprea
W

 

 

Appendix A

FIELD MICROCOSM, AND COMBINED MODELS

MULTIPLE REGRESSION COEFFICIENTS FOR INCUBATOR,

In the tables below, the “spread” rows contain the equation coefficients used
to calculate standard deviations of the life history character immediately above.
Input and output ranges refer to the coded sizes/stages.

 

Table 23. Multiple regression coeflicients for D. octaedra incubator model.

 

 

 

 

 

 

COCOONS
development - - - - T[a] = -2.7000 Slope= 1746.0
Input Output Soil Soil Tempx Initial
RQQe Range a Temp Moisture Temp2 Moisture2 Moisture Size
fecundity - - - - 1.8640 -0.2233 -0.1316 0.0221
fertility - - - - -0.3418 0.1973 -0.0084
survivorship - - - - 1.0000
IMMATURE WORMS
growth 1 4 1 4 ~0.3268 0.1296 0.0180 -0.0042 -0.2580
spread 1.2620 -0.0527 -0.0528 0.0006 0.0028
stage 3 14 -2.1680 0.2223 0.1113 ~0.0112 -0.0019
spread -5.3240 0.2360 0.3939 -0.0074 -0.0074 -0.0040
survivorship - - 0.8672 -0.0854 0.0022 0.2003
ACLITELLATE WORMS
growth 13 15 13 15 2.0160 0.2544 -0.0101 -0.2259
spread -1.2260 0.0526 0.1466 -0.0036 -0.0036 0.0011
stage 13 15 3 25 0.8823 -0.1192 -0.1494 0.0092 0.0031 0.0842
spread 0.3299 -0.1141 0.0151 -0.0014 0.0061
survivorship - - -1.9170 0.0853 0.1370 -0.0023 -0.0040 0.0682
CLITELLATE WORMS
growth 24 26 24 26 8.4290 0.1552 -0.1419 -0.0075 0.0039 -0.3283
spread -0.8824 0.0198 0.1208 -0.0026
stage 24 26 14 26 -0.3264 0.0216 0.0109 ~0.0009 0.0104
spread 2.2430 -0.0671 -0.1531 0.0025 0.0027
survivorshij 24 26 - - -0.7560 -0.0102 0.0806 -0.0017 0.0312
158

 

Table 2‘

 

00000
developr

*

tecundit
fertility
survivor

IMMATl
growth
sprea
stage
sprea
survive

ACLITE
growth
Sprea
stage
Spree
survive

CLlTE
9”)er
Spre:
stage
spre:
sUlViVI

\

 

159

 

Table 24. Multiple regression coefficients for L. rubellus incubator model.

 

 

COCOONS
development - - - - T[a] = 0.3900 Slope= 1967.0
Input Output Soil Soil Moisture Temp x Initial
Range Range a Temp Moisture Temp2 2 Moisture Size
fecundity - - - - —46.530 0.2204 0.1430 1.6990
fertility - ~ - - 0.3892 0.0364
survivorship - - - - 1.0000

IMMATURE WORMS

growth 1 4 1 4 0.9731 0.0375 -0.1024 0.0023 0.0506
spread -0.3480 0.0497 0.0299 -0.0018

stage 4 4 4 14 -0.5904 0.1494
spread -2.2740 0.1088 0.1408 -0.0048 -0.0021

survivorship 1 4 - - 0.5947 0.0406 -0.0028 0.0678

ACLITELLATE WORMS

growth 14 15 14 15 -3.2000 0.2168 0.2720 -0.0108 -0.0059 -0.0375
spread -1.6190 0.2031 -0.0050

stage 14 15 4 25 -7.5050 0.0341 0.3674 -0.0079 0.2184
spread -5.3110 -0.0794 0.5405 0.0042 -0.0120 0.0001
survivorship 14 15 - - 0.7977 0.0118
CLITELLATE WORMS

growth 24 26 24 26 13.600 -0.0349 0.0072 -0.3829
spread -0.1949 0.0111 0.0036
stage 24 26 14 26 -3.0180 0.0278 0.7081 -0.0158 -0.1939
spread 2.2590 -0.0404 -0.1722 0.0010 0.0039 -0.0003
survivorship 24 26 - - -1.0730 0.0746

 

 

Table 2

COCOC
develop

 

fecundi
fertility
survivor

IMMAT

growth
sprea

stage
sprea

survive

ACLIT
growth
Spre:
stage
Spre:
Survive

CLITE
growtl
Spre
stage
Spre
surviv

\

160

 

 

Table 25. Multiple regression coefficients for A. tuberculata incubator model.

 

 

 

COCOONS
development - - - - T[a] = 0.59 Slope= 1037
Input Output Soil Soil Tempx Initial
Range Range :3 Temp Moisture Tempz Moistt_r_re2 Moisture Size
fecundity - - - - -2.73 -0.16 -0.065 0.0094 0.154
fertility - - - - 0.6483 0.0106
survivorship - - - - 1
IMMATURE WORMS
-0.0967
growth 1 6 1 6 -1.216 0.0336 0.1077 -0.002 4
spread 1.818 0.0006 -0.134 0.0032
stage 4 6 4 16 -1.993 0.0403 0.0348 -0.002 -7e-04 0.2858
spread -2.589 0.0903 0.1923 -0.005 -0.004 0.0007
survivorship 1 6 0.5741 0.0632 -0.004 0.03599
ACLITELLATE WORMS
growth 15 17 15 17 3.244 0.0215 0.0304 -0.2597
spread 0.872 -0.029 -0.094 0.0027 0.0027
stage 15 17 5 27 -0.85 -0.224 -0.055 0.0041 0.0074 0.1402
spread -1.684 0.0273 0.141 -0.002 -0.003 0.0018
survivorship 15 17 -0.143 0.0379 -0.004 -0.003 0.07066
CLITELLATE WORMS
growth 26 28 26 28 5.221 0.1557 0.0331 -0.007 -0.2583
spread 4.06 -0.209 -0.259 0.0057 0.0051 0.0041
stage 26 28 16 28 -4.222 0.0056 0.3292 -0.005 -0.008 0.0054
spread 2.352 -0.015 -0.159 0.0033 0.0037 -0.003
survivorship 26 28 1.35 0.0547 -0.048 -0.004 0.0011

 

 

Table 2

COCOt
develo;

 

fecundi
fertility
survivo

IMMAT
growth
spree
stage
spre:
survive

ACLIT
QIOth
spre
stage
spre
surviv

CLITE
QIOWt
Spit
stage
Spre
SUIViI.

\

161

 

 

Table 26. Multiple regression coeflicients for D. octaedra field microcosm model.

 

 

 

COCOONS
development - - - - T[a] = -1.3600 Slope= 1857.0
Input Output Soil Soil Temp x Initial
Rgnge Rm fia Temp Moisture Temp2 Moisture2 Moisttire Size
fecundity - - - - -56.960 0.4522 -0.1509 2.3530
fertility - - - - 0.7764 0.0116
survivorship - - - - 0.9946
IMMATURE WORMS
growth 1 4 1 4 -4.3830 0.7997 -0.0305 -0.0995
spread 0.0496 0.0866
stage 3 4 3 14 -4.2490 0.0551 0.2100 -0.0047 0.4820
spread -0.4168 0.1734
survivorship 1 4 - - -0.2767 0.0193 0.1534
ACLITELLATE WORMS
growth 13 15 13 15 -8.8520 2.3170 -0.0910 -0.3341
spread -1.1010 0.0609 0.0325
stage 13 15 3 25 -5.8340 0.1102 0.1285 -0.0030 0.2454
spread 0.1898
survivorship 13 15 - - 0.8988
CLITELLATE WORMS
growth 24 26 24 26 7.2120 1.4480 -0.2238 -0.0558 0.0038 -0.5205
spread 0.3000
stage 24 26 14 26 -0.3179 0.0122
spread 0.8878 -0.0054 -0.0314
survivorship 24 26 - - -0.0224 0.1424 -0.0052

 

 

 

Table I

COCOt
develo;

 

fecund
fertility
survivo

IMMAT
growth
spre:
stage
spre;
surviv:

ACLIT
QIOWtI
spre
stage
spre
surviv

CLITE
QIOWt'
Spre
stage
spre
surviv

\

162

 

 

Table 27. Multiple regression coefficients for L. rubellus field microcosm model.

 

 

 

COCOONS
development - T[a] = 0.1100 Slope= 1675.0
Output Soil Soil Temp x Initial
Range 3 Temp Moisture Temp2 Moisture2 Moisture Size
fecundity - -30.920 -2.4210 -0.7853 0.0189 0.0010 0.0846 2.0640
fertility - 0.9073
survivorship - 0.9876
IMMATURE WORMS
growth 4 -3.6900 0.5454 0.0306 -0.0201 -0.0754
spread -0.2814 0.0265 0.0929
stage 14 -5.4130 0.7513 0.0196 -0.0292 0.1414
spread 3.9340 1.3430 -1.0940 0.0387
survivorship - -2.9770 0.1470 0.1849 -0.0022 -0.0056 0.1421
ACLITELLATE WORMS
growth 14 15 14 15 -8.9350 1.6890 -0.1175 -0.0388 0.0090 -0.0265
spread -31.281 3.0490 -0.0210 -0.2981 -0.0457 0.1890 1.0650
stage 25 -9.5270 0.4232 0.1970 -0.0160 0.3279
spread -20.131 0.9264 -0.0182 0.6122
survivorship - 2.4090 0.2993 -0.0112 0.0926
CLITELLATE WORMS
growth 24 26 24 26 2.7990 0.4166 -0.0164 -0.2056
spread 7.2840 0.0594 -0.3220
stage 24 26 14 26 -1.1020 0.1643 -0.0061
spread 0.3000
survivorship - -4.9470 0.0177 0.2157 -0.0042 0.1113

 

 

 

Table 2
model.

COCOC
develop

 

fecundi
fertility
survivo

IMMAT
growth
sprea
stage
spre:
Survive

ACLIT
growtt
spre
stage
spre
surviv

CLITE
QIOWI
SprE
stage
Spre
Survit

\

163

 

Table 28. Multiple regression coeflicients for A. tuberculata field microcosm

 

 

 

 

model.
COCOONS
deveIOpment - - - - T[a]=2.6030 Slope= 746.10
Input Output Soil Soil Tempx Initial
Remge Ra_nge a Temp Moistui Temp2 Moisture2 Moisture Size
fecundity - - - - -8.0880 1.5130 -0.0552
fertility - - - - 1.0120 -0.0069
survivorship - - - - 0.9863
IMMATURE WORMS
growth 1 6 1 6 0.8585 -0.0230
spread 0.3844
stage 4 6 4 16 4.2760 -0.2991 0.0052
spread 4.2312 -0.2923 0.0050
survivorship 1 6 -0.6084 0.1224 0.0441 -0.0051 0.1116
ACLITELLATE WORMS
growth 15 17 15 17 -1.1970 0.6198 -0.1905 -0.0262 0.0042
spread 0.3095
stage 15 17 5 27 0.0572 -0.0123 0.0022 0.1157
spread -1.1260 -0.0157 0.1086
survivorship 15 17 -0.3364 0.1879 -0.0083 0.0182
CLITELLATE WORMS
growth 26 28 26 28 0.4256
spread 0.4352
stage 26 28 16 28 4.1530 -0.1459
spread 4.3060 -0.1515
survivorship 26 28 2.1370 -0.0799 -0.0454 0.0032

 

 

Table 2
microo

COCOC
develop

 

fecundi
fertility
survivor

IMMAT
1-2
growth
sprea
stage
sprea
survivo

IMMAT
3-4

growth
Spree

stage
Spree

survive

AC LIT
91°th
Spre
stage
Spre
surviv.

CLITE
9'0th

spre
stage

Spre
Surviv

\

 

164

 

Table 29. Multiple regression coefficients for D. octaedra combined incubator and

 

 

 

microcosm model.
COCOONS
development - - - - T[a] = -6.8800 Slope= 2394.9
Input Output Soil Soil Moisture Tempx Initial
Ra_nge Range g Temp Moisture Temp2 2 Moistgr_e Size
fecundity - - - - -38.481 -0.8422 0.6710 0.0267 -0.0192 0.0257 1.3870
fertility - - — - 0.4527 0.0325
survivorship - - - - 0.9990
IMMATURE WORMS classes
1-2
growth 1 2 1 3 0.2832 0.0442 -0.1818
spread 2.3903 -0.1752 -0.0737 0.0058
stage - - - -
spread
survivorship 1 2 - - -0.4456 0.0181 0.2741
IMMATURE WORMS classes
3-4
growth 3 4 2 4 -0.4633 0.2231 -0.0078 -0.2120
spread 0.3857
stage 3 4 3 14 -2.7603 0.0266 0.1320 -0.0029 0.3769
spread -0.3588 0.1661
survivorship 3 4 - - 0.7677
ACLITELLATE WORMS
growth 13 15 13 15 2.1609 0.4794 -0.0186 -o_3145
spread 0.4898
stage 13 15 3 25 -2.9750 -0.0634 0.0160 0.0062 0.2520
spread -1.1674 0.0958
survivorship 13 15 - - 0.9047
CLITELLATE WORMS
growth 24 26 24 26 11.6610 0.2274 -0.0236 -0.0084 -0.5036
spread 0.3298
stage 24 26 14 26 -0.4248 0.0165
0.1.159
spread 9 -0.0451
24 26 - - 0.7761 0.0102

survivorship

 

IF

Table I
microe

COCO!
develol

 

fecund
fertility
survivo

IMMAT
growth
spree
stage
sprea
survivc

ACLIT
QIOWIh
spre:
stage
spre:
SUIVch

CLITE
growtr
spre:
stage
spre:
survive

\

 

 

Table 30. Multiple regression coefficients for L. rubellus combined incubator and

microcosm model.

 

 

COCOONS
development - - - - T[a] = -2.2100 Slope= 2271.6
Input Output Soil Soil Moisture Tempx Initial
Range Range a Temp Moisture Temp2 2 Moisture Size

fecundity - - - - -32.520 -0.5447 -0.1351 0.0298 1.4010
fertility - - - - 0.5866 0.0241

survivorship - - - — 0.9983

IMMATURE WORMS

growth 1 4 1 4 -0.9114 0.1536 0.0164 -0.0048 -0.0853
spread 0.1396 0.0638
stage 4 4 4 14 1.3785 —0.0086 -0.0969 -0.0083 0.0090

spread 0.1905

survivorship 1 4 - - 0.5171 -0.0137 0.1382
ACLITELLATE WORMS

growth 14 15 14 15 -1.5020 0.3268 -0.0135

spread 0.3135

stage 14 15 4 25 -8.5967 0.0290 0.2556 -0.0051 0.4029
spread -4.7103 0.3106
survivorship 14 15 - - -0.4902 0.0965
CLITELLATE WORMS

growth 24 26 24 26 5.9323 -0.2339
spread 0.4275

stage 24 26 14 26 -1.3557 0.0162 0.0966 -0.0020

spread 1.7940 0.0646 -0.1710 -0.0042 0.0037

survivorship 24 26 - - -2.3118 0.1241

 

IF

 

Table
and m

COCO
develo

 

fecunc
fertility
survive

IMMAT
growtt
spre:
stage
spre;
survivr

IMMAI
growtt
spre
stage
spre
survivI

IMMA‘
QIOWtI
spre
stage
Spre
SUI’Vin

ACLIT
9'0th
Spre
stage
Spre
surviv.

CLITE
9'0th

Spre
Stage

spre

sUrviv

\

 

 

166

 

Table 31. Multiple regression coefficients for A. tuberculata combined incubator

and microcosm model.

 

 

COCOONS
development - - - - T[a] = 0.5500 Slope= 938.64
Input Output Soil Soil Tempx Initial
Rme Range g Temp Moisture Temp2 Moisture2 Moisture Size
fecundity - - - - -15.803 0.0841 0.5995
fertility - - - - 0.8814
survivorship - - - - 0.9968
IMMATURE WORMS class 1
growth 1 1 1 2 -7.4285 1.4330 -0.0632
spread 0.4751
stage - - - -
spread
survivorship 1 6 0.3167 0.1968 -0.0109
IMMATURE WORMS classes 2-4
growth 2 4 1 5 -1.9768 0.0427 0.1553 -0.0032
spread 0.4005
stage - - - -
spread
survivorship 2 4 0.9650 -0.0189 0.1065
IMMATURE WORMS classes 5-6
growth 5 6 4 6 2.3081 0.3667 -0.0146 -0.9297
spread 0.3927
stage 5 6 5 16 -1.4909 0.0166 0.2748
spread 0.1238
survivorship 5 6 0.9548
ACLITELLATE WORMS
growth 15 17 15 17 1.7030 0.0243 -0.1291
spread -1.2255 0.0963
stage 15 17 5 27 -2.6167 0.0188 0.1602
spread 0.2432
survivorship 15 17 -0.3991 0.0126 -0.0273 -0.0042 0.0025 0.0605
CLITELLATE WORMS
growth 26 28 26 28 1.6522 0.1586 0.3289 -0.0144 ~0.0079 0.0077 -0.2429
spread -0.0997 0.0189
stage 26 28 16 28 -2.9088 0.0067 0.2214 -0.0072 -0.0059 0.0072
spread 2.9655 -0.1022
-0.025
survivorship 26 28 1.5007 4 -0.0256 -0.0012 0.0020

 

 

Table 3

D. octt

 

Appendix B

MODEL-GENERATED MONTHLY POPULATION STRUCTURES

 

Table 32. Monthly changes in modelled population structure of a cohort of class 1
D. octaedra, starting on May 1 (day 1, month 1) of a typical year.

 

 

 

Temper- Moisture Imma- Aclitel- Clitel- Popu- Cocoons per
Year Month ature (°C) (%) Cocoons tures Iates Iates Iation clitellate
0 10000 0 0 10000 ---
1 1 10 33 0 4255 0 0 4255 ---
1 2 13 30 0 2247 0 0 2247 ---
1 3 15 24 0 1296 32 0 1328 ---
1 4 16 25 0 773 189 23 985 0.00
1 5 15 25 70 393 239 150 782 0.47
1 6 11 30 144 237 236 178 651 0.81
1 7 8 25 109 129 228 189 546 0.58
1 8 5 25 55 80 201 181 462 0.30 .
1 9 1 25 22 55 170 158 383 0.14
1 10 0 25 1 34 153 124 311 0.01
1 11 0 25 0 18 138 96 252 0.00
1 12 1 30 0 11 124 75 210 0.00
1 13 3 33 0 13 107 60 180 0.00
2 1 10 33 0 10 96 52 158 0.00
2 2 13 30 53 7 82 53 142 1.00
2 3 15 24 155 3 14 110 127 1.41
2 4 16 25 466 2 1 115 118 4.05
2 5 15 25 412 1 0 107 108 3.85
2 6 11 30 123 0 0 95 95 1.29
2 7 8 25 63 0 0 81 81 0.78
2 8 5 25 24 0 0 67 67 0.36
2 9 1 25 8 0 0 52 52 0.15
2 10 0 25 0 0 0 40 40 0.00
2 11 0 25 0 0 0 31 31 A 0.00
2 12 1 30 0 0 0 24 24 0.00
2 13 3 33 0 0 0 19 19 0.00

167

 

Table

r
a
e
V1

 

33.533333333334444.44444444444\

Table 32 (cont’d).

 

 

 

 

Temper- Moisture Imma- Aclitel- Clitel- Popu— Cocoons per
Year Month ature(°C) (%) Cocoons tures lates lates Iation clitellate
3 1 10 33 0 0 0 17 17 0.00
3 2 13 30 17 0 0 15 15 1.13
3 3 15 24 44 0 0 15 15 2.93
3 4 16 25 63 O 0 14 14 4.50
3 5 15 25 50 0 0 13 13 3.85
3 6 11 30 15 0 0 12 12 1.25
3 7 8 25 9 0 0 10 10 0.90
3 8 5 25 3 0 0 8 8 0.38
3 9 1 25 1 0 0 6 6 0.17
3 10 0 25 0 0 0 5 5 0.00
3 11 0 25 0 0 0 4 4 0.00
3 12 1 30 0 0 0 3 3 0.00
3 13 3 33 O 0 0 2 2 0.00
4 1 10 33 0 0 0 1 1 0.00
4 2 13 30 0 0 0 1 1 0.00
4 3 15 24 0 O 0 1 1 0.00
4 4 16 25 0 0 0 1 1 0.00
4 5 15 25 0 0 0 1 1 0.00
4 6 11 30 0 0 0 1 1 0.00
4 7 8 25 0 0 0 1 1 0.00
4 8 5 25 0 0 0 1 1 0.00
4 9 1 25 0 O 0 1 1 0.00
4 10 0 25 0 0 0 1 1 0.00
4 11 0 25 0 0 0 1 1 0.00
4 12 1 30 0 0 0 1 1 0.00
4 13 3 33 0 0 0 1 1 0.00

 

Table 3'
L.rube

Year N

 

\3333333333333

169

 

Table 33. Monthly changes in modelled population structure of a cohort of class 1
L. rubellus, starting on May 1 (day 1, month 1) of a typical year.

 

 

 

 

Temper- Moisture Imma- Aclitel- Clitel- Popu- Cocoons per
Year Month ature (°C) (%) Cocoons tures lates lates Iation clitellate
0 10000 0 0 10000

1 1 9.5 33 0 5248 0 0 5248 ---

1 2 12.5 30 0 3007 0 0 3007 ---

1 3 15 24 0 1925 0 0 1925 ---

1 4 16 25 0 1311 49 0 1360 ---

1 5 15 25 0 920 110 36 1066 0.00
1 6 11 30 48 688 165 71 924 0.68
1 7 8 25 33 615 82 143 840 0.23
1 8 5 25 28 592 44 140 776 0.20
1 9 1 25 12 571 32 116 719 0.10
1 10 0 25 10 521 29 91 641 0.11
1 11 0 25 9 428 25 73 526 0.12
1 12 1 30 3 313 21 60 394 0.05
1 13 3 33 12 213 26 41 280 0.29
2 1 9.5 33 92 127 24 32 183 2.88
2 2 12.5 30 96 80 12 40 132 2.40
2 3 15 24 80 54 4 42 100 1.90
2 4 16 25 104 38 3 37 78 2.81
2 5 15 25 83 27 4 32 63 2.59
2 6 11 30 70 20 6 27 53 2.59
2 7 8 25 25 18 3 26 47 0.96
2 8 5 25 12 17 1 22 40 0.55
2 9 1 25 4 17 1 18 36 0.22
2 10 0 25 2 16 1 14 31 0.14
2 11 0 25 2 13 1 12 26 0.17
2 12 1 30 1 9 1 10 20 0.10
2 13 3 33 3 6 2 7 15 0.43
3 1 9.5 33 17 4 2 5 11 3.40
3 2 12.5 30 16 2 0 5 7 3.20
3 3 15 24 13 1 0 4 5 3.25
3 4 16 25 12 0 0 4 4 3.00
3 5 15 25 11 0 0 4 4 2.75
3 6 11 30 10 0 0 3 3 3.33
3 7 8 25 4 0 0 3 3 1.33
3 a 5 25 2 0 o 3 3 0.67
3 9 1 25 1 0 0 3 3 0.33
3 10 0 25 1 0 0 3 3 0.33
3 11 0 25 1 0 0 2 2 0.50
3 12 1 30 0 0 0 2 2 0.00
3 13 3 33 1 O 0 2 2 0.50

 

Table

6
VI

4444444444444

 

170

Table 33 (cont’d).

 

 

Temper- Moisture Imma— Aclitel— Clitel- Popu- Cocoons per
Year Month ature(°C) (%) Cocoons tures lates lates Iation clitellate
4 1 9.5 33 8 0 0 2 2 4.00
4 2 12.5 30 10 O 0 2 2 5.00
4 3 15 24 8 0 0 1 1 8.00
4 4 16 25 0 O 0 1 1 0.00
4 5 15 25 0 0 0 1 1 0.00
4 6 11 30 0 0 0 1 1 0.00
4 7 8 25 0 0 0 1 1 0.00
4 8 5 25 0 0 0 1 1 0.00
4 9 1 25 0 0 0 1 1 0.00
4 10 0 25 0 0 0 1 1 0.00
4 11 0 25 0 0 0 1 1 0.00
4 12 1 30 0 0 0 1 1 0.00
4 13 3 33 0 0 0 1 1 0.00

 

 

Table 3

A. tube

 

1 Al 1 1 1 Al Al Al Al Al 1 4| 1 2 2 2 2 2 2 2 2 2 2 2 2 2\

 

3333333333333

171

 

Table 34. Monthly changes in modelled population structure of a cohort of class 1
A. tuberculata, starting on May 1 (day 1, month 1) of a typical year.

 

 

 

 

Temper- Moisture lmma- Aclitel- Clitel- Popu- Cocoons per
Year Month ature (°C) (%) Cocoons tures lates lates Iation clitellate
0 10000 0 0 10000

1 1 10 33 0 9108 0 0 9108 ---

1 2 13 30 0 8118 0 0 8118 ---

1 3 15 24 0 6168 0 0 6168 ---

1 4 16 25 0 4630 0 0 4630 ---

1 5 15 25 0 3865 1 0 3866 ---

1 6 11 30 0 3289 10 0 3299 ---

1 7 8 25 0 2869 26 1 2896 0.00
1 8 5 25 0 p 2565 29 1 2595 0.00
1 9 1 25 0 2194 23 1 2218 0.00
1 10 0 25 0 1890 16 1 1907 0.00
1 11 0 25 0 1680 12 0 1693 0.00
1 12 1 30 0 1474 8 0 1482 0.00
1 13 3 33 0 1301 6 0 1308 0.00
2 1 10 33 0 1176 7 1 1184 0.00
2 2 13 30 2 1044 48 2 1095 0.93
2 3 15 24 3 885 132 10 1027 0.30
2 4 16 25 16 747 178 30 954 0.53
2 5 15 25 45 659 179 56 894 0.80
2 6 11 30 72 600 182 75 857 0.96
2 7 8 25 85 558 178 90 827 0.94
2 8 5 25 85 531 169 92 792 0.93
2 9 1 25 45 508 179 33 720 1.36
2 10 0 25 8 486 149 7 642 1.21
2 11 0 25 0 465 107 5 578 0.00
2 12 1 30 0 444 71 4 519 0.00
2 13 3 33 0 423 53 4 480 0.00
3 1 10 33 2 401 45 11 458 0.18
3 2 13 30 12 366 52 21 439 0.58
3 3 15 24 34 310 75 32 417 1.06
3 4 16 25 56 260 84 44 388 1.26
3 5 15 25 73 230 80 57 366 1.29
3 6 11 30 80 211 78 64 353 1.25
3 7 8 25 79 196 75 71 341 1.12
3 8 5 25 72 187 72 69 328 1.04
3 9 1 25 39 178 91 27 297 1.43
3 10 0 25 7 171 85 4 259 1.95
3 11 0 25 0 164 61 3 229 0.00
3 12 1 30 0 157 41 2 200 0.00
3 13 3 33 0 150 29 2 182 0.00

 

Table

 

r

 

6666666666666

172

Table 34 (cont’d).

 

 

 

 

Temper- Moisture Imma- Aclitel- Clitel- Popu- Cocoons per
Year Month ature(°C) (%) Cocoons tgres lates lates Iation clitellate
4 1 10 33 1 143 25 6 174 0.17
4 2 13 30 7 130 26 11 167 0.61
4 3 15 24 18 110 31 16 157 1.09
4 4 16 25 30 93 33 22 147 1.35
4 5 15 25 37 82 30 27 139 1 .38
4 6 11 30 39 75 29 30 134 1.31
4 7 8 25 38 70 28 32 130 1.17
4 8 5 25 33 66 27 31 124 1.06
4 9 1 25 18 63 36 13 112 1.42
4 10 0 25 4 60 36 1 97 2.75
4 11 0 25 0 58 25 1 84 0.00
4 12 1 30 0 56 17 1 74 0.00
4 13 3 33 0 54 13 1 67 0.00
5 1 10 33 0 51 10 3 64 0.00
5 2 13 30 2 47 11 4 62 0.46
5 3 15 24 7 40 12 6 59 1.08
5 4 16 25 11 34 12 8 54 1.34
5 5 15 25 14 30 11 10 51 1.41
5 6 11 30 14 27 11 11 48 1.24
5 7 8 25 15 25 10 12 48 1.25
5 8 5 25 12 24 10 11 45 1.10
5 9 1 25 6 23 13 4 41 1.40
5 10 0 25 1 22 14 1 36 1.74
5 11 0 25 0 21 10 1 31 0.00
5 12 1 30 0 20 7 0 28 0.00
5 13 3 33 0 19 5 0 24 0.00
6 1 10 33 0 18 4 1 23 0.00
6 2 13 30 2 16 5 2 23 0.97
6 3 15 24 3 14 4 3 21 1.04
6 4 16 25 6 12 4 4 20 1.62
6 5 15 25 5 11 4 4 18 1.35
6 6 11 30 6 9 4 4 17 1.34
6 7 8 25 6 8 4 4 17 1.38
6 8 5 25 4 8 4 4 15 1.01
6 9 1 25 3 7 5 2 13 1.50
6 10 0 25 0 7 4 0 11 0.00
6 11 0 25 0 7 3 0 10 0.00
6 12 1 30 0 7 3 0 9 0.00
6 13 3 33 0 7 1 0 8 0.00

'Tablei

Year

 

 

 

——.--—--- -.—s—u~—n.—g.‘_=' "' '.‘ .A. 4.- ’7 he,» \W)~ » .,_.. .-

173

Table 34 (cont’d).

 

 

Temper- Moisture lmma- Aclitel- Clitel- Popu- Cocoons per
Year Month Ange (°C) (%) Cocoons tures lates lates Iation clitellate
7 1 10 33 0 7 1 0 8 0.00
7 2 13 30 0 6 1 0 8 0.00
7 3 15 24 0 5 1 0 7 0.00
7 4 16 25 0 4 2 0 6 0.00
7 5 15 25 0 4 2 0 6 0.00
7 6 11 30 0 3 2 0 5 0.00
7 7 8 25 0 2 2 0 4 0.00
7 8 5 25 0 1 2 0 3 0.00
7 9 1 25 0 1 1 0 2 0.00
7 10 0 25 0 1 0 0 1 0.00
7 11 0 25 0 1 0 0 1 0.00
7 12 1 30 0 1 0 0 1 0.00
7 13 3 33 0 1 0 0 1 0.00

 

 

Tabl
temI

 

aaaaaaaaaaaaaaaaaaaaaa

 

Appendix C

DATA SUMMARIES FOR INCUBATOR AND FIELD MICROCOSM

STUDIES

 

Table 35. L. rubellus incubator cocoon development summary for each of five

 

 

temperatures.
3°C 6°C 9°C 12°C 15°C

DAY SCORE DAY SCORE DAY SCORE DAY SCORE DAY SCORE
0 0.00 0 0.00 0 0.00 O 0.00 0 0.00
31 0.00 31 0.00 28 0.00 31 0.11 31 0.50
46 0.00 48 0.00 42 0.20 49 1.22 53 2.22
63 0.00 62 0.00 54 0.40 64 1.67 63 2.50
77 0.00 74 0.00 63 0.80 76 2.11 74 2.80
89 0.00 83 0.00 75 1.20 83 2.44 83 3.00
98 0.00 95 0.00 83 1.20 95 2.78 95 3.10
110 0.11 102 0.00 95 2.00 102 2.89 109 3.30
129 0.22 114 0.00 109 2.00 114 3.00 123 3.60
144 0.22 129 0.00 122 2.20 128 3.56 137 3.70
157 0.22 152 0.80 136 2.60 142 3.67 147 3.70
171 0.33 166 1.20 150 2.80 152 3.67 161 3.80
199 0.67 180 1.80 160 3.20 166 3.78 175 3.90
209 1.00 193 2.20 174 3.40 180 3.89 186 4.00
223 1.33 211 2.40 188 3.80 211 4.00

237 1.67 219 2.80 201 3.80

250 1.67 232 3.00 219 4.00

268 2.22 246 3.20

275 2.33 267 4.00

289 2.33

303 2.56

324 2.67

338 3.11

352 3.33

366 3.44

380 3.67

394 3.78

408 4.00

 

 

 

 

 

174

Tabll

temp

n l A
A/«Irk (1111111122224

 

175

 

Table 36. D. octaedra incubator cocoon development summary for each of five

 

 

temperatures.
3°C 6°C 9°C 12°C 15°C
DAY SCORE DAY SCORE DAY SCORE DAY SCORE DAY SCORE
0 0.00 0 0.00 O 0.00 O 0.00 0 0.00
31 0.00 31 0.00 29 0.88 29 1.14 31 1.33
48 0.00 48 0.25 42 1.14 42 1.67 51 2.29
62 0.00 62 0.50 54 1.38 52 2.00 63 2.86
74 0.50 74 1.00 63 1.63 66 2.33 74 3.14
89 1.50 89 1.50 75 1.88 76 2.67 88 4.00
98 1.50 98 2.00 84 2.00 90 3.11
112 2.00 112 2.50 98 2.33 110 4.00
126 2.00 126 2.75 108 2.67
139 2.00 135 3.00 122 3.11
157 3.00 143 3.44 142 4.00
178 3.00 156 4.00
192 3.50
213 4.00

 

 

 

 

 

 

Table 37. A. tuberculata incubator cocoon development summary for each of five

 

 

temperatures.
3°C 6°C 9°C 12°C 15°C

DAY SCORE DAY SCORE DAY SCORE DAY SCORE DAY SCORE
0 0.00 0 0.00 0 0.00 0 0.00 0 0.00
28 0.00 28 0.00 14 1.00 14 0.00 15 0.50
46 0.38 42 0.25 26 2.00 26 0.33 31 2.00
63 0.50 54 0.50 35 2.25 35 1.00 53 2.00
77 0.63 63 0.75 47 2.75 47 2.33 63 2.75
96 0.75 82 1.25 54 3.00 54 2.67 74 3.88
110 0.88 109 1.25 66 3.25 66 3.00 83 4.00
124 1.13 122 1.25 81 3.50 81 3.33

140 1.25 136 1.25 94 3.75 94 3.67

154 1.75 150 1.25 108 4.00 108 4.00

164 2.25 discontinued

178 2.38

192 3.13

205 3.25

223 3.75

244 3.88

258 3.88

279 4.00

 

 

 

 

 

176

 

Table 38. D. octaedra incubator worm summary by developmental stage.

 

 

TEMP MOIST BEGINNING ENDING STAGES
(°C) (%) STAG N IMM ACL CLI DIED
E
3 22.5 IMM 54 33 3 o 18
ACL 14 0 12 1 1
CLI 27 o o 27 o
5 17.2 IMM 54 28 1 o 25
ACL 25 0 21 1 3
CLI 30 o 2 21 7
5 27.8 IMM 52 39 2 o 11
ACL 22 o 19 1 2
CLI 28 o o 27 1
10 15 IMM 42 27 3 o 12
ACL 46 1 31 5 9
CLI 4 0 o 3 1
10 22.5 IMM 39 20 4 0 15
ACL 45 o 37 6 2
CLI 24 o o 21 3
10 3o IMM 32 16 9 o 7
ACL 39 o 28 9 2
CLI 10 o o 8 2
15 17.2 IMM 29 12 2 0 15
ACL 3 0 o 3 o
CLI 22 o o 20 2
15 27.8 IMM 28 2o 2 0 6
ACL 6 0 2 2 2
CLI 24 o 0 2o 4
17 22.5 IMM 47 25 3 o 19
ACL 8 o 2 o
CLI 42 o 0 34 8

 

 

 

IMM = immature, ACL = aclitellate, CLI = clitellate. .
Ending stages are the fates of each ofthe N worms at the end ofa 28-day perlod.

in

177

 

Table 39. L. rubellus incubator worm summary by developmental stage.

 

 

TEMP MOIST BEGINNING ENDING STAGES
(°C) (%) STAG N IMM ACL CLI DIED
E
3 22.5 IMM 69 63 0 o 6
ACL 9 3 4 o
CLI 13 o o 10 3
5 17.2 IMM 66 6O 0 o 6
ACL 5 o 5 o o
CLI 22 0 2 20 o
5 27.8 IMM 62 53 o 0 9
ACL 4 o 4 0 0
CLI 23 o 2 16 5
10 15 IMM 64 53 3 o 8
ACL 2o 6 12 1 1
CLI 7 o 3 2 2
10 22.5 IMM 77 69 3 o 5
ACL 5 o 3 2 0
CLI 14 o o 14 o
10 30 IMM 72 65 0 0 7
ACL 17 o 15 2 o
CLI 7 o 4 3 0
15 17.2 IMM 71 63 o o 8
ACL 12 o 9 2 1
CLI 10 o o 10 o
15 27.8 IMM 95 82 1 o 12
ACL 10 8 2 0
CLI 17 o o 10 7
17 22.5 IMM 28 12 0 0 16
ACL 4 0 1 2 1
CLI 5 o o o 5

 

 

 

IMM = immature, ACL = aclitellate, CLI = clitellate.
Ending stages are the fates of each ofthe N worms at the end ofa 28-day period.

t

 

 

Table 40. A. tuberculata incubator worm summary by developmental stage.

 

 

TEMP MOIST BEGINNING ENDING STAGES
(°C) (%) STAG N IMM ACL CLI DIED
E
3 22.5 IMM 57 48 o 0 9
ACL 10 o 10 o o
CLI 20 0 7 13 o
5 17.2 IMM 72 71 o o 1
ACL 4 o 4 o 0
CLI 20 o 13 7 o
5 27.8 IMM 68 66 2 o o
ACL 12 o 12 o o
CLI 17 o 13 4 o
10 15 IMM 61 55 6 0 o
ACL 29 5 24 0 o
CLI 6 o 4 2 o
10 22.5 IMM 97 89 7 o 1
ACL 80 6 69 3 2
CLI 15 o 1 14 0
10 3o IMM 61 50 9 o 2
ACL 23 o 19 2 2
CLI 10 o 2 8 o
15 17.2 IMM 63 59 0 o 4
ACL 36 3 28 1 4
CLI 17 o 5 12 o
15 27.8 IMM 53 48 0 0 5
ACL 12 0 5 7 o
CLI 22 0 2 20 o
17 22.5 IMM 103 65 2 0 36
ACL 43 o 17 11 15
CLI 21 o 1 10 10

 

 

 

IMM = immature, ACL = aclitellate, CLI = clitellate.

Ending stages are the fates of each ofthe N worms at the end ofa 28-day period.

 

 

Tabl
tem;
with
with

TE

179

Tables 41 to 43: N DAYS = days elapsed since the previous sampling date;
temperature and moisture are means over this period. Each cocoon series begins
with a boldface capital letter and the number of cocoons in parentheses, and ends
with a horizontal line. The numbers are mean developmental codes.

 

Table 4 1. D. octaedra field microcosm summary by date.

 

 

 

 

 

 

 

 

N TEMPMOIST BEGINNING ENDING STAGES
DATEDAYS (°C) (%) STAGE N IMM ACL CLI DIED COCOON SERIES
1 183 1.5 IMM 161 41 0 o 120
ACL 11 3 3 1 4
CLI 4 0 1 1 2
2 27 11.1 19.9 IMM 44 12 5 o 27
ACL 4 0 2 1 1
CLI 2 o 0 0 2
3 28 11.9 25.4 IMM 53 32 9 0 12
ACL 20 0 15 1 4
CLI 1 o 0 o 1 A(90)
4 28 13.3 22.8 IMM 32 17 0 0 15 0
ACL 24 0 9 13 2
CLI 3 0 0 3 o 3197)
5 28 13.8 28.9 IMM 17 6 4 0 7 1.55 0
ACL 9 0 2 6 1
CLI 14 0 1 11 2 c1101)
6 28 12.4 25.4 IMM 8 4 1 0 3 2.12 0.97 o
ACL 9 0 2 6 1
CLI 19 0 o 16 3 0156)
7 27 8.6 27.6 IMM 29 24 1 o 4 2.5 1.46 0.88 o
ACL 5 0 1 3 1
CLI 26 0 3 1o 13 E(13)
8 210 2.0 26.2 IMM 39 11 12 1 15 3.97 2.79 2.31 0.78 0
ACL 5 o 1 3 1
CLI 14 0 o 10 4
9 27 10.0 25.5 IMM 36 18 1 o 17 3.99 3.5 3.34 2.36 0.63
ACL 11 0 5 4 2
CLI 14 0 0 1o 4 HBO)
10 29 14.1 26.2 IMM 24 11 5 o 8 4 3.78 4 4 1.62 0
ACL 8 0 2 4 2
CLI 14 o o 11 3 G182)
11 27 15.2 27.2 IMM 27 8 4 0 15 o 3.97 4 1.48
ACL 8 0 5 3 0 —
CLI 15 0 o 13 2 H1125)
12 29 16.2 27.4 IMM 33 20 3 o 10 1.86 4 0 2.13
ACL 9 0 2 6 1
CLI 15 0 0 13 2 I(82)
13 27 9.7 26.9 IMM 37 15 2 o 20 1.95 0.97 o 2.23
ACL 5 o 0 4 1
CLI 20 0 o 19 1
14 214 2.6 coc 3.5 2.96 1.99 3.67
15 29 9.8 coc __4_ ___4_3.62 4
16 30 12.1 coc 3.99
17 28 15.1 coc 4
L

 

 

1111‘

 

 

Table 42. L. rubellus field microcosm summary by date.

 

 

 

 

 

N TEMP MOIST BEGINNING ENDING
DATEDAYS °c ° STAGE N IMM ACL COCOON SERIES
1 1831.51 IMM1610 41 o o
ACL 11 3 3 1 4
CLI 4 0 1 1 2
2 27 11.071988 IMM 44 12 5 o 27
ACL 4 0 2 1 1
CLI 2 o 0 0 2 M37)
3 28 11.9 25.45 IMM 53 32 9 0 12 0
ACL 20 o 15 1 4
CLI 1 o 0 0 1 B(71)
4 28 13.25 22.84 IMM 32 17 0 0 15 0.91 o
ACL 24 o 9 13 2
CLI 2 o 0 0 2 C(46)
5 28 13.84 28.91 IMM 12 1 4 0 7 2.63 0.39 o
ACL 9 0 2 6 1
CLI 14 0 1 11 2 D(6)
6 28 12.4 25.45 IMM 8 4 1 0 3 3.031.49 0 o
ACL 9 0 2 6 1
CLI 19 o o 16 3
7 27 8.57 27.55 IMM 29 24 1 o 4 4 1.94 0.13 0
AOL 6 0 1 3 2
CLI 26 o 3 10 13 E(9)
8 210 2 26.22 IMM 2911 2 1 15 0 1.92 0.3
ACL 5 0 1 3 1
CLI 13 0 2 7 4 H7)
9 27 9.97 25.51 IMM 36 18 1 0 17 0.63 2.93 3.33 2.7 o
ACL 11 o 5 4 2
CLI 14 0 0 1o 4
10 29 14.11 26.16 IMM 24 11 5 0 8 1.45 3.24 3.91 3.37
ACL 8 0 2 4 2
CLI 14 o o 11 3
11 27 15.23 27.18 IMM 27 8 4 0 15 3.213.81 4 4 2.73 0.77
ACL 8 o 5 3 0
CLI 15 0 0 13 2 H112
12 29 16.24 27.39 IMM 33 20 3 0 10 3.73 3.86 o 2.94 2.62
ACL 9 0 2 6 1
CLI 15 o 0 13 2
13 27 9.69 26.91 IMM 37 15 2 0 20 3.89_4_1.3 4
ACL 5 o 0 4 1
CLI 20 0 0 19 1
14 214 2.57 coc 4 1.73
15 29 9.75 000 3.41
16 30 12.1 000 3.83
17 28 15.11 coc 4

 

1F

 

 

181

 

Table 43. A. tuberculata field microcosm summary by date

 

 
 

   

 

   

 

 

 

 

 

N TEMP MOIST BEGINNING ENDING STAGES
DATEDAYS ° ° MACL CLI DIED COCOONSERIES
1 183 1.5 91 5 0 39
5 20 0 6
CLI 0 0 0 0 0
2 27 11.1 19.92 IMM 96 76 6 0 14
ACL 25 2 23 0 0
CLI 0 0 o 0 0
3 28 11.9 24.50 IMM 116 76 4 0 36
ACL 46 2 34 8 2
CLI 3 0 0 3 0 M35)
4 28 13.3 22.47 IMM 79 44 4 0 31 0
ACL 38 0 14 20 4
CLI 11 0 5 6 o B(19)
5 28 13.8 29.43 IMM 45 26 0 o 19 1.72 0
ACL 22 o 1 15 6
CLI 26 0 1 25 0 0129)
6 28 12.4 24.65 IMM 47 35 0 0 12 3.311.47 0
AOL 2 0 1 1 0
CLI 40 0 5 23 12 D(12)
7 27 8.6 28.86 IMM 59 35 3 0 21 3.9 2.06 0.6 0
ACL 29 1 11 12 5
CLI 24 0 3 13 8 E16)
8 210 2.0 25.52 IMM 36 24 5 0 7 4 3.441.77 0.79 0
ACL 17 1 11 5 o
CLI 15 0 2 13 o F(22)
9 27 10.0 24.62 IMM 42 27 3 0 12 0 3.95 3.15 2.63 0.22
ACL 20 2 9 5 4
CLI 0 0 o 0 o G134)
1o 29 14.1 26.09 IMM 51 25 5 0 21 1.15 4 4 3.12 0
ACL 18 0 7 7 4
CLI 22 0 0 17 5 H(17)
11 27 15.2 27.24 IMM 57 25 3 0 29 3.73 0 4 2.46
ACL 15 0 6 4 5
CLI 28 0 1 23 4 I114)
12 29 16.2 27.30 IMM 47 17 3 0 27 4 1.05 0 4
ACL 10 o 3 5 2
CLI 27 0 4 23 0
13 27 9.7 26.90 IMM 63 34 6 0 23 3.08 0.21
ACL 10 o 6 4 0
CLI 28 o 3 23 2
14 214 2.6 000 27 27 0 0 0 3.48 3.17
15 29 9.8 coc 6 6 0 0 o 4 3.4
16 30 12.1 0001212 0 0 0 4
C.

Ba

Ba

Be

Bi

Br

B1

B1

BI

_ — _ ._*2 -4..-—

 

LITERATURE CITED

Anderson, J.M. 1988. Spatiotemporal effects of invertebrates on soil processes.
Biol. Fertil. Soils 6:216-227.

Barley, KB and AC. Jennings. 1959. Earthworms and soil fertility. III. The
influence of earthworms on the availability of nitrogen. Aust. J. Agric. Res.
10:364-370.

Bather, EA. 1920. Protoscolex laz‘us, a new worm from Lower Ludlow beds.
Ann. Mag. Nat. Hist. 9:5.

Beadle, LC. 1957. Respiration in the African swampworm Alma emini Mich. J.
Exp. Biol. 34:1-10.

Bierzychudek, KR 1982. Demography of Jack-in-the~pulpit, a forest perennial
that changes sex. Ecol. Monogr. 52:335-351.

Bouché, MB. 1972. Lombriciens de France. Ecologie et Syste’matique. INRA
Pub]. 7 2-2. Institut National des Recherches Agiculturelles, Paris.

Bouché, MB. 1977. Strategies lombriciennes. [71: Soil Organisms as
Components of Ecosystems (U. ohm and T. Persson, eds.). Ecol. Bull.
(Stockholm) 25: 122-132.

Bouché, MB. 1980. Objectifs, compartimentation et faisabilite' du modele
R.E.A.L. (Rfile Ecologique et Agronomique des Lombriciens. Pedobiologia
20: 197-211.

Bouche’, M. B. and A. Kretschmar. 197 7 . REAL: Un modele du rfile écologique
et agronomique des lombriciens. In: Soil Organisms as Components of
Ecosystems (U. Lohm and T. Persson, eds.). Ecol. Bull. (Stockholm)
25:402-408.

Bouyoucos, G. J. 1927. The hydrometer as a new and rapid method for
determining the colloidal content of soils. Soil Sci. 23:319-331.

182

 

 

 

 

Brt

Bul

Ca:

Ca

Cl:

Cl:

Cr

C1

C1

 

 

183

Brusca, RC. and G]. Brusca. 1990. Phylum Annelida: the segmented worms. In:
Invertebrates. Pp. 381-436. Sinauer, Sunderland.

Butt, K.R., J. Frederickson and R.M. Morris. 1992. The intensive production

of Lumbricus terrestris L. for soil amelioration. Soil Biol. Biochem.
24:1321-1325.

Caswell, H. 1989. Matrix Population Models: Construction, Analysis, and
Interpretation. Sinauer Associates, Sunderland MA. 328 pp.

Catriona, M., K. Gardner, J.P. Bell, J.D. Cooper, T.J. Dean, and M. Hodnett.
1991. Soil Water Content. in: Soil Analysis: Physical Methods, K.A.
Smith and CE. Mullins, eds. Marcel Dekker, New York. pp. 1-74.

Clancy, K.M. and R.M. King. 1993. Defining the western spruce budworm's
nutritional niche with response surface methodology. Ecology 74(2):442-
454.

Clark, R.B. 1978. Composition and Relationships. In: Physiology of Annelids
(P.J. Mill, ed.), pp. 1-32. Academic Press, London.

Crouse, D.T., L.B. Crowder and H. Caswell. 1987. A stage-based population
model for loggerhead sea turtles and implications for conservation. Ecology
68: 1412-1423.

Crowley, PH. 1992. Resampling methods for computation—intensive data
analyses in ecology and evolution. Ann. Rev. Ecol. Syst. 23:405-447.

Curry, J. P. and D.C.F. Cotton. 1983. Earthworms and land reclamation. In:
Earthworm Ecology (J.E. Satchell, ed.), pp. 215-228. Chapman & Hall,
London.

Dales, RP. 1978. Defence mechanisms. In: Physiology of Annelids (P.J.
Mill, ed.), pp. 479-507. Academic Press, London.

Daniel, 0. 1991. Leaf-litter consumption and assimilation by juveniles of
Lumbricus terrestris L. (Oligochaeta, Lumbricidae) under different
environmental conditions. Biol. Fertil. Soils 12:202-208.

Darwin, C .R. 1881. The formation of vegetable mould through the actions of
worms, with observations on their habits. Appleton, New York. 326 pp.

Deevey, E.S. Jr. 1947. Life tables for natural populations of animals. Q. Rev.
Biol. 22:283-314.

 

 

 

D0

D0

Ea

Ed

EC

Er

F.

184

Dobrolyubov, A.I. 1986. The mechanism of locomotion of some terrestrial
animals by travelling waves of deformation. J. Theor. Biol. 119:457—
466.

Dotson, DB. and P.J. Kalisz. 1989. Characteristics and ecological
relationships of earthworm assemblages in undisturbed forest soils in the
southern Appalachians of Kentucky, USA. Pedobiologia 33:211—220.

Easton, E.G. 1983. A guide to valid names of Lumbricidae (Oligochaeta). In:
Earthworm Ecology. From Darwin to Vermiculture (J.E. Satchell, ed.),
pp 475—485. Chapman and Hall, London.

Edwards, C.A., and K.E. Fletcher 1988. Interactions between earthworms and
microorganisms in organic—matter breakdown. Agric. Ecosyst. Environ.
24:235-247.

Edwards, C.A. and J.R. Lofty. 1972. Biology of Earthworms. Chapman and
Hall, London. 283 pp.

Edwards, W.M., S.J. Shipitalo, S.J. Traina, C.A. Edwards and LB. Owens.
1992. Role of Lumbricus terrestris (L.) burrows on water quality of
infiltrating water. Soil Biol. Biochem. 24:1555—1561.

Ehlers, W. 1975. Observations on earthworm channels and infiltration on tilled
and untilled loess soils. Soil Science 119:242—249.

Evans, A.C. and W.J.McL. Guild. 1947. Studies on the relationship between
earthworms and soil fertility. 1. Biological studies in the field. Ann.
App. Biol. 34:307—330

Evans, A.C. and W.J.McL. Guild. 1948. Studies on the relationship between
earthworms and soil fertility. 4. On the life cycles of some British
Lumbricidae. Ann. App. Biol. 35:471—484.

Fischer, E. and L. Molnar. 1992. Environmental aspects of the chloragogenous
tissue of earthworms. Soil Biol. Biochem. 24:1723—1727.

Food and Agriculture Organization (FAO). 1977. "Assessing Soil
Degradation". FAO Bull. 34, Rome.

Fragoso, C. and P. Lavelle. 1992. Earthworm communities oftropical
rainforests. Soil Biol. Biochem. 24:1397-1408.

 

Fn

Ge

G:

G:

185

Frenot, Y. 1992. Introduced populations of Dendrodrilus rubidus ssp.
(Oligochaeta: Lumbricidae) at Crozet, Kergeulen and Amsterdam Islands:
Effects of temperature on growth patterns during the juvenile stages. Soil

Biol. Biochem. 24:1433-1439.
Gates, GE. 1970. Miscellanea megadrilogica VII. Megadrilogica 1(2):1-14.

Gates, GE. 1975. Contributions to a revision of the earthworm family
Lumbricidae. XII. Enterion mammale Savigny, 1826 and its position in
the family. Megadrilogica 2(1):1—5.

Gates, GE. 1979. South Dakota does have earthworms! Megadrilogica 3:165—

166.
Gates, GE. 1982. Farewell to North American megadriles. Megadrilogica
4:12—77.

Germann, P.F., W.M. Edwards and L.B. Owens. 1984. Profiles of bromide
and increased soil moisture after infiltration into soils with macropores.
Soil Science Soc. Am. I. 48: 237—244.

Gill, J.L. 1978. Design and Analysis of Experiments in the Animal and
Medical Sciences: Volume 2. Iowa State University Press, Ames, Iowa

USA. 301 pp.

Gjelstrup, P. and NB. Hendriksen. 1991. Histz'osoma murchiei Hughes and
Jackson (Anoetidae) as a parasite in the cocoons of some Danish
earthworms. In: The Acari. Reproduction, Development and Life—history

Strategies (R. Schuster and RW. Murphy, eds), pp 441-445. Chapman and
Hall, London.

Glaessner, M.F., W. V. Priess and H.R. Walter. 1969. Precambrian columnar
stromatolites in Australia: Morphological and stratigraphic analysis.

Science 164:1056-158.

Graff, O. 1953. Die Regenwiirmer Deutschlands. Schrift. Forsch. Land.

Brauschweig-Volkenrode 7.

Grant, W.C. 1955. Studies on moisture relationships in earthworms. Ecology
36:400—407.

Granval, P. and R. Aliaga. 1988. Analyse critique des conaissances sur les
predateurs de lombriciens. Gibier Faune Sauvage 5:71—94.

 

 

Hai

Hai

Ha

Ha

H(

186

Haimi, I., and M. Boucelham 1991. Influence of a litter feeding earthworm,
Lumbricus rubellus, on soil processes in a simulated coniferous forest
floor. Pedobiologia 35:247-256.

Haimi, J., and M. Einbork 1992. Effects of endogeic earthworms on soil
processes and plant growth in coniferous forest soil. Biol. Fertil. Soils
13:6—10.

Hall, PG. 1922. The vital limit of desiccation of certain animals. Biol. Bull.
(Woods Hole) 42:31-51.

Hartenstein, R. 1986. Earthworm biotechnology and global biogeochemistry.
Adv. Ecol. Res. 15:379—409.

Holmstrup, M., B.T. Hansen, A. Nielsen, and I.K. Ostergaard 1990. Frost
tolerance of lumbricid earthworms. Pedobiologia 34:361—366.

Holmstrup, M., I.K. Ostergaard, A. Nielsen, and B.T. Hansen 1991. The
relationship between temperature and cocoon incubation time for some
lumbricid species. Pedobiologia 35:179-184.

Hoogerkamp, M., H. Rogaar and H.J.P. Eijsackers. 1983. Effects of

earthworms on grassland on recently reclaimed polder soils in the
Netherlands. In: Earthworm Ecology (J.E. Satchell, ed.), pp.88—105.

Chapman & Hall, London.

Joschko, M., W. Séchtig and O. Larink. 1992. Functional relationship between
earthworm burrows and soil water movement in column experiments.

Soil Biol. Biochem. 24:1545—1547.

Judas, M. 1989. Predator—pressure on earthworms: field experiments in a
beechwood. Pedobiologia 33:339—354.

Kalisz, P.J., and DB. Dotson 1989. Land-use history and the occurrence of
exotic earthworms in the mountains of eastern Kentucky. Am. Midl. Nat
122:288—297.

Knollenberg, W.G., R.W. Merritt, and D.L. Lawson 1985. Consumption of
leaf litter by Lumbricus terrestris (Oligochaeta) on a Michigan woodland
floodplain. Am. Midl. Nat. 11311—6.

1991. Weight response to the soil water

Kretschmar, A., and C. Bruchou.
potential of the earthworm Aporrectodea longa. Biol. Fertil. Soils

12:209—212.

 

187

Lavelle, P. and J.A. Meyer. 1977. Modélisation et simulation de la dynamique,
de la production et de la consommation des populations du Ver de terre

géophage Millsonia anomala (OIigochetes-Acanthodrilidae) dans la
Savane de Lamto (Cote de Ivoire). In: Soil Organisms as Components of

Ecosystems (U. Lohm and T. Persson, eds.). Ecol. Bull. 25:420—430.

Lavelle, P. and J.A. Meyer. 1982. Allez-les—vers, a simulation model of
dynamics and effect on soil of populations of Millsom'a anomala
(Oligochaeta — Megascolecidae). In: New Trends in Soil Biology. Proc.
VII Intl. Colloquium on Soil Zoology, Lebrun Ph., André H.M., de
Medts A., Grégoire-Wibo C., Wauthy G., eds., pp. 503-517,
Dieu—Brichart, Ottignes-Louvain-la Neuve, FR.

Lavelle, P., R. Schaefer, and Z. Zaidi 1989. Soil ingestion and growth in

Millsom'a anomala, a tropical earthworm, as influenced by the quality of
the organic matter ingested. Pedobiologia 33:379-388.

Laverack, M.S. 1961. Tactile and chemical perception in earthworms -— 11.
Responses to acid pH solutions. Comp. Biochem. Physiol. 2:22-34.

Laverack, M.S. 1963. The Physiology of Earthworms. Macmillan, New York.

206 pp.
Lee, K.E. 1972. Biology of Earthworms. Chapman and Hall, London. 283

PP-

Lee, K.E. 1983. The influence of earthworms and termites on soil nitrogen
cycling. In: New Trends in Soil Biology (Ph. Lebrun, H.M. André, A.
de Medts, C. Grégoire—Wibo and G. Wauthy, eds.), pp. 35—48. Proc. 8th
Int'l Colloq. Soil 2001., Louvain—la—Neuve, 1982. Dieu—
Briehart,Ottignies—Louvain-la—Neuve.

Lee, K.E. 1985. Earthworms. Their Ecology and Relationships with Soils and
Land Use. Harcourt Brace Jovanovich, New York. 511 pp.

Leslie, RH. 1945. On the use of matrices in certain population mathematics.
Biometrika 33:183—212.
1948. Some further notes on the use of matrices in population

Leslie, P.H.
Biometrika 35:213—245.

mathematics.
Liscinsky, S. 1965. The American woodcock in Pennsylvania. Pennsylvania
Game Commission, Harrisburg, PA. 32 pp.

 

 

L01

Ma

M2

188

Loreau, M. 1988. Determinants of the seasonal pattern in the niche structure of
a forest carabid community. Pedobiologia 31:75-87.

MacArthur, R.H. and EC. Wilson. 1967. The theory of island biogeography.
Princeton University Press, Princeton, NJ. 203 pp.

Macdonald, D.W. 1980. The red fox, Vulpes vulpes, as a predator upon
earthworms, Lumbricus terrestris. Z. Tierpsychol. 52:171—200.

Mangum, GP. 1978. Temperature adaptation. In: Physiology of Annelids
(P.J. Mill, ed.), pp. 447-478. Academic Press, London.

Martin, NA. 1982. The interaction between organic matter in soil and the
burrowing activity of three species of earthworms (Oligochaeta:

Lumbricidae). Pedobiologia 24: 185—190.

Martin, S. and P. Lavelle. 1992. A simulation model of vertical movements of
an earthworm population (Millsom'a anomala Omodeo, Megascolecidae)
in an African savanna (Lamto, Ivory Coast). Soil Biol. Biochem.
24:1419—1424.

McKey—Fender, D. and W.M. Fender. 1982. Arctiostrotus (gen. nov.). 1. The

identity of Plutellus perrl’eri Benham 1892 and its relation to glacial
refugia, Megadrilogica 4:82-85.

Meglitsch, RA. and ER. Schram. 1991. Annelida. In: Invertebrate zoology.
Pp. 302—346. Oxford University Press. Oxford.

Michon, J. 1954. Influence de l’isolement a partir de la maturité sexuelle sur la
biologic des Lumbricidae. C. r. hebd. Séanc. Acad. Sci., Paris
238:2457—2458.

Mitchell, M.J. 1983. A simulation model of earthworm growth and population

dynamics: Application to organic waste conversion. In: Earthworm
Ecology. From Darwin to Vermiculture (J.E. Satchell, ed.), pp. 339

349. Chapman and Hall, London.
Myers, R.H. 1971. Response Surface Methodology. Allyn and Bacon, Boston.

Nielsen, C. 0. 1962. Carbohydrases in soi] and litter invertebrates. Oikos 13:200-
2 I 5.

 

Nit

Nt

189

Nielsen, GA, and ED. Hole 1964. Earthworms and the development of
coprogenous Al horizons in forest soils of Wisconsin. Soil Science Society

Proceedings 28:426-430.

Nordstrom, S. 1975. Seasonal activity of lumbricids in southern Sweden. Oikos
26:307-315.

Nordstr'o'm, S. and S. Rundgren. 1974. Environmental factors and lumbricid
associations in southern Sweden. Pedobiologia 14:1-27.

Oglesby, LC. 1978. Salt and water balance. In: Physiology of Annelids (P.J.
Mill, ed.), pp. 555-65 8. Academic Press, London.

Olive, P.J.W. and RB. Clark. 1978. Physiology of Reproduction. In: Physiology
ofAnnelids (P.J. Mill, ed.), pp. 271-368. Academic Press, London.

Oliver, J.H. 1962. A mite parasitic in the cocoons of earthworms. J. Parasitol.
48:120-123.

Omodeo, P. 1956. Contributo alla revisione dei Lumbricidae. Arch. 2001. Ital.
41: 131-212.

Omodeo, P. 1963. Distribution of the terricolous oligochaetes on the two shores
of the Atlantic. In: Love and Love (eds.). North Atlantic biota and their
history. New York, Pergamon Press. Pp. 127-141.

Pakkala, T. And T. Kolstrom. I988. Simulating the development of Norway
spruce stands using a transition matrix. For Ecol. Manage. 25:255—267.

Parmelee, R.W. and DA. Crossley Jr. 1988. Earthworm production and role in
the nitrogen cycle of a no-tillage agroecosystem on the Georgia Piedmont.

Pedobiologia 32:353-361.

Pashanasi, B. and P. Lavelle. 1992. Effect of inoculation with the endogeic
earthworm Pontoscolex corethurus (Glossoscolecidae) on N availability,
soil microbial biomass and the growth ofthree tropical fruit tree seedlings in
a pot experiment. Soil Biol. Biochem. 24:1655-1659.

Perel’, TS. 1977. Differences in lumbricid organization connected with
ecological properties. In: Soil Organisms as Components of Ecosystems
(U. Lohm and T. Persson, eds.) Ecol. Bull. (Stockholm) 25:56-63.

Petersen, H. and M. Luxton. 1982. A comparative analysis of soil fauna
populations and their role hi decomposition processes. Oikos 39:288-388.

190

Phillipson, J ., R. Abel, J. Steel, and S.R.J. Woodell 197 8. Earthworm numbers,
biomass and respiratory metabolism in a beech woodland -- Wytham Woods,
Oxford. Oecologia 33:291-309.

Piearce, T. G. 1981. Losses of surface fluids from lumbricid earthworms.
Pedobiologia 21:417-426.

Press, W.H., B.P. Flannery, S.A.Teukolsky and W. T. Vetterling. 1986. Numerical
recipes: the Art of Computing. Cambridge University Press, Cambridge
England.

Raw, F. 1962. Studies on earthworm populations in orchards. Ann. Appl.
Biol. 50:389-404.

Reichle, DE. 1971. Systems analysis as applied to ecological problems: a
method for synthesis, integration and interpretation of IBP woodlands
ecosystem research. In: Systems Analysis in Northern Coniferous
Forests -— IBP Workshop (T. Rosswall, ed.). Bull. Ecol. Res. Comm.
NPR 14: 12-28. Swedish National Science Research Council, Stockholm.

Reinecke, A.J. 1975. The influence of acclimation and soil moisture on the
temperature preference of Eisenia rosea (Lumbricidae). In: Progress in
Soil Zoology, (J. Vanek, ed.), pp. 341—349. Proc. 5th Int'l Colloq. Soil
2001., Prague, 1973. Junk, The Hague.

Reinecke, A.J., S.A. Viljoen and RI. Saayman. 1992. The suitability of
Eudrilus eugeniae, Perionyx excavatus and Eisenz'a fetida (Oligochaeta)

for vermicomposting in southern Africa in terms of their temperature
requirements. Soil Biol. Biochem. 24:1295-1307.

Reynolds, J.W. 1973a. The earthworms of Rhode Island (Oligochaeta:
Lumbricidae). Megadrilogica 1(6):1-4.

Reynolds, J.W. 1973b. The earthworms of Delaware (Oligochaeta:
Acanthodrilidae and Lumbricidae). Megadrilogica 1(5):1-4.

Reynolds, J.W. 1974. Earthworms of Maryland (Oligochaeta: Acanthodrilidae,
Lumbricidae, Megascolecidae, SparganOphilidae). Megadrilogica
1(11):1-12.

Reynolds, J.W. 1975a. [The earthworms (Oligochaeta: Lumbricidae) of Cape
Breton] Megadrilogica 2(6): 1—7.

 

191

Reynolds, J.W. 1975b. The earthworms (Oligochaeta: Lumbricidae) of Prince
Edward Island. Megadrilogica 2(7):4-11.

Reynolds, J.W. 1976. The distribution and ecology of the earthworms of Nova
Scotia. Megadrilogica 2(8): 1-7.

Reynolds, J.W. 1977a. The Earthworms (Lumbricidae and Sparganophilidae)
of Ontario. Life Sciences Miscellaneous Publications, Royal Ontario
Museum, Toronto. 141 pp.

Reynolds, J .W. 1977b. The earthworms of Massachusetts (Oligochaeta:

Lumbricidae, Megascolecidae, and Sparganophilidae). Megadrilogica
3:49-54.

Reynolds, J.W. 19770. The earthworms of Tennessee (Oligochaeta). II.
Sparganophilidae, with the description of a new species. Megadrilogica
3:61-64.

Reynolds, J .W. 1977d. The earthworms of Tennessee (Oligochaeta). III.
Komarekionidae, with notes on distribution and biology. Megadrilogica
3:65-69.

Reynolds, J.W. 1978a. A contribution to our knowledge of the earthworm
fauna of North Dakota. Megadrilogica 32148-149.

Reynolds, J.W. 1978b. The earthworms of Tennessee (Oligochaeta). IV.
Megascolecidae, with notes on distribution, biology, and a key to the
species in the state. Megadrilogica 3:81—116.

Reynolds, J.W. 1995. The distribution of earthworms (Annelida, Oligochaeta)
in North America. Pp. 133—153 in: Mishra, P.C., N. Behera, B.K.
Senapati and BC. Guru (eds.). Advances in Ecology and Environmental
Science. Ashihs Publishing House, New Delhi. 651 pp.

Reynolds, J.W., E.E.C. Clebsch and W.M. Reynolds. 1974. Contributions to
North American Earthworms (Oligochaeta). No. 13. The earthworms of
Tennessee (Oligochaeta). Lumbricidae. Bull. Tall Timbers Res. Sta.

17:1-133.

Reynolds, J.W. and GA. Jordan. 1975 . A preliminary conceptual model of
megadrile activity and abundance in the Haliburton highlands.
Megadrilogica 2(2): 1-8.

 

I92

Rhee, J .A. van. 1977. A study of the effects of earthworms on orchard
productivity. Pedobiologia 17:107—114.

Richards, K.S., and C. Arme 1982. Integumentary uptake of dissolved organic
materials by earthworms. Pedobiologia 23:358—366.

Rigby, B.J. 1968. Temperature relationships of poikilotherms and the melting
temperature of native collagen. Biol. Bull. (Woods Hole) 135:223—229.

Rohlf, F.J. and RR. Sokal. 1995. Statistical Tables. 3rd ed. Freeman, New York.

Rosenberg, AA. and R.W. Doyle. 1986. Analyzing the effect of age structure and
stock recruitment in herring (Clupea harengus). Can. J. Fish. Aquatic Sci.
43:674-679.

Ruppert, BE. and RD. Barnes. 1994. The Annelids. In: Invertebrate Zoology
(3rd ed.), pp. 505-593. W.B. Saunders, Philadelphia.

Rushton, SP. and ML. Lufl‘. 1984. A new electrical method for sampling
earthworm populations. Pedobiologia 26:15-19.

Satchell, J.E. 1967. Lumbricidae. In: "Soil Biology" (A. Burges and F. Raw,
eds.), pp.259-322. Academic Press, London.
Satchell, J.E. 1980. r worms and K worms: a basis for classifying lumbricid

earthworm strategies. In: Soil Biology as Related to Land Use Practices
(D.L. Dindal, ed.), pp. 848—854. Proc. 7th Intl. Colloq. Soil 2001.,

Syracuse, 1979. EPA, Washington DC.

Satchell, J.E. and D.G. Lowe. 1967. Selection of leaf litter by Lumbricus
terrestris. In: Progress in Soil Biology (0. Graff and J.E. Satchell,

eds.), pp. 102—119. Vieweg, Braunschweig.

Savitzky, A. and M.J.E. Golay. 1964. Smoothing and differentiation of data by
simplified least squares procedures. Analyt. Chem. 36: 1627- 1639.

Schmidt, P. 1918. Anabiosis of the earthworm. J. Exp. Zool. 27:55—72.
Semenova, L.M. 1967. Biological significance of the chloragogenous tissue of
earthworms. J. Evol. Biochem. Physiol. 3:115—123.

Seymour, M. 1978. The infinite variety of worms. New Scientist 77:650-652

 

I93

Sims, R.W. 1983. The scientific names of earthworms. In: Earthworm
Ecology. From Darwin to Vermiculture (J .E. Satchell, ed.), pp 467-474.

Chapman and Hall, London.

Skoczen, S. 1970. Food storage of some insectivorous mammals. Przegl.
2001. 14:243-248.

Smettem, K.R.J. 1992. The relation or earthworms to soil hydraulic properties.
Soil Biol. Biochem. 24:1539-1543.

Smith, P]. 1963. Topics in geophysics. MIT Press, Cambridge. 246 pp.

Snider, R.M. 1991. Checklist and distribution of Michigan earthworms. Mich.
Acad. 24:105-114.

Snider, R.M. 1994. A simple technique for field incubation of earthworms
used to assess potential effects of low-level electric fields. Pedobiologia

38:115-124.

Snider, R.J. and R.M. Snider. 1987. ELF ecological monitoring in Michigan.
1. Description of sites for soil biological studies. Pedobiologia 30:241-

250.

Snider, R.M., and RI. Snider 1988. ELF ecological monitoring in Michigan.
I]. The earthworm communities of test and control sites. Pedobiologia

32:335-342.

Séchtig, W. and O. Larink. 1992. Effect of soil compaction on activity and
biomass of endogeic lumbricids in arable soils. Soil Biol. Biochem.
24:1595—1599.

Sokal, R.R. and F.J. Rohlf. 1995. Biometry. Third edition. W.H. Freeman,

New York. 887 pp.

Springett, J.A., R.A.J. Gray and J.B. Reid. 1992. Effect of introducing
earthworms into horticultural land previously denuded of earthworms.
Soil Biol. Biochem. 24:1615-1622.

Stephenson, J. 1930. The Oligochaeta. Clarendon Press, Oxford. 978 pp.

Stewart, V.l. and J. Scullion. 1988. Earthworms, soil structure, and the
rehabilitation of former opencast coal-mining land. In: Earthworms in
Waste and Environmental Management (C.A. Edwards and F. Neuhauser,

eds.), pp. 263-272. SPB, The Hague.

 

I94

Stine, R. 1990. An introduction to bootstrap methods. Examples and Ideas. In:
J. Fox and J. Scott Long, eds. Modern Methods ofData Analysis. Sage
Publications, London. Pp. 325-373.

Stockdill, S.M.J. 1982. Effects of introduced earthworms on the productivity
of New Zealand pastures. Pedobiologia 24:29-35.

Syers, I.K., A.N. Sharpley and DR. Keeney. 1979. Cycling of nitrogen by
surface—casting earthworms in a pasture ecosystem.. Soil Biol. Biochem.
11:181—185.

Thielemann, U. 1986. Glasrohrchenmethode zur Lebendbestimmung von
Regenwiirmern. [A glass—tube method for live determination of species of
earthworms.] Pedobiologia 29:341—343.

Topp, G.C., J.L. Lewis and AP Annan. 1980. Electromagnetic determination of
soil water content: measurements in coaxial transmission lines. Water
Resour. Res. 16:574-582.

Tsukamoto, J. and H. Watanabe. 1977. Influence of temperature and hatching
on growth of Eiseniafoetida (Oligochaeta, Lumbricidae). Pedobiologia
17:338-342.

Viljoen, S.A. and A.J. Reinecke. 1992. The temperature requirements of the
epigeic earthworm species Eudrilus eugeniae (Oligochaeta) —— A
laboratory study. Soil Biol. Biochem. 24:1345—1350.

Viljoen, S.A., A.J. Reinecke and L. Hartman. 1992. The influence of
temperature on the life-cycle of Dendrobaena veneta (Oligochaeta). Soil
Biol. Biochem.24:1341~1344.

Vimmerstedt, J. P. and J.H. Finney. 1983. Impact of earthworm introduction
on litter burial and nutrient distribution in Ohio stripmine banks. Proc.
Soil Science Soc. Am. 37:388—391.

Walther, PB. and R.M. Snider. 1984. Techniques for sampling earthworms
from leaf litter, humus and soil. Pedobiologia 27:293—297.

Weber, RE. 1978. Respiratory pigments. In: "Physiology of Annelids" (P.J.
Mill, ed.), pp. 393—446. Academic Press, London.

Wolters, V. and W. Stickan. 1991. Resource allocation of beech seedlings
(Fagus sylvatica L.) —— relationship to earthworm activity and soil
conditions. Oecologia 88:125-131.

195

Woolhouse, M.E.J. and R. Harmsen. 1989. A transition matrix model of the
population dynamics of a two~prey two-predator acarid complex. Ecol.
Modelling 39:307—323.

Wright, HE. and D.G. Frey. 1965. The Quaternary of the United States.
Princeton University Press, New Haven. 922 pp.

Yahnke, W. and J.A. George. 1972. Rearing and immature stages of the cluster
fly, Pollenia rudis (Diptera: Calliphoridae) in Ontario. Can. Ent.
104:567—576.

Zajonc, I. 1972. La distribution quantitative des lombrices (Lumbricidae,
Oligochaeta) dans les grand types mondieux d’ecosystem forestries. In:
Productivity ofForest Ecosystems (P. Duvigneaud, ed.), pp. 453-462.
UNESCO, Paris.

r"

 

= “11111111llllllllf ,