| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
From * Hamilton Thorne Biosciences, Beverly,
Massachusetts; Kuster Research and Consulting,
Geneseo, Illinois; IVF Center, Department of
Reproductive Medicine, VUMc Academic Hospital, Vrije Universiteit, Amsterdam,
The Netherlands; and University of Pennsylvania,
School of Veterinary Medicine, Department of Clinical Studies, New Bolton
Center, Kennett Square, Pennsylvania.
| Correspondence to: Gary C. Althouse, University of Pennsylvania, School of Veterinary Medicine, Department of Clinical Studies, New Bolton Center, Kennett Square, PA 19348-1692 (e-mail: gca{at}vet.upenn.edu). |
| Received for publication May 25, 2004; accepted for publication September 8, 2004. |
 |
Abstract |
|---|
 Top Abstract Materials and MethodsResultsDiscussionConclusionReferences |
|---|
Key words: Spermatozoa, capillary slide, concentration, hemacytometer, Poiseuille, Segre-Silberberg
During capillary loading, particles suspended in the flow are exposed to a velocity gradient, which causes particles to move transversely to the flow direction (Vasseur and Cox, 1976). This phenomenon, first reported by Segre and Silberberg (1961) and called the Segre-Silberberg (SS) effect, was observed for particles initially evenly suspended in a cylindrical and planar Poiseuille flow, which then became unevenly concentrated at a well-defined distance from the walls. For a sample of semen (neat or diluted) entering a capillary chamber, sperm can therefore be expected to cross the flow and congregate in faster flowing layers of fluid. The transfer of sperm into the faster flowing layers causes a concomitant reduction in unit area sperm concentration and, thus, an accumulation of sperm at the meniscus. Consequently the measured concentration at the meniscus will be higher, and behind the meniscus (during or after capillary flow) will be lower, than in the original sample. Indeed, previous works using 20-µm capillary slides (Johnson et al, 1996b; Mahmoud et al, 1997) have noted differences in latex-bead concentration across the slide.
We have predicted previously the expected difference between actual sample concentration and that measured in low-volume capillary-loaded slides, similar to those used in computer-automated semen-analysis systems (Douglas-Hamilton et al, 2005). We found that the velocity of transverse flow and hence significance of the SS effect depends on the sample viscosity, surface tension, chamber depth, and cell size. For low-viscosity samples in which the SS effect is completely developed, our model predicted that a factor of 1.30 is necessary to compensate for the SS effect. The purpose of this study was to test and validate our prior predictions on the measured sperm concentration in capillary-loaded slides. Three experiments were designed using low-viscosity samples. Experiment I was performed to verify the predicted meniscus concentration peak in diluted boar and human sperm concentrations within a chamber. In experiment II, inflow velocities of entrained beads and sperm into a capillary chamber were assessed. Last, experiment III was a field study of 2 boar studs, in which boar-sperm concentrations measured using hemacytometry and capillary slides were compared. We found the predicted compensation factor allowed for consistent and accurate agreement between the 2 methodologies.
|
Materials and Methods |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
= 630 nm incident at the Brewster
angle, which is reflected from the top and bottom walls of the chamber and the
beams recombined. Loaded and unloaded chamber depths are determined with
accuracy to within 0.5 µm of actual depth. Along with verifying chamber
depth, interferometry was also used to ensure that the hemacytometer depth did
not significantly change when repeatedly loaded with a solution of viscosity
1.0 centipoise to a level that visually filled the chamber.
|
Both porcine and human semen samples were used in the initial experiments. Porcine semen was obtained from Monsanto Choice Genetics (St Louis, Mo), where the semen was diluted to 40 x 106 sperm per mL (abbreviated as 40 M/mL) and shipped overnight at approximately 17°C to the laboratory for subsequent analyses. Donor human semen samples were washed and resuspended to 22 M/mL using a modified HTF medium with 0.5% human serum albumin (Biowhittaker Europe, Verviers, Belgium). For the field studies, both diluted-chilled (trial A) and fresh-extended (trial B) porcine semen was used, specifics of which are further outlined below.
Experiment I. Intrachamber Sperm Concentration
Porcine extended semen samples (N = 4 boars) were subdivided by placing
1.0-mL aliquots of sample into each of 5 covered test tubes, and by placing
0.5 mL into a sixth test tube. Extender media was added to each tube to
achieve final predicted sperm concentrations from the measured weight dilution
relative to samples of 1.0, 0.837, 0.672, 0.498, 0.332, and 0.165,
respectively. Weights were verified using a calibrated digital balance
(Mettler; Fisher Sci). Human semen samples (N = 3), suspended to 22 M/mL in a
modified HTF medium, were immobilized by heat inactivation following loading
into the Leja slide and prior to data acquisition. All subsequent procedures
remained the same.
Capillary chambers were loaded using an adjustable digital pipette (P-10,
Rainin Instrument Co Inc, Woburn, Mass) to place 1.5-1.7 µL of sample at
the capillary slide entry port. Our observation has been that the sample moves
smoothly into the chamber until the rear meniscus of the sample globule
reaches the entry port, at which point further flow ceases. The forward
meniscus is located 10.3-14.8 mm from the entry port. Sample load time for
diluted boar semen with a viscosity of 1.2 cP (determined by viscometry
[Gilmont GV-2100; VWR, Boston, Mass]) is 1.3-2.5 seconds, with this time
variation due to small differences in the glass-sample capillary attraction.
The cessation of flow is abrupt, and the distribution of nonmotile sperm
inside the flow will not change significantly thereafter. All sperm introduced
into the chamber are visible. Nonmotile sperm can only be moved by drift or
Brownian motion: because both of these are negligible, the sperm remain in
their positions, and a frozen picture is obtained of particle distribution at
the moment when flow ceases. For motile sperm, the inflow velocity (
8 000
µm/s) is greater than the sperm swimming velocity (200 µm/s), thus,
sperm distribution is controlled by the chamber flow dynamics. Rapid analysis
after cessation of flow ensures that motile sperm do not have time to change
their positions significantly.
Photomicroscopic capture of sperm concentration in the 20-µm slide following the loading operation was measured using a CEROS analyzer (Version 12.1; Hamilton-Thorne Biosciences) with a 10x negative phase contrast objective on an Olympus CH-2 microscope (Optical Analysis, Nashua, NH) and a Sony XC-75 camera (Sony Corp, Tokyo) interfaced with a computer. Each field was immediately checked visually for correct sperm acquisition and recognition by CEROS using the playback feature. Standard setup values for diluted boar sperm were used (Table 1). Because the main variation in concentration is in the direction of flow (and lateral concentration variation was shown to be negligible in preliminary experiments), we measured concentration down the centerline of the chamber at 1-mm intervals, starting at 1 mm from the entry and continuing up to the forward meniscus. In each of the samples, sequential measurements were replicated 4 times in each of 4 separately loaded chambers.
|
Partial chamber filling through the control of load volume was used to allow for viewing of the entire loaded sample in the chamber and derivation of sample concentration through counting the total number of particles present. Total concentration was defined as the count of all particles divided by the volume of all fields counted in the chamber. Similarly, partial concentration was defined as the count of all particles from each field except the 3 fields nearest the meniscus and the 1 field nearest the entry port, divided by the volume of the fields counted. Partial concentration represented the concentration a fully loaded 20-µm chamber would give with the meniscus wave concentration no longer visible on the slide.
To determine the dependence of the SS effect on capillary flow velocity, we reduced the flow rate into the 20-µm 4-chambered slide. This was achieved by plugging the chamber outlet with 0.2 µL of white ink correction fluid, which is composed of suspended particles in a volatile liquid. When the correction fluid evaporated, the slightly porous material left (which extended about 2 mm into the exit port) provided an effective porous plug. Air could still escape slowly through the pores. Sperm density of 4 aliquots from a porcine extended semen sample (26.4 M/mL) was measured as a function of position, normalized to the meniscus, which was set to 18 mm. Measurements were obtained from each of 4 separately loaded chambers, with the mean and standard deviation at each position computed.
The effect of using a deeper capillary chamber was briefly examined. Three
aliquots (each 7.0 µL of boar semen, concentration 1 M/mL) were loaded
into three 100-µm-deep chambers (courtesy of Leja Products) of identical X
and Y dimensions as the 4-chamber slide described above. The chamber was
filled to allow the meniscus to reach close to 18 mm from entry, and then
sperm counted down the centerline using the same CEROS setup with a 4x
objective. Cell acquisition was verified by playback for each field. This
technique was performed on 3 separate chambers, with the concentrations at
each position relative to the meniscus averaged across the chambers.
Experiment II. Entrained Sperm Velocity During Loading
Stroboscopic illumination techniques were utilized to determine particle
velocity at inflow (Figure 2).
The 20-µm 4-chambered slide was placed on the stage of an Olympus CH 30
microscope with a 10x NH objective, and a Sony XC-75 camera (Sony Corp,
Tokyo) interfaced with a CEROS sperm-motion analyzer. The microscope was
focused on the central plane of the chamber between the slide and its fixed
coverslip. Illumination was provided by a 1 j/pulse Xenon source (Perkin
Elmer, Salem, Mass) directed at the phase plate and strobed at 60 Hz.
Strobe-flash duration was set to less than 5 µs so that clear images of the
moving particle could be obtained. After setup, the chamber was loaded with 3
µL of sample, and 100 images are acquired at 60 Hz, digitized, and stored
using the interfaced software (IVOS V12.2c; Hamilton Thorne Biosciences). The
interval between each image was 16.67 msecond, during which the flow travels
approximately 130 µm. Images were printed out on vugraph acetate and
sequentially compared to determine particle velocity. With proper timing, the
meniscus is captured as it passes through the field of view (350 x 550
µm), and, thus, its velocity derived. Initial velocity determinations were
made on 2 types of latex beads during their inflow into the 20-µm chamber.
A mixture of 4-µm and 1.9-µm beads (Bangs Laboratories, Fishers, Ind)
was made in an aqueous medium with viscosity of 1.0 cP and at a concentration
of 5 M/mL and 3 M/mL, respectively; at these densities, bead size can easily
be distinguished by intensity and individual beads tracked without ambiguity
between successive frames. Inflow velocities in the 20-µm slide were then
determined for extended boar semen (6 M/mL) utilizing the same
procedure.
|
Experiment III. Field Studies
Two studies were undertaken to compare porcine sperm concentration
generated from use of the 4-chamber slide on a computer-automated semen
analysis system internally corrected for the SS effect (UltiMate [ULT],
Version 12.1; Hamilton Thorne Biosciences) and with a hemacytometer (Fisher
Sci) using standard techniques (Althouse et
al, 1995; WHO,
1999). In both studies and with each technique, each
sample/ejaculate was analyzed in duplicate. The analysis was double blind;
that is, the ULT and hemacytometer results were each unknown to the other
technician and were only compared by an independent third party at the
conclusion of the trial.
In the first field trial, diluted (40 M/mL) and cooled specimens (N =
12 boars) less than 6 hours postprocessing were hand delivered to the
laboratory for further analysis. Each sample was given an arbitrary
identifying number and then subdivided into 2 equal aliquots for determination
of sperm concentration using the ULT and hemacytometer. The second field trial
took place on-site at a commercial boar stud farm (Iowa Select Farms, Iowa
Falls, Ia). Gel-free ejaculates (N = 47) were collected from boars using the
gloved-hand technique. The ejaculate was given an arbitrary identifying number
and then a subsample diluted 1:40 using isothermal extender (IMV Intl,
Minneapolis, Minn). For ULT measurement in each trial, prepared samples (1.9
µL each) were loaded into two 20-µm chambers. After slide loading into
the ULT, concentration was determined in each chamber by obtaining
concentration measurements in 6 adjacent horizontal fields, with these 6
fields averaged to obtain sample concentration in that chamber. The playback
feature was used to confirm proper cell identification and absence of
agglutination. Readings from the 2 chambers within 10% of each other were
averaged and the final concentration ascribed for that sample. For both field
trials, hemacytometry was performed from subsamples of each sample used in the
20-µm chamber and further diluted 1: 4 using isothermal extender.
Estimation of sample concentration was then performed as previously outlined
(Althouse et al, 1995). Counts
from both sides of the hemacytometer were required to be within 10% of one
another to be accepted, with the reported sperm concentrations being the
average of the 2 counts.
Statistical Analysis
Fundamental descriptive statistics were utilized in all experiments to
describe quantitative data. Student's t test for paired samples was
used for performing pairwise comparisons using a 1-tailed distribution. A
1-tailed distribution was chosen for our analysis based on our understanding
of the SS effect and the results of the theoretical model
(Douglas-Hamilton et al, 2005). The power of the t test for detecting differences was established at
P < .05 in this study.
|
Results |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
|
Values of the mean total concentrations averaged over the 4 chambers for each dilution are provided in Figure 3 and summarized in Table 2. Also given in Table 2 is the partial concentration, averaged over 4 chambers. The theoretical weight dilution concentration agrees well with the total concentration estimated by CASA on the 20-µm slide, confirming that all the sperm in the 20-µm chamber are accounted for. Our results show that the average total to partial concentration ratio can be estimated at approximately 1.17, indicating that the actual concentration is approximately 17% higher than the value measured from a fully loaded 20-µm chamber.
|
Blocking the exit port with a porous plug effectively reduced inflow and prolonged filling time from approximately 2 seconds to 65-180 seconds. This change greatly diminished the size of the concentration peak at the meniscus, as expected from the analysis. The concentration averaged over 4 chambers is presented in Figure 4, showing almost flat concentration in the partially filled 20-µm slides. The meniscus concentration wave was greatly reduced, as predicted, and the measured partial and total concentrations are much closer, supporting the concept that the SS effect has not had time to develop at the low sample-flow rates.
|
Sperm concentration in the 100-µm chamber when flow ceases after 18 mm is shown in Figure 5 as a function of position. No sharp meniscus wave was observed in the 100-µm chamber, although cell concentration did increase gradually from entry to exit.
|
Assessment for an SS effect on human sperm samples (N = 3) with a viscosity of 1.02 cP, is presented in Figure 6. The high-concentration wave at the meniscus is clearly visible in all cases and is similar to that seen in the case of motile boar sperm (see Figure 3).
|
Experiment II. Inflow Entrained Sperm Velocity
Velocity distribution of the 4-µm-diameter beads is shown in
Figure 7. A strong velocity
peak for the beads was found near the position where flow velocity
approximates 0.8 cm/s, with no slow-velocity beads present. This finding
suggests a distinct transport mechanism that draws beads from the low-velocity
to the high-velocity planes in the Poiseuille flow. It is consistent with
beads transported in the high-velocity SS-preferred planes in the 20-µm
slide, at the fractional positions of ß = .27 and ß = .73, where
ß is the distance from plane to wall normalized to the chamber depth.
|
Velocity distribution of the 1.9-µm beads is shown in Figure 8. Although only 34 beads were counted at the 3 M/mL concentration, there is a clear trend for the 1.9-µm beads to establish low- and high-velocity regions. The solid line represents a prediction of the velocity frequency curve for uniformly distributed beads in Poiseuille flow. The distribution of beads observed in this experiment is close to uniform: that is, the 1.9-µm beads have not had enough time to segregate in the SS planes because the transverse velocity varies as the cube of the bead radius, showing large standard deviation in velocity. No leading-edge concentration peak was observed.
|
The inflow velocity distribution of live, motile boar sperm is shown in Figure 9. The sperm show the same distribution as the 4-µm latex-bead spheres, with the sperm concentration increasing sharply near the meniscus and decreasing behind it. The mean inflow velocity (ie, meniscus velocity) is shown as a solid vertical line. As observed, almost all the sperm in the postmeniscus region are moving at a velocity higher than the mean fluid flow. These high-velocity sperm accumulate at the meniscus leading edge thereby reducing the concentration of sperm measured behind the meniscus. Denoting the maximum velocity along the central plane of the chamber as Vmax, the velocity corresponding to the SS layer is approximately 0.8 Vmax, or about 7400 µm/s.
|
Experiment III. Field Studies
Results of the 2 field studies are shown in
Figure 10. When comparing
concentrations determined by hemacytometry to the ULT that has been corrected
for the SS effect, the correlation coefficient was r2 =
.936, with a regression gradient of 0.984.
|
|
Discussion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
When determining total, partial, and dilution concentrations within
chambers using total counts, we expect that the ratio of the total mean
concentration to the expected weight percentage dilution concentration
approaches 1, if a representative sample of sperm was counted. Therefore, the
ratio of the total to the partial measured concentration reflects the
difference between the actual number of sperm present and what would normally
be counted in a capillary slide. A total-to-partial concentration ratio of
1.17 was found, establishing that the actual concentration was approximately
17% higher than the value measured from a fully loaded 20-µm chamber. Note
that this ratio is only approximate because it depends on estimation of the
number of cells in the meniscus peak. This number is of limited accuracy
because the size of the measuring area (
mm) is similar to the
width of the meniscus concentration peak (1 mm; see
Figure 3), and the cells in the
concentration peak will be undercounted. We can, however, conclude that these
results provide direct evidence that the meniscus high-concentration wave
exists and that the concentration behind the wave is lower than the sample
concentration.
Blocking the exit port with a porous plug was performed to greatly reduce inflow velocity. As expected, reduction of velocity effectively eliminated the SS effect (B3 in Table 2; Figure 4) in the 20-µm chamber. Likewise, particle concentration did not develop a significant meniscus wave in the 100-µm slide because particles had not had time to migrate sufficiently to the faster regions of flow. This is in accordance with the prediction that the time to develop the SS segregation effect varies as 1/L2, where L is chamber depth. Therefore, a negligible SS effect can also be expected in the 100-µm depth hemacytometer, in which the flow travels less than 10 mm from point of entry. Indeed, when we repeated these experiments using boar sperm in a hemacytometer, we found that the inflow was irregular because of the V loading incision in the Improved Neubauer hemacytometer. Nevertheless, no concentration increase at the leading edge of the meniscus wave was observed.
In experiment II, inflow velocity distribution of boar sperm and 2.0-µm- and 0.95-µm-radius beads was assessed. Results confirmed that the SS effect occurs for the larger sperm and 2.0-µm beads as they flow into the chamber and show that, as expected, the beads and sperm are in fact entrained into more rapidly moving layers of the flow, move in more rapidly than the mean fluid velocity, and as a result are measured behind the meniscus with reduced concentration. The SS-stable planes are expected to be at the fractional positions of ß = 0.27 and ß = 0.73. On the other hand, the larger spread in velocity of the 1.9-µm beads is expected because the predicted degree of segregation increases monotonically with the figure of merit, which is proportional to the cube of the particle radius (Douglas-Hamilton et al, 2005). The smaller particles do not have time to segregate into a particular stream, and they remain spread through all velocities.
Last, 2 field studies were performed comparing diluted semen concentrations estimated by standard hemacytometry and in capillary-loaded 20-µm slides using a computer-automated semen-analysis system designed to account for the SS effect (eg, ULT). A very high degree of correlation (r2 = .936) was found between estimates obtained from the ULT and hemacytometer. Additionally, high amounts of precision, accuracy, and repeatability were observed between the 2 sites.
Sperm segregation can be expected to occur whenever a sperm sample of low viscosity is introduced into a narrow capillary-loaded slide. Therefore, a correction is required if these slides are to be used to estimate sperm concentration. This segregation is due to the SS effect and is important to consider in industries where samples examined are always diluted and of low viscosity. Oversight in accounting for this SS effect can lead to faulty estimation of sperm concentration when using thin, capillary-loaded slides, which induce the SS effect. The oversight of the SS effect in capillary-loaded slides most likely accounts for the differences reported in prior published articles in which variability existed in estimates of sperm concentration between hemacytometry and CASA (Johnson et al, 1996a; Seaman et al, 1996; Mahmoud et al, 1997; Brazil et al, 2004; and others).
We have derived a simple model to predict the concentration change and compared it with experiments using 20-µm chambers. The model derived is based on spheres suspended in Poiseuille flow. Sperm behavior in gradient flow is complicated by the presence of the tail: this suppresses rotation and increases the sperm effective area. However, the model derived for spheres gives relatively good prediction of sperm behavior in the experiments, and SS segregation of sperm is observed similar to that expected for 4-µm-diameter spheres. Our model and observations also show that the SS effect is insignificant in hemacytometers due to the larger chamber depth and small size. The hemacytometer remains the gold standard. Using SS-compensated UltiMate CASA analysis, good agreement was obtained in a field test between hemacytometer and UltiMate for boar sperm analysis.
|
Conclusion |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
References |
|---|
|
TopAbstractMaterials and MethodsResultsDiscussionConclusionReferences |
|---|
Brazil C, Swan SH, Drobnis EZ, Liu F, Wang C, Redmon JB, Overstreet
JW. Standardized methods for semen evaluation in a multicenter research study.
J Androl. 2004; 25:635-644.
Douglas-Hamilton DH, Smith N, Kuster CE, Vermeiden JPW, Althouse
GC. Particle distribution in low volume capillary-loaded chambers.
J Androl. 2005; 26:107-114.
Johnson JE, Boone WR, Blackhurst DW. Manual versus computer-automated semen analyses, Part I. Comparison of counting chambers. Fertil Steril. 1996a; 65:150-55.[Medline]
Johnson JE, Boone WR, Blackhurst DW. Manual versus computer-automated semen analyses, Part III. Comparison of old versus new design MicroCell chambers. Fertil Steril. 1996b; 65:446-447.[Medline]
Mahmoud AMA, Depoorter B, Piens N, Comhaire FH. The performance of 10 different methods for the estimation of sperm concentration. Fertil Steril. 1997; 68:340-345.[Medline]
Mortimer D, Shu MA, Tan R. Standardization and quality control of
sperm concentration and sperm motility counts in semen analysis.
Hum Reprod. 1986; 1:299-303.
Seaman EK, Goluboff E, BarChama N, Fisch H. Accuracy of semen counting chambers as determined by the use of latex beads. Fertil Steril. 1996; 66:662-665.[Medline]
Segre G, Silberberg A. Behavior of macroscopic rigid spheres in Poiseuille flow. J Fluid Mech. 1961; 14:115.
Vasseur P, Cox RG. The lateral migration of a spherical particle in two-dimensional shear flows. J Fluid Mech. 1976; 78:385-413.
World Health Organization (WHO). Laboratory Manual for the Examination of Human Semen and Sperm-Cervical Mucus Interaction. 4th ed. Cambridge: Cambridge University Press; 1999 :14-17.
This article has been cited by other articles:
|
|
 |
|
  S. Rijnders, J. G. M. Bolscher, J. McDonnell, and J. P. W. Vermeiden Filling Time of a Lamellar Capillary-Filling Semen Analysis Chamber Is a Rapid, Precise, and Accurate Method to Assess Viscosity of Seminal Plasma J Androl, July 1, 2007; 28(4): 461 - 465. [Full Text] [PDF]
|
|
|
|
 |
|
 H. D. Guthrie and G. R. Welch Use of fluorescence-activated flow cytometry to determine membrane lipid peroxidation during hypothermic liquid storage and freeze-thawing of viable boar sperm loaded with 4, 4-difluoro-5-(4-phenyl-1,3-butadienyl)-4-bora-3a,4a-diaza-s-indacene-3-undecanoic acid J Anim Sci, June 1, 2007; 85(6): 1402 - 1411. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. L. Vadnais and K. P. Roberts Effects of Seminal Plasma on Cooling-Induced Capacitative Changes in Boar Sperm J Androl, May 1, 2007; 28(3): 416 - 422. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
L. Bjorndahl and C. L. R. Barratt Semen Analysis: Setting Standards for the Measurement of Sperm Numbers J Androl, January 1, 2005; 26(1): 11 - 11. [Full Text] [PDF]
|
|
|
|
|
|
D. H. Douglas-Hamilton, N. G. Smith, C. E. Kuster, J. P. W. Vermeiden, and G. C. Althouse Particle Distribution in Low-Volume Capillary-Loaded Chambers J Androl, January 1, 2005; 26(1): 107 - 114. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
