Abstract
Variation in the quality of ejaculate produced by male American lobsters, Homarus americanusMilne Edwards, 1837, has been previously described, but never quantified. This study examined the size and composition of ejaculates produced by 111 males ranging from 60 to |$108\,{\rm{mm}}$| in carapace length (CL). Ejaculates were obtained via electrical stimulation, photographed and then processed for histology. Half of the males produced an ejaculate from each gonopore, 29% produced only one ejaculate, and the remainder (21%) produced none. Males as small as |$64\,{\rm{mm\, CL}}$| produced an ejaculate containing sperm. Ejaculate weight increased with male size, but there was a negative relationship between ejaculate weight and the percent of the ejaculate that was composed of sperm mass. Variation observed in the size and composition of ejaculates produced by similarly-sized males indicates that not all males invest equally in reproduction. Additionally, larger males may invest disproportionately more in the sperm plug (acellular component), possibly as paternal assurance.
Introduction
The potential for sperm limitation in exploited crustaceans has been gaining attention over the past few decades, as researchers and fisheries scientists are beginning to appreciate the extent to which male contributions can affect the reproductive output of females (MacDiarmid and Butler, 1999; Jivoff, 2003; Sato et al., 2007; Sainte-Marie et al., 2008). A positive relationship between male size and ejaculate load (sperm number or spermatophore size) has been documented in several decapod species including the Atlantic blue crab Callinectes sapidusRathbun, 1896 (Jivoff, 2003), spiny king crab Paralithodes brevipes (H. Milne Edwards and Lucas, 1841) (Sato et al., 2005), coconut crab Birgus latro (Linnaeus, 1767) (Sato et al., 2008), spiny rock lobster Jasus edwardsii (Hutton, 1875) (MacDiarmid and Butler, 1999), Caribbean spiny lobster Panulirus argusLatreille, 1804 (Butler et al., 2011), and American lobster Homarus americanus H. Milne Edwards, 1837 (Gosselin et al., 2003). However, the metric used to describe male contributions or reproductive potential differed somewhat in these studies. For example, while there was a positive relationship between male body size and vas deferens weight for the spiny rock lobster, there was no clear correlation between the number of sperm within the vas deferens and male body size (MacDiarmid, 1989). Moreover, in some species ejaculate did not scale with male body size, and its composition (ratio of sperm to seminal fluids) varied between male morphometric types, e.g., snow crab Chionoecetes opilio (Fabricius, 1788) (Sainte-Marie and Lovrich, 1994). Thus, a simple relationship between the amount of ejaculate and male body size might not accurately represent male reproductive contributions.
American lobsters generally mate shortly after the female has molted (reviewed by Atema and Voigt, 1995). The male uses his first pair of pleopods to form a gonopod and transfer the ejaculate to the seminal receptacle of the female, where it is stored until she spawns. Spawning can occur several months to a year after mating, so the sperm must remain viable for a lengthy period of time. Gosselin et al. (2003) reported differences in seminal receptacle loads (based on weight) of females that mated with small versus large males. However, results from previous electro-stimulation studies showed that ejaculate size was not necessarily related to sperm quantity (some ejaculates contained no sperm, Aiken et al., 1984; based on gross visual description), or quality (Talbot et al., 1983). Thus, while females may receive more ejaculate from larger males (Gosselin et al., 2003), weight may not accurately reflect actual sperm content.
While males as small as |$41\,{\rm{mm}}$| carapace length (CL) produce sperm (Krouse, 1973; Conan et al., 2001; data based on vasa deferentia dissections), it seems unlikely that males this small are capable of mating with even the smallest mature females (between 80 and |$90\,{\rm{mm\, CL}}$| in Boothbay, Maine; Krouse, 1973). Although small males may have sperm in their vasa deferentia, they may not be capable of producing an ejaculate. Émond et al. (2010) found that male size-at-maturity estimates, based on the dimensions of the vas deferens and gonopods, resulted in larger size-at-maturity values than those based solely on the presence of sperm in their vasa deferentia. Furthermore, females generally choose to mate with dominant males (see Atema, 1986 for review), and mating success is greater when males are larger than females (Templeman, 1934). Hence, male mating activity may be more related to lobster behavior than actual physical ability, with small males losing out to larger, more dominant males (Stein et al., 1975; Karnofsky and Price, 1989). Nevertheless, it is possible that, in the absence of larger males, small males might gain mating opportunities, if they are capable of producing ejaculates containing viable sperm.
The purpose of this study was to investigate the production of ejaculates by male American lobsters, quantify their composition, and examine the relationships between male size and ejaculate characteristics. To our knowledge this is the first study to examine the relationship between the size of male lobsters and the composition of the ejaculate they produce.
Materials and Methods
Lobster Collection and Holding
Male lobsters were collected using traps from the inshore waters of coastal New Hampshire to northern Massachusetts (USA). Legal-sized lobsters |$(\geqslant83\, {\rm{mm\, CL}})$| were obtained primarily from commercial lobstermen, while sublegal-sized lobsters, between 60 and |$82\,{\rm{mm\, CL}}$|, were obtained during scientific surveys conducted by staff from the University of New Hampshire (UNH), New Hampshire Fish & Game, and Massachusetts Division of Marine Fisheries. Lobsters were collected during summer and fall months of 2011 and 2012, and held with other males in laboratory tanks supplied with flow-through seawater. Holding times prior to testing ranged from 3 to 36 days. Water temperatures ranged from 12.2 to |$18.8^\circ {\rm{C}}$| in 2011, and from 13.7 to |$19.3^\circ {\rm{C}}$| in 2012. All males were measured and tagged (knuckle band cinch-up tags with individual IDs) upon arrival at the laboratory. In 2011 and 2012, work was conducted at the UNH Coastal Marine Laboratory in New Castle, NH, USA and the University of Massachusetts Hodgkin’s Cove Marine Station in Gloucester, MA, USA, respectively.
The ejaculate of a male American lobster is a single cohesive structure, often referred to as a spermatophore, which is composed of an internal tubular-shaped sperm mass encapsulated by the primary spermatophoric layer and surrounded by acellular materials (Kooda-Cisco and Talbot, 1982). There appears to be some confusion in the literature as to whether the entire structure, including the acellular materials, should be referred to as a ‘spermatophore’ (see, e.g., Kooda-Cisco and Talbot, 1982; Aiken et al., 1984; Subramoniam, 1991), or whether the term ‘spermatophore’ should/could refer solely to the sperm mass and primary spermatophoric layer (Matthews, 1954; Haley, 1984; Martin et al., 1987). Thus, for clarity we use the term ‘ejaculate’ to refer to the entire structure (including the acellular materials) in this paper.
Stimulation Technique
Individual males were secured ventral-side up for the electro-stimulation procedure described by Kooda-Cisco and Talbot (1983). Positive and negative electrodes were held against the membranes surrounding the gonopore at the base of the fifth pereiopod (Fig. 1) and a 12-16 V stimulus was applied using a SD9 Grass Stimulator (Grass Instruments, Quincy, MA, USA). If an ejaculate was extruded, it was collected with forceps and immediately processed. The procedure was then repeated on the opposite gonopore. If no ejaculate was produced, the procedure was repeated again twice, for a maximum total of three attempts per gonopore. Immediately after testing, each male was molt-staged using the pleopod staging technique (Aiken, 1973) and classified as premolt |$({\rm{pleopod\, stage}}\, \geqslant 3.0)$| or intermolt |$({\rm{pleopod\, stage}}\, \lt 3.0)$|.
Placement of the electrodes used to deliver the electrical stimulus around the gonopore at the base of the 5th pereiopod (walking leg). This figure is published in colour in the online edition of this journal, which can be accessed via http://booksandjournals.brillonline.com/content/journals/1937240x.
Upon extrusion, each ejaculate was immediately weighed and photographed along with a reference scale bar using a digital camera (Fig. 2). The ejaculate was then submerged in seawater in a petri dish and photographed with transmitted light under a dissecting microscope.
Macroscopic image of a freshly extruded lobster ejaculate. The internal sperm mass (indicated by the white arrow) extends from the left tip (leading edge) and tapers off just past the center of the ejaculate and is surrounded by an acellular material (clear substance). The scale bar is marked in |$1{\text{-mm}}$| increments. This figure is published in colour in the online edition of this journal, which can be accessed via http://booksandjournals.brillonline.com/content/journals/1937240x.
Male lobsters can produce an ejaculate from each gonopore, and two ejaculates from the same male cannot be considered independent data points. Thus, for those males that produced two ejaculates, we randomly selected either the ejaculate produced from the right or left side to include in further analysis.
Histology
Each ejaculate was fixed in Bouin’s solution for 3 to 10 days, then rinsed and stored in 70% ethanol. Fixed ejaculates were dehydrated in a series of graded concentrations of ethanol, cleared in a xylene agent, and then embedded in paraffin wax within a vacuum chamber. The embedding time lasted from 12 to 24 hours to allow complete penetration of paraffin wax into the tissues. Blocks were sectioned longitudinally at a thickness of |$5 \,\mu{\rm{m}}$| on a rotary microtome. Serials sections were mounted on glass slides and stained with modified Masson’s trichrome, a modification by Goldener (Martoja and Martoja-Pierson, 1967). Final mounts were made with glass cover slips using a mounting resin.
Image Analysis
The composition of each ejaculate was determined using two different techniques. The first method used the two-dimensional image taken of the entire structure, while the second used the serial histological images of the sectioned ejaculate (three-dimensional technique).
Two-Dimensional Analysis (2D)
Digital images were analyzed using ImageJ image processing software (Rasband, 2014). Using the polygon selection tool, the outline of the entire ejaculate (to calculate the total area; Fig. 3A) and then the sperm mass (sperm mass area; Fig. 3B) were manually selected. Separating the sperm mass from the acellular material was subjective and based on distinguishing the darker centrally located area of the ejaculate containing sperm from the lighter surrounding area of the acellular material. In addition to the digital images, the macroscopic photos of fresh ejaculate (see Fig. 2), in which the sperm mass appeared milky white in contrast to the translucent acellular material, were consulted to assist with determining the border of the sperm mass. The composition of the ejaculate, in terms of the percent of sperm mass, was calculated as follows:
Two-dimensional image analysis of the ejaculate from male 6135L. Manually selected outline of (A) the entire ejaculate and (B) the sperm mass. |${\rm{Percent\, sperm\, mass}}\, =\, (13,718\, {\rm{pixels/40,668\, pixels}}) \,\times \,100\, = \,33.7\%$|. This figure is published in colour in the online edition of this journal, which can be accessed via http://booksandjournals.brillonline.com/content/journals/1937240x.
Three-Dimensional Analysis (3D)
Each stained section of the ejaculate was photographed under a dissecting microscope with transmitted light, producing a series of images that together created a three-dimensional representation. The photograph of the first serial section of the entire ejaculate was opened in ImageJ (Fig. 4A), and the scale for each series of images was set using the reference scale within the image and the “global” option. For each section, the image was split into color channels (red, blue, green) and the resulting black and white image with the best contrast was selected. If necessary, contrast was further adjusted, then the threshold tool was engaged to outline the entire section (Fig. 4B), and the area (pre-selected as a measurement option) was calculated by selecting “analyze particles.”
Three-dimensional technique used to quantify material in histological sections of male ejaculate. This is one of 14 sections from the left-side ejaculate of male 6135 (same ejaculate as shown in Fig. 3). (A) Stained section of the entire ejaculate |$(5 \,\mu{\rm{m\, thick}})$|; the sperm mass appears blue and granular and is surrounded by the red-stained primary spermatophoric layer; scale bar shows |$1\,{\rm{mm}}$| increments. (B) Outline of the entire ejaculate section after thresholding in ImageJ. (C) Outline of the sperm mass imported from Photoshop into ImageJ. This figure is published in colour in the online edition of this journal, which can be accessed via http://booksandjournals.brillonline.com/content/journals/1937240x.
Each image was also opened in Photoshop (Adobe Photoshop Elements 9.0), and the “magic wand” tool was used to identify and select the sperm mass within each histological section of the ejaculate. This selection was then copied and opened in ImageJ (maintaining the correct scale). The color image was split into channels, and the darkest and most consistently contrasted channel was used to measure the area of the sperm mass (Fig. 4C). Results were copied into an Excel spreadsheet, and the process was repeated with the images of all the serial sections.
The total area of the ejaculate for each serial section was multiplied by the section’s thickness |$(0.005\, {\rm{mm}})$| to generate a volume and these were added together to estimate the volume of the ejaculate sampled. Similarly, volumes were calculated for the sperm mass within each section and summed to represent the total volume of sperm. The percentage of the sperm mass was then calculated for each ejaculate as follows:
For ejaculates with more than approximately 25 serial sections, every other section was analyzed, and a 2-section running average (using only the sections immediately preceding and following the blank section) was used to gap-fill those sections not analyzed.
Since a complete set of sections was not obtained for each sample, and the amount of each ejaculate processed varied, a standardized calculation of sperm mass volume within the central section of the ejaculate was generated. This calculation was based on the first ten serial sections of the ejaculate, representing its longitudinal central-most portion. Sections were |$0.005\,{\rm{mm}}$| thick, so this central portion from each ejaculate was |$0.05\,{\rm{mm}}$| thick. The volume of the sperm mass was calculated for this portion, and presented as “central sperm mass volume.”
To estimate the total volume of the sperm mass within the ejaculate, measurements of the length and an average width (based on multiple (3-4) measurements of width) of each freshly extruded ejaculate were taken in ImageJ from the macroscopic image (Fig. 2, e.g.). Assuming that the ejaculate is roughly the shape of a cylinder, the volume was calculated as follows:
The percent of sperm generated from the 3D analysis was then scaled up to estimate the total volume of the sperm mass using the following equation:
To generate preliminary estimates of the number of sperm cells within a sperm mass, the volume of an individual sperm cell was estimated using length and width measurements of 25 sperm cells (5 each from 5 lobsters) taken under |$400\times\, {\rm{magnification}}$|. Assuming that sperm cells are roughly cylindrical in shape (excluding the spikes), sperm cell volume was calculated using the formula for a cylinder. This value |$(0.00019\, {\rm{mm}}^3)$| was then used to calculate the number of sperm within a sperm mass as follows:
Statistics
An exponential model was used to describe the relationships between male size and ejaculate characteristics, and between ejaculate weight and composition. One ejaculate with weight |$\lt 0.0001\, {\rm{g}}$| (this was below the detection threshold of the scale) was excluded from modeling. The model was fit using the generalized linear model (GLM) with gamma error distribution and a log-link function in R (Version 3.0.3; R Core Development Team, 2014). The gamma error distribution was selected based on the skewed variance distribution. The model took the form:
Statistical significance of the model and parameter estimates were generated (Analysis of Deviance table, |$F$| and |$p$| values reported) and a generalized |${R^2}$| (Naglekerke’s adjustment of the Cox-Snell |${R^2}$|, see Faraway, 2006) was calculated for each model fit, which describes the proportion of the deviance explained by the model. Model fit was checked using a goodness-of-fit test and plots of the residuals (Faraway, 2006).
Results
Production of Ejaculates
A total of 111 males were tested, ranging in size from 60 to |$108\,{\rm{mm\, CL}}$| (Table 1). Half (50.5%, 56 males) of the males tested produced an ejaculate from each gonopore, 28.8% (32 males) produced an ejaculate from only one gonopore, and 20.7% (23 males) did not produce an ejaculate (Table 1). The proportion of males that produced at least one ejaculate was highest in the |$91{\text{-}}95\, {\rm{mm}}$| size bin (0.94), while the proportion of males that produced 2 ejaculates was highest in the |$86{\text{-}}90\, {\rm{mm}}$| size bin (0.71) (Table 1). The smallest male to produce an ejaculate was |$61\,{\rm{mm\, CL}}$|. For those males that produced two ejaculates, the weights of the right vs left-side ejaculates were correlated (Pearson’s |$r = 0.58,\,p \lt 0.01$|). Within each |$5\,{\rm{mm\, CL}}$| size bin, there was no difference in the mean weight of ejaculates from males that produced one versus two (one-way ANOVAs for each size bin, |$p \gt 0.1$| for each bin). Finally, there was no correlation between the time males were held prior to the electro-stimulation procedure (3 to 36 days, mean |$13.4\, \pm \,0.4$|) and the number of ejaculates produced (Pearson’s |$r = - 0.26,\,p = 0.34$|).
In each |$5\,{\rm{mm\, CL}}$| size bin, the total number of males tested, and the number of males that produced 0, 1, or 2 ejaculates is presented. Shown in parentheses for males that produced 1 or 2 ejaculates is the number of ejaculates included in analyses. The total number of ejaculates included in analyses and the mean weight |$(\pm{\rm{SE}})$| of the ejaculates are also shown for each |$5\,{\rm{mm\, CL}}$| size bin.
| Size bin (mm CL) | No. of males tested | No. Produced 0 ejaculates | No. of produced 1 ejaculate (No. of ejaculates included) | No. of produced 2 ejaculates (No. of ejaculates included) | Total ejaculates included in analyses | Mean (SE) ejaculate weight (g) |
|---|---|---|---|---|---|---|
| 60-65 | 12 | 5 | 4 (3) | 3 (2) | 5 | 0.003 (0.0008) |
| 66-70 | 12 | 4 | 4 (2) | 4 (4) | 6 | 0.006 (0.0019) |
| 71-75 | 13 | 4 | 4 (4) | 5 (4) | 8 | 0.007 (0.0011) |
| 76-80 | 14 | 3 | 3 (3) | 8 (6) | 9 | 0.013 (0.0015) |
| 81-85 | 13 | 2 | 3 (3) | 8 (6) | 9 | 0.016 (0.0049) |
| 86-90 | 14 | 1 | 3 (1) | 10 (7) | 8 | 0.018 (0.0038) |
| 91-95 | 18 | 1 | 7 (5) | 10 (9) | 14 | 0.024 (0.0044) |
| 96+ | 15 | 3 | 4 (3) | 8 (6) | 9 | 0.032 (0.0057) |
| Total | 111 | 23 | 32 (24) | 56 (44) | 68 | 0.017 (0.0018) |
| Size bin (mm CL) | No. of males tested | No. Produced 0 ejaculates | No. of produced 1 ejaculate (No. of ejaculates included) | No. of produced 2 ejaculates (No. of ejaculates included) | Total ejaculates included in analyses | Mean (SE) ejaculate weight (g) |
|---|---|---|---|---|---|---|
| 60-65 | 12 | 5 | 4 (3) | 3 (2) | 5 | 0.003 (0.0008) |
| 66-70 | 12 | 4 | 4 (2) | 4 (4) | 6 | 0.006 (0.0019) |
| 71-75 | 13 | 4 | 4 (4) | 5 (4) | 8 | 0.007 (0.0011) |
| 76-80 | 14 | 3 | 3 (3) | 8 (6) | 9 | 0.013 (0.0015) |
| 81-85 | 13 | 2 | 3 (3) | 8 (6) | 9 | 0.016 (0.0049) |
| 86-90 | 14 | 1 | 3 (1) | 10 (7) | 8 | 0.018 (0.0038) |
| 91-95 | 18 | 1 | 7 (5) | 10 (9) | 14 | 0.024 (0.0044) |
| 96+ | 15 | 3 | 4 (3) | 8 (6) | 9 | 0.032 (0.0057) |
| Total | 111 | 23 | 32 (24) | 56 (44) | 68 | 0.017 (0.0018) |
In each |$5\,{\rm{mm\, CL}}$| size bin, the total number of males tested, and the number of males that produced 0, 1, or 2 ejaculates is presented. Shown in parentheses for males that produced 1 or 2 ejaculates is the number of ejaculates included in analyses. The total number of ejaculates included in analyses and the mean weight |$(\pm{\rm{SE}})$| of the ejaculates are also shown for each |$5\,{\rm{mm\, CL}}$| size bin.
| Size bin (mm CL) | No. of males tested | No. Produced 0 ejaculates | No. of produced 1 ejaculate (No. of ejaculates included) | No. of produced 2 ejaculates (No. of ejaculates included) | Total ejaculates included in analyses | Mean (SE) ejaculate weight (g) |
|---|---|---|---|---|---|---|
| 60-65 | 12 | 5 | 4 (3) | 3 (2) | 5 | 0.003 (0.0008) |
| 66-70 | 12 | 4 | 4 (2) | 4 (4) | 6 | 0.006 (0.0019) |
| 71-75 | 13 | 4 | 4 (4) | 5 (4) | 8 | 0.007 (0.0011) |
| 76-80 | 14 | 3 | 3 (3) | 8 (6) | 9 | 0.013 (0.0015) |
| 81-85 | 13 | 2 | 3 (3) | 8 (6) | 9 | 0.016 (0.0049) |
| 86-90 | 14 | 1 | 3 (1) | 10 (7) | 8 | 0.018 (0.0038) |
| 91-95 | 18 | 1 | 7 (5) | 10 (9) | 14 | 0.024 (0.0044) |
| 96+ | 15 | 3 | 4 (3) | 8 (6) | 9 | 0.032 (0.0057) |
| Total | 111 | 23 | 32 (24) | 56 (44) | 68 | 0.017 (0.0018) |
| Size bin (mm CL) | No. of males tested | No. Produced 0 ejaculates | No. of produced 1 ejaculate (No. of ejaculates included) | No. of produced 2 ejaculates (No. of ejaculates included) | Total ejaculates included in analyses | Mean (SE) ejaculate weight (g) |
|---|---|---|---|---|---|---|
| 60-65 | 12 | 5 | 4 (3) | 3 (2) | 5 | 0.003 (0.0008) |
| 66-70 | 12 | 4 | 4 (2) | 4 (4) | 6 | 0.006 (0.0019) |
| 71-75 | 13 | 4 | 4 (4) | 5 (4) | 8 | 0.007 (0.0011) |
| 76-80 | 14 | 3 | 3 (3) | 8 (6) | 9 | 0.013 (0.0015) |
| 81-85 | 13 | 2 | 3 (3) | 8 (6) | 9 | 0.016 (0.0049) |
| 86-90 | 14 | 1 | 3 (1) | 10 (7) | 8 | 0.018 (0.0038) |
| 91-95 | 18 | 1 | 7 (5) | 10 (9) | 14 | 0.024 (0.0044) |
| 96+ | 15 | 3 | 4 (3) | 8 (6) | 9 | 0.032 (0.0057) |
| Total | 111 | 23 | 32 (24) | 56 (44) | 68 | 0.017 (0.0018) |
Characteristics of the Ejaculate
For those males that produced two ejaculates, only the ejaculate from the randomly selected side was used for the rest of the analyses presented. A total of 68 ejaculates (out of 144), produced by 68 individual males, were completely extruded and could be included in further analyses (Table 1).
Ejaculates varied in weight from a minimum of |$\lt0.0001\, {\rm{g}}$| (below the detection threshold of the scale) to a maximum of 0.065 g. There was a significant positive relationship between male size and ejaculate weight |$(F = 61.981,\,p \lt 0.001,\,n = 64)$|, with male size explaining roughly half of the deviance |$(R_{{\rm{gen}}}^2 = 0.55)$| (Fig. 5). Variability in ejaculate weight was greater at larger male sizes. Ejaculates produced by the six males that were pre-molt (|${\rm{pleopod\, stage}}\, \geqslant 3.0$|; Aiken, 1973) followed the same general pattern as those produced by inter-molt males (Fig. 5).
The relationship between ejaculate weight (g) and male size (mm CL); each data point represents one ejaculate (symbols indicate male molt stage) and the solid line represents a GLM fit to the data (“Model fit”); model parameter estimates: |${\beta _0} = - 9.227\,(\pm0.58),\,{\beta _1} = 0.060\,(\pm0.007),\,R_{{\rm{gen}}}^2 = 0.55$|.
Ejaculate Composition-2D Analysis
The majority (82%) of ejaculates were composed of 20-70% sperm mass, with only two ejaculates composed of |$\gt70\%$| sperm mass (71 and 74% sperm mass) (Fig. 6A). Several ejaculates were composed of |$\lt20\%$| sperm mass, most of which (6/9) were produced by males |$\gt85\, {\rm{mm\, CL}}$|. Although there was no apparent relationship between male size and ejaculate composition (% of ejaculate composed of sperm mass), no male |$\gt92\, {\rm{mm\, CL}}$| produced an ejaculate that was |$\gt50\%$| sperm mass. Ejaculates produced by the pre-molt males appeared to follow the same general pattern in composition as those produced by inter-molt males (Fig. 6A, B).
Composition of ejaculates, as percentages of sperm (% sperm mass) for data generated from the two-dimensional (2D) and three-dimensional (3D) image analysis techniques. (A) Ejaculate composition (% sperm mass, 2D technique) in relation to male size (mm CL) |$(n = 60)$|. (B) Relationship between ejaculate composition (% sperm mass, 2D technique) and weight (g): each data point represents one ejaculate |$(n = 60)$| and the solid line represents a GLM fit to the data (“model fit”); model parameter estimates: |${\beta _0} = 3.813\,(\pm0.07),\,{\beta _1} = - 12.763\,(\pm3.18),\,R_{{\rm{gen}}}^2 = 0.213$|. (C) Ejaculate composition (% sperm mass, 3D technique) by male size (mm CL) |$(n = 60)$|. (D) Relationship between the ejaculate composition (% sperm mass, 3D technique) and ejaculate weight (g): each data point represents one ejaculate |$(n = 60)$| and the solid line represents a GLM fit to the data (“model fit”); model parameter estimates: |${\beta _0} = 3.321\,(\pm0.14),\,{\beta _1} = - 23.401\,(\pm6.80),\,R_{{\rm{gen}}}^2 = 0.15$|.
Heavier ejaculates were generally composed of a lower percentage of sperm mass (Fig. 6B). There was a significant negative relationship between the percentage of sperm mass and the ejaculate weight |$(F = 15.051,\,p \lt 0.001,\,R_{{\rm{gen}}}^2 = 0.214)$| when one ‘outlier’ data point (where |$\%{\text{ sperm mass }}= 0$|) was excluded (required for the model to run), and model diagnostics indicated reasonable fit |$(p = 0.373)$| (Fig. 6B, note outlier data point not removed from plot). The ejaculate weight explained a small portion of the deviance in ejaculate composition. The variability in ejaculate composition appeared to decrease with increasing ejaculate weight, although there were fewer heavy ejaculates.
Ejaculate Composition-3D Analysis
Based on the three-dimensional (3D) technique, ejaculate composition varied from 0 to 52% sperm mass. Similar to the 2D results, the ejaculate composition as determined by the 3D technique showed no pattern in relation to male size (Fig. 6C). Also like the 2D method, the % sperm mass within the ejaculate generally decreased at heavier weights (again, after exclusion of ejaculates with 0% sperm mass; |$F = 9.266,\,p = 0.004$|), and model diagnostics indicated an acceptable but poor fit |$(p = 0.06)$| (Fig. 6D). There was a significant positive correlation between the ejaculate composition determined by the 3D technique versus the 2D technique (Pearson’s |$r = 0.605,\,p \lt 0.01$|), however the 2D technique yielded consistently higher values of % sperm mass than the 3D method (Fig. 7).
Relationship between the results of the 2D and 3D techniques for determining ejaculate composition (% sperm mass) (|$n = 52$| ejaculates analyzed with both techniques). Pearson’s |$r = 0.60$|. Dashed line represents a 1:1 reference line.
The actual volume of the sperm mass was calculated for the first ten serial sections of each ejaculate, representing a standard thickness of |$0.05\,{\rm{mm}}$| sampled from the center of each ejaculate. Central sperm mass volume generally increased with increasing male size |$(F = 6.42,\,p = 0.014)$|, although again, there was a large amount of variation resulting in a poor model fit |$(p = 0.005)$| (Fig. 8). Only about a tenth of the deviance in central sperm mass volume was explained by the model |$(R_{{\rm{gen}}}^2 = 0.098)$|.
The relationship between central sperm volume (sperm mass volume |${\rm{mm}}^3$|; central |$50 \,\mu{\rm{m}}$| of ejaculate) and male size (mm CL). Each data point represents one ejaculate |$(n = 59)$|. The dashed line represents a GLM fit to the data (“Model fit”); model parameter estimates: |${\beta _0} = - 5.542\,(\pm0.685),\,{\beta _1} = 0.039\,(\pm0.008),\,R_{{\rm{gen}}}^2 = 0.24$|.
Since only a small portion of each ejaculate could be processed for histology, it was desirable to describe the proportion of the ejaculate that was represented by the results of the 3D technique. An estimate for the total volume of each ejaculate was calculated based on length and width measurements taken from digital macroscopic images. This was possible for 60 of the 68 ejaculates, and total volumes ranged from 2.3 to |$75.5\, {\rm{mm}}^3\,({\rm{mean}}\, = \,25.3 \pm 2.4\, {\rm{mm}}^3)$|. The estimated total volume of ejaculate was positively correlated with its weight (Pearson’s |$r = 0.892,\,p \lt 0.01$|). Based on these calculations an average of |$6.9\, \pm \,0.68\%$| of ejaculate volume was sampled with the histology procedure (range 0.2-26%).
Using the estimated total ejaculate volume and the results from the 3D technique, it was possible to estimate the total volume of the sperm mass within an ejaculate. Estimates of total sperm mass volume within an ejaculate increased with increasing male size |$(F = 12.61,\,p \lt 0.001)$| and the model fit the data reasonably well |$(p = 0.204)$|, explaining roughly a quarter of the deviance in sperm mass volume |$(R_{{\rm{gen}}}^2 = 0.257)$| (Fig. 9A). Again, there was increased variation at larger male sizes (Fig. 9A). Estimates of total sperm mass volume ranged from 0 to |$20 \,{\rm{mm}}^3$|. The estimated number of sperm cells within an ejaculate ranged from a minimum of 1,002,181 (excluding those with 0 sperm) to a maximum of 107,364,726, and generally increased with male size (Fig. 9B).
Estimates of ejaculate composition (|$n = 48$| ejaculates): (A) the estimated total volume of sperm mass |$({\rm{mm}}^3)$| within an ejaculate in relation to male size (mm CL). The dashed line represents a GLM fit to the data (“Model fit”); model parameter estimates: |${\beta _0} = - 2.615\,(\pm1.054),\,{\beta _1} = 0.049\,(\pm0.013),\,R_{{\rm{gen}}}^2 = 0.26$|. (B) Box plots showing the estimated number of sperm cells (median, 10th, 25th, 75th, and 90th percentiles) within ejaculates produced by males in each |$5\,{\rm{mm}}$| size bin.
Discussion
Production of Ejaculates
The smallest male to successfully produce an ejaculate containing sperm in this study was |$64\,{\rm{mm\, CL}}$|, while the smallest male to produce an ejaculate (albeit without sperm) was |$61\,{\rm{mm\, CL}}$|. Krouse (1973) and Conan et al. (2001) reported that even smaller males (approximately |$41\,{\rm{mm\, CL}}$|) have sperm in their vasa deferentia, but this is the first known documentation that relatively small males have the ability to produce an ejaculate. There is likely a difference in the size of functional maturity (capable of producing an ejaculate) and the size of males actually participating in mating, which may be related to the size structure of the population and a given male’s competitive abilities. In the U.S. fishery, the smallest minimum legal size is |$83\,{\rm{mm\, CL}}$|, so there should be sufficient males larger than |$60{\text{-}}65\,{\rm{mm}}$| that would surely outcompete these smaller males for mating shelters (Stein et al., 1975; Karnofsky and Price, 1989), and thus monopolize mating opportunities (Stein et al., 1975; Atema et al., 1979; Cowan, 1992; Bushmann and Atema, 2000). However, it may be possible for small males to mate with small mature females, as there is likely at least some degree of size-assorted mating taking place.
The majority of males (79%) in the size range tested |$(60{\text{-}}108\,{\rm{mm\, CL}})$| produced either one or two ejaculates. This success rate is similar to those previously reported using the same electro-stimulation technique (77%, Kooda-Cisco and Talbot, 1983; 60-85%, Talbot et al., 1983; 100%, Aiken et al., 1984). Failure to extrude an ejaculate in this study does not necessarily indicate that the male was immature or incapable of producing an ejaculate. Rather, it is possible that some males did not have a fully formed ejaculate within the vas deferens at the time of the experiment, or had mated previously and had insufficient time to fully generate a new ejaculate. To avoid potential depletion issues resulting from recent matings, males were typically held prior to testing for at least a week (range 3 to 36 days). However, the post-mating recharge rate for the production of new ejaculate is unknown in this species. Aiken et al. (1984) described production rates of up to six ejaculates from a single male within a five-day span, and Kooda-Cisco and Talbot (1983) reported collecting additional ejaculates from 69% of electro-stimulated males tested a second time after 1-2 weeks. However, neither study reported on the composition or quality of these subsequent ejaculates. Further work is required to document the production rate of fully formed new ejaculates following mating, and how time between mating events might affect the quality of the ejaculate. Recharge rates in decapods are generally not well known or studied; reported rates range from 9 to 20 days for C. sapidus (Kendall and Wolcott, 1999; Kendall et al., 2001) to nearly a month or longer for P. brevipes (Sato et al., 2005). For P. brevipes, recovery of sperm production happens more quickly than the accessory seminal product (Sato et al., 2005), suggesting that the production of accessory products requires a greater physiological investment.
Failure of some males to produce an ejaculate in this study may also have occurred due to issues with the stimulus delivery. When an ejaculate was not immediately produced at the target stimulus strength (approximately 12 V), small adjustments in the placement of the electrodes frequently elicited an ejaculate without further adjusting the stimulus strength. It is possible that if electrodes were not in the ideal location for an individual, the stimulus may not have reached the intended musculature, resulting in no contractions and, thus, no ejaculate. In 2011, 62 males (from this and other experiments) that did not produce an ejaculate with this technique were subsequently sacrificed and their vasa deferentia were removed. Sperm enclosed within an ejaculate were found in the distal vas deferens in 79% of these males (T. Pugh, unpublished data), suggesting that the male should have been able to produce an ejaculate if properly stimulated (see Kooda-Cisco and Talbot, 1983).
Time in captivity did not affect the production of ejaculates in this study. Previous studies reported that lengthy captivity and/or prolonged stress negatively affected males’ abilities to produce ejaculates and the relative quality of those ejaculates (Talbot et al., 1983; Aiken et al., 1984). Most males in this study were tested within 3 weeks of capture, so were likely not in captivity long enough to result in reduced reproductive output. Additional males that were subjected to the electro-stimulation procedure after being used in other experiments (mating experiments) had experienced longer captivity times than those included in this study (29 to 132 days), and their ejaculate output was similarly high (72% success for 26 males, T. Pugh, unpublished data).
Ejaculate Size and Relative Composition
While there was no apparent relationship between male size and ejaculate composition (% sperm mass), central sperm mass volume did increase with male size. Thus, the larger ejaculates typically produced by larger males contained more total sperm. However, larger ejaculates were composed of a lower percentage of sperm, as shown with both the 2D and 3D analysis techniques. This decreasing ratio of sperm to acellular material as the ejaculate size increased suggests that larger males were disproportionately increasing the acellular material component of the ejaculate. Results from the 2D technique suggested a lower asymptote for sperm composition at around 10-20% sperm mass. However, comparison of results from the two techniques shows that the 2D technique tended to overestimate sperm composition, so this lower asymptote for sperm composition may be closer to 5-10% sperm mass, as shown with the 3D technique.
The 3D technique indicated that some ejaculates contained no sperm, which was not apparent with the 2D technique. For some ejaculates, portions of the acellular material may have looked like a darkened area or outline using the 2D technique, and were likely mis-identified as sperm mass, whereas the 3D technique did not detect sperm in these ejaculates. This is a weakness of the 2D technique, since interpretation of the darker areas within the ejaculate is relatively subjective. In comparison, sperm cells (or the sperm mass) are quite obvious with the 3D technique because of the histological staining, making the 3D analysis technique less subjective and more accurate for determining components of an ejaculate. Nevertheless, the good correlation in composition results generated by the two techniques shows the usefulness of the more easily performed 2D technique, as long as some caution is used in the interpretation of results. The 2D technique would be useful for relative comparisons of ejaculate composition since the observed patterns were similar with the two techniques. However, the 3D technique, while more demanding (materials, expertise and labor), likely produced more accurate measurements of ejaculate and sperm volume. Ideally a sub-sampling protocol for the 3D technique can be developed to produce accurate results while minimizing costs.
There was a positive allometric relationship between the sizes (weight) of ejaculates and males size in this study (i.e., larger males produced larger ejaculates). This increase in ejaculate size with increasing male size is consistent with a previous report for American lobster (Gosselin et al., 2003), as well as studies with several other lobster species (MacDiarmid and Butler, 1999; Butler et al., 2011; Robertson and Butler, 2013).
Male Investment in Ejaculate Composition
There was a large amount of variation in the percentage of sperm mass (0-52%) within an ejaculate based on the 3D analysis technique. Interestingly, the variation in ejaculate composition appeared to be higher for mid-sized males (approximately |$76{\text{-}}90\,{\rm{mm\, CL}}$|) in our study. The patterns observed suggest adaptive investment strategies in relation to male size. At smaller sizes, males produced ejaculates with increasing amounts of sperm, and some acellular material. At intermediate sizes, males produced more sperm, and started to increase the amount of the acellular material within the ejaculate. The ability of mid-sized males to invest in this acellular material may vary in relation to other characteristics of individual males, such as age, health (Talbot et al., 1983), or prior mating history. The acellular material may be more costly to produce than sperm, as previously suggested for some crabs (Sato et al., 2005). At larger sizes, males appeared to increasingly devote resources to production of the acellular component of the ejaculate, although, once again, health and prior history could influence this ability. This investment strategy would explain why heavier ejaculates, produced by larger males, are composed of higher percentages of the acellular material.
Both the sperm mass and the acellular material of the ejaculate have important biological roles. Obviously, sufficient sperm are required to ensure fertilization success (although sperm:egg ratios for American lobsters are currently unknown). The acellular material on the other hand serves several purposes (Kooda-Cisco and Talbot, 1982, 1983; Hinsch, 1991; Subramoniam, 1991). First, it protects sperm from sea water and bacteria while it is stored within the female’s receptacle (Hinsch, 1991; Erkan et al., 2009). Second, it may provide nutrition and possibly antimicrobial properties to allow sperm to survive lengthy storage times (Subramoniam, 1991), often as long as one year after mating for the American lobster. Finally, the acellular material of the ejaculate hardens to form a sperm plug within the female’s seminal receptacle, and likely serves to ensure the male’s paternal investment. If the ejaculate fills the female’s receptacle, the hardened plug leaves no room for additional males to inseminate the female; otherwise, the female may mate again and sperm competition would result in offspring of mixed paternity (Nelson and Hedgecock, 1977; Gosselin et al., 2005).
Based on existing knowledge of American lobster mating (see Atema and Voigt, 1995 for review), a hypothetical male investment strategy may involve an ontogenetic shift from a focus on producing sperm to investing more in acellular material (sperm plug) as a form of paternity assurance. Smaller males with low expectations of immediate reproductive opportunities with smaller, less fecund females (Estrella and Cadrin, 1995) produce ejaculates with relatively high sperm:plug ratios and thus a minimal investment in the acellular material. As males get larger, their expected reproductive opportunities increase, the females they tend to mate with are likely larger, and they start to invest more in the acellular material, making larger ejaculates with relatively more “sperm plug” component that now incorporates paternity assurance. Finally, at larger sizes, males invest in increasing the total size of the ejaculate (thus increasing sperm volume and total ejaculate volume), to ensure they not only can fertilize the entire clutch of increasingly fecund larger females, but also completely fill the female’s receptacle to protect their paternal investment. The increased variation observed in ejaculate sizes at larger male sizes may reflect variability in male fitness, suggesting that not all males are equal in terms of reproductive investment.
Larger males provided females with larger ejaculates in laboratory mating experiments (Gosselin et al., 2003), corroborating our results. Additionally, Gosselin et al. (2003) reported that larger males performed longer duration post-copulatory mate-guarding behavior than did smaller males, which also suggests increased investment in paternity assurance by larger males. Due to the lengthy time period between mating and spawning in this species, however, an additional form of paternity assurance is required after the mate-guarding period ends, since females have been shown to mate again in the hard-shelled condition (“intermolt mating” Dunham and Skinner-Jacobs, 1978; Waddy and Aiken, 1990). The decreasing sperm:plug ratio observed here suggests that American lobster males solve the problem of long-term paternal assurance by increasing their investment in a sperm plug. Sainte-Marie and Lovrich (1994) reported male snow crabs differed in the sperm:seminal fluid ratios in ejaculates based on their morphometric maturity, and decreased sperm:fluid ratios were later suggested as a mechanism for paternal assurance in older morphometrically mature males (Sainte-Marie et al., 2008).
Male body size is often not the only factor influencing ejaculate size, as males in some species reportedly scaled the size of their ejaculate to the size of their partner: P. argus (MacDiarmid and Butler, 1999); H. americanus (Gosselin et al., 2003); Austropotamobius italicus (Faxon, 1914) (Rubolini et al., 2006); in some cases they may adjust ejaculate sizes based on perceived mating opportunities (Pitnick and Markow, 1994; Rondeau and Sainte-Marie, 2001). In these studies, male ejaculate size was measured after mating took place and their ejaculate was subsequently removed from the female. While these techniques provide good evidence of male reproductive contributions to mates, it is difficult to separate external effects on male reproductive capacity. A major advantage of the electro-stimulation technique is that the ejaculate is produced without the need for, or influence of, intraspecific interactions. Thus, our results represent male potential capabilities in the absence of female cues. However, males in our study were held communally with other males, so the potential for a ‘rival males’ effect cannot be ruled out, and should be controlled in future experiments. An interesting next step would be to test males using this technique after exposure to the presence of receptive (premolt) females of different sizes.
Estimation of Sperm Volume
The estimates of total sperm mass volume based on calculations of ejaculate total volume and composition (% sperm mass, 3D technique) suggest that ejaculates may have no sperm or contain upwards of |$20\, {\rm{mm}}^3$| of sperm mass. These are the first quantitative estimates of ejaculate composition for this species. The obvious next question is how many sperm cells are contained within a unit of sperm mass volume. American lobster sperm cells readily undergo an acrosomal reaction; so readily that even slight mechanical agitation will elicit the reaction (Talbot and Helluy, 1995). Unfortunately, this sensitivity makes traditional methods of homogenizing the ejaculate and counting sperm cells unreliable, because once the sperm have undergone the acrosomal reaction they have essentially been turned inside-out and degenerate to the point that they are unrecognizable under a compound microscope (personal observation). While histological sections containing the sperm mass are too thick to allow for accurate counts of individual sperm cells, which are randomly oriented (Kooda-Cisco and Talbot, 1982) and in more than one layer within the section (K. Benhalima, personal observation), it was possible to measure the diameter and length of individual sperm cells and then calculate their volumes. This value was then used to estimate the number of sperm within ejaculates produced by males in this study, which ranged over two orders of magnitude, from a minimum of 1,002,180 to a maximum of 107,364,726 sperm. These estimates were generated assuming that cylindrically-shaped sperm cells are packaged with maximum economy within the sperm mass, and do not account for the spikes present on lobster sperm cells. While these are the first reported estimates for the number of sperm within American lobster ejaculates, they are preliminary and should be confirmed by some method of direct counting, possibly using a DNA-targeting stain to help identify individual sperm cells within a histological section.
In conclusion, this is the first study to quantify the production and composition of ejaculates relative to male size in the American lobster H. americanus. Males as small as |$64\,{\rm{mm\, CL}}$| successfully produced an ejaculate containing sperm, indicating that small males might be able to mate if they were behaviorally competent and could gain mating opportunities. Decreasing ratios of sperm:plug with increasing ejaculate weight indicates that larger males invest disproportionately more in the acellular component of the ejaculate (the sperm plug). This “investment” is likely a form of paternity assurance over the long interval between mating and spawning. The large degree of variation observed in ejaculate composition makes it difficult to assess male reproductive contributions based solely on ejaculate weight, and indicates that not all males are capable of investing equally in reproduction.
Acknowledgements
We would like to thank Audra Chaput, Elizabeth Morrissey, Jill Carr, Kelly Whitmore, and Bonnie Pugh for assistance with the electro-stimulation testing. Nate Rennels provided lab support and maintained the sea water systems at the UNH Coastal Marine Lab. Numerous MADMF staff members assisted with tanks and pump maintenance at the UMass Marine Station. Josh Carloni, Jason Goldstein, Audra Chaput, Elizabeth Morrissey and Tom Langley assisted with lobster collection and maintenance. MADMF Assistant Director Michael Armstrong and the staff of the MADMF Age & Growth Laboratory generously provided microscope time and advice. Gary Nelson and Robert Glenn provided statistical advice, and Gary Nelson assisted with R programming. Michael Armstrong, Robert Glenn, Genevieve Nesslage, Hunt Howell, and Kari Lavalli provided valuable critiques of an earlier version of the manuscript. Funding for this work was provided by New Hampshire Sea Grant (No. NA06OAR4170109) to WHW. This is contribution No. 52 from the Massachusetts Division of Marine Fisheries.
