the gastropod on individual hermit crabs appears to be food-related. Field surveys, however, suggested that the hermit crab population is limited by shell number ...
Oecologia 9 Springer-Verlag1986
Oecologia (Berlin) (1986) 69:213-216
Positive abundance and negative distribution effects of a gastropod on an intertidal hermit crab Peter T. Raimondi i and Curtis M. Lively 2 .
1 Department of Biological Sciences, University of California, Santa Barbara, CA 93106, USA 2 Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, AZ 85721, USA
Summary. Field experiments were used to determine the
effect of a common intertidal snail (Nerita funiculata) on the use of space for foraging by the hermit crab Clibanarius digueti. Removals of Nerita resulted in an increased density of foraging Clibanarius, while additions of the gastropod had the opposite effect. The observed negative effect of the gastropod on individual hermit crabs appears to be food-related. Field surveys, however, suggested that the hermit crab population is limited by shell number, rather than food. Because Nerita contributes to the shell resource, its effect on the hermit crab population is positive. Nerita, therefore, has a negative effect on the distribution of foraging hermit crabs, but a positive effect on their abundance. Such decouplings of distribution and abundance effects are rare.
Factors that affect the distribution of a species usually affect its abundance in the same way (see Brown 1984). Lower-intertidal barnacle species, for example, have been shown to restrict the distribution of barnacle species which live higher on the shore (e.g. Connell 1961, Dungan 1985); and as a consequence of this, they reduce the realized density of the upper-shore species. In the present study, we present evidence of an exception to the general positive correlation between distribution and abundance effects. A gastropod is shown to negatively affect the local distribution of an intertidal hermit crab, but positively affect its abundance. Hermit crabs have been the subjects of many ecological and behavioral studies, most of which have focused on their utilization and exchange of coiled gastropod shells (reviews by Reese 1969; Hazlett 1981). Population size is most commonly limited by the availability of shells (e.g. Hazlett 1970; Vance 1972a; Childress 1972; Fotheringham 1976; Bach et al. 1976), which are used by most species as transportable shelters that afford protection from unspecialized predation (Vance 1972b; Bertness 1981) as well as desiccation (Bollay 1964; Bertness 1981) and wave action (Reese 1969). Because of the obvious importance of coiled shells to hermit crabs, the gastropods that produce them may be regarded as having a positive effect on hermit crab populations. Offprint requests to: P.T. Raimondi * Present address. Department of Zoology, University of Canterbury, Christchurch 1, New Zealand
The hermit crab Clibanarius digueti (henceforth Clibanarius) is endemic to the Gulf of California and is the most common and active hermit crab on rocky intertidal shores in this region (Snyder-Conn 1980). The herbivorous gastropod, Nerita funiculata (henceforth Nerita), overlaps with Clibanarius in the midintertidal zone and is especially common in boulder fields and on rocky outcrops protected from direct wave action (Raimondi and Lively, pers observ). At our study site, the vertical range for Clibanarius was - 0 . 6 to + 3 m above mean low water (MLW), while the range for Nerita was + 0.6 to + 4.3 m above MLW. Nerita, therefore, occur over 0.67% of Clibanarius' range. Because Clibanarius occupy Nerita shells, Nerita is assumed to have a positive effect on the Clibanarius population. The primary purpose of the present study was to test the hypothesis that this gastropod also has a negative effect on the hermit crab. Materials and methods
Study site. Field experiments were conducted at Punta Pelicano (Pelican Point), a granitic shore in the northern Gulf of California, 8 km northwest of Puerto Pefiasco in Sonora, Mexico. The northern Gulf of California is known for its strong seasonality (Hendrickson 1973), extreme tidal amplitude (Brusca 1980; Levington 1982), high diversity and endemism (Brusca 1980). Three adjacent and interconnected tidepools were selected in the mid-intertidal at Punta Pelicano and designated a s " landward", "middle" a n d " seaward". The water level of the pools (when exposed by low tides) was 60 cm above MLW. All three pools had smooth granite walls (approx 60 ~ slope) on their north sides, extending at least 75 cm above and roughly 30 cm below water level. Several species of branching, red coralline algae were the primary sessile occupants of the granite walls in the tidepools, and Clibanarius was the most abundant mobile species in the pools. Nerita was the most abundant mobile species above pool level and foraged on a thin mat of blue-green bacteria (Callothrix spp). The pools were protected from direct wave action by a granitic outcrop on the south (seaward) side. Removal experiment. In order to determine the effect of Nerita on Clibanarius, we manually removed Nerita from one of the tidepools for two separate, one-month periods during the summer of 1982. During the first removal period (27 May-24 June), Nerita were removed from the middle
214 pool and the landward pool served as a control. During the second removal period (31 July-22 Aug), Nerita were removed from the landward pool and the middle pool served a control. The removal and control sites were switched between periods in order to decouple site effects from treatment effects (see Hurlbert 1984). We also removed Nerita from the third (seaward) pool for the first eight days of the second period and then allowed the gastropod to return to that pool. For both removal periods, Nerita were removed every day during spring low tides. The numbers of Nerita and Clibanarius in the sites were estimated at the end of the first removal period by counting the numbers of both species in each of four (nonoverlapping) 730 cm 2 areas. They were sampled in the same way during the second removal period except that only three samples were taken in each site and the sites were sampled daily during spring tidal series throughout the duration of the experiment. Nerita were sampled just prior to inundation of the pools by incoming evening tides, and Clibanarius were sampled when the pools had been inundated to a depth of 30 cm. This was because Nerita begin foraging when exposed by evening low tides, while Clibanarius delay foraging until they are inundated (Raimondi and Lively, pets observ). Addition experiment. To further determine the effect of Nerita on Clibanarius, eight granite boulders of approximately equal sizes (c 20 x 20 x 10 cm) were collected from the midintertidal at Punta Pelicano and randomly distributed between two plastic pools (1 m diameter in 12 cm of seawater) so that there were 4 boulders in each pool. Nerita (n = 200) were added to one of the pools (randomly determined), and the other pool served as a control. After 2 h the Nerita were removed and two boulders in each pool were randomly selected and transplanted to the other pool. Both pools, therefore, contained two treatment (Nerita added) and two control boulders. Hermit crabs (n= 100) were then added to each pool and the number foraging on each of the eight boulders was determined after I h. Interference versus exploitation. The purpose of the previous experiments was to determine whether Nerita have any effect on Clibanarius foraging under natural conditions. This effect might be due to either exploitation of a common resource or interference with the hermit crab. To determine whether Nerita inhibit foraging by Clibanarius, through either chemical (allelopathy) or aggressive means, we placed 50 Nerita in 5 cm of water in each of three basins (40 cm diameter) and restricted them to a randomly chosen half (with respect to compass direction) using a partition. After 2 h, the partitions were removed and 30 Clibanarius were placed (haphazardly) in the basins. Nerita were manually restricted to their original sides of the basins and Clibanarius were counted on both sides of the basins at 5 min intervals (beginning 15 min and ending 45 min after the addition of the hermit crabs). If Nerita do not interfere with foraging by the hermit crab they may still affect them by removing food. To determine the response of Clibanarius to material left by receding tides, we attached five pieces of carpet (10 by 20 cm) in the intertidal zone at Punta Pelicano and collected them when they became exposed by the following low tide. As a control, we placed five additional pieces of carpet in filtered seawater while the experimental pieces were inun-
dated. One control and one experimental carpet were then placed 5 cm apart in each of five different basins (40 cm diameter) in 5 cm of seawater. Fifty Clibanarius were placed in the gap between the carpets. The number of Clibanarius on control and experimental carpets was determined at 15 min intervals (beginning 15 rain and ending 90 rain after the addition of the hermit crabs). Shell surveys. Surveys of shell use by Clibanarius were made at two different sites: one (site B) near the site of the removal experiment and another (site A) 200 m east of that site. At site B, we collected all dead gastropod shells (n = 344) found during a single low tide, examined them for the presence of Clibanarius, and then sorted them according to species. At site A we also measured the aperture width of each of the shells collected (n= 646). This measurement was taken to determine the size distribution of each of the shell species commonly used by Clibanarius. Results
Removal experiment. The manual removals of Nerita indicated a negative effect of Nerita on the utilization of space by foraging Clibanarius (Table 1). More Clibanarius were observed in removal sites than in control sites at the ends of both removal periods. The difference between the mean number of Clibanarius in removal and control sites was statistically significant (ANOVA: F1, 2 = 74.07; P < 0.025). The analysis was performed on the means in Table 1 to avoid the problem of having replicate samples of the same experimental unit or "pseudoreplication" (see Hurlbert 1984). The results of daily samples during the second removal period (Fig. 1) give a clearer picture of the response of Clibanarius to the Nerita removals. The hermit crab increased in both pools where Nerita were removed during the first spring tides of August 1982, but no such increase was observed in the control pool (which was nested between the two removal pools and in which Clibanarius had increased during the first removal period). Clibanarius densities then decreased in the seaward pool during the second spring tidal series when Nerita were allowed to reinvade, but continued to increase slightly in the landward pool were Nerita removals were reinitiated following the neap tidal series (Fig. 1). As previously, no change was observed in Nerita densities in the control pool. Finally, hermit crab density declined in the landward pool (as previously observed in the seaward pool) following termination of Nerita removals following the second spring tidal series (Fig. 1). Hence the removal of Nerita resulted in local increases in hermit crab Table 1. The mean number (_ SE) of Clibanarius per ring sample
(730 crn2) in Nerita removal and control sites after approximately one month Nerita
Middle pool Landward pool
removal
control
9.25 -t-3.35 (period 1) 11.00 _ 3.46 (period 2)
1.33 _+0.88 (period 2) 0.00 +_0.00 (period 1)
215
iiI 66
Interference Experiment
o s
Table 2. Results of the "interference" and "carpet" experiments. The means are the average differences between the number of hermit crabs on the experimental and control sides of the basins
CONTROL
.
.
.
.
.
Removal 133 t < - ~ - ~ l 66
.
.
Time (m/n)
~'~--,',
REMOVAL / INVASION
;~
/II
.....
Remora[ 1(
mean SE paired t
--2 (4) 0.50
20
25
30
35
40
45
2.7 - 0 . 7 --0.7 (3.3) (2.4) (1.8) 0.82 0.29 0.39
8 (4.6) 1.79
8 (6) 1.33
2 (4) 0.50
30
75
90
Carpet Experiment
- -
Removol
133 66 1
15
R E MMO V A L / R E M O V A L
Time (m/n)
15
mean SE pairedt
10.2 9.4 9.2 13.2 12.2 8.4 (1.9) (2.4) (3.4) (3.9) (2.7) (2.8) 5.37* 3.92* 2.71" 3.38* 4.52* 3.0*
45
60
tll
* probability
