Discussion

Community structure
The bimodal size-distribution of lobsters observed in Penobscot Bay is consistent with the pattern seen at Lowell's Cove, a long-term intertidal study site in Casco Bay, Maine (Cowan 1999a). The size class of young-of-the-year (4-15 mm CL) is clearly distinguishable from older year classes.

The male-biased sex ratio observed in Pen Bay has also been found consistently in the intertidal zone. Similar ratios have been observed at Lowell's Cove, 8 other sites in Harpswell, and 1 site in New Hampshire (Table 4). The reasons for this biased sex ratio are not known. In lobsters, sex is not determined until several months after settlement. It is possible that environmental factors such as temperature or salinity in the intertidal zone lead to the development of more males than females. Alternatively, the male-biased sex ratio may reflect differential rates of survival in, or tolerance to, this environment. It would be interesting to compare sex ratio between intertidal and subtidal nursery grounds as these data become available.

The incidence of injury was similar to patterns seen at other intertidal sites, in that the majority of lobsters had both claws. At 8 sites in Harpswell, 24% of lobsters had lost at least one claw (20% loss of one claw, 4% loss of both claws; n = 358) while in New Hampshire the percentage of claw loss was higher (25% loss of one claw, 10% loss of both claws; n = 214) (Ellis and Cowan 1999a, 1999b). Lobsters will drop a claw as a protective measure when attacked by a predator. It would be interesting to compare the incidence of injury in subtidal versus intertidal areas, to assess the relative protection from predators in each habitat, however, to the best of our knowledge, these data have not been collected in subtidal studies.

The peak molting period of juveniles in the intertidal zone occurred in August. This coincides with the peak molting season of adults in Maine, when landings are high and the greatest numbers of soft shell lobsters are caught. Growth rates of lobsters increase proportionally with increased temperature (Waddy et al. 1995). Thus it is not surprising that peak molting occurred in August, since temperatures generally increase as summer progresses.

The timing of settlement appears to extend beyond September and into the fall. Settlers were found at two intertidal sites in May, while others were found in September and October (Figure 11). Newly-settled lobsters (< 6 mm CL) have been observed in the intertidal zone of Lowell's Cove in all months of the year except May and June (Cowan unpublished). Young-of-the year < 10 mm CL have been observed at Lowell's Cove all 12 months of the year.

Habitat Use
Rocks that provided shelter to juvenile lobsters in Penobscot Bay (Table 2) had similar dimensions to those sampled at the long-term study site at Lowell's Cove, where average rock dimensions were approximately 30 x 30 x 10 cm (Cowan 1999a). In terms of horizontal dimensions, rocks in Pen Bay were rectangular (40 x 29 cm) rather than square as in Lowell's Cove. In both cases, however, rocks were generally flattened horizontally, i.e., height was less than 1/2 the length or width. Cobble, the preferred subtidal habitat of juvenile lobsters (Wahle and Steneck 1991) is more spherical or cube-shaped. In the intertidal zone somewhat flattened rocks would be less likely to be overturned by wave action than would cube-shaped or round rocks.

An alternative lobster habitat was documented on Vinalhaven, in which lobsters were found to burrow in eel grass. Lobsters have previously been observed in eel grass (e.g., Hudon 1987; Heck et al. 1989; Wahle and Steneck 1991). On Vinalhaven, no young-of-the-year were found in eel grass and the size distribution was biased toward larger animals than in the nearby rocky intertidal zone (Figure 9). This is consistent with subtidal studies in which eel grass beds sheltered lobsters of all sizes, but supported very low densities of lobsters <40 mm CL (Wahle and Steneck 1991).

Patterns of Distribution and Abundance
No juvenile lobsters were found at any of the study sites in inner regions of Penobscot Bay (Figure 2). This same pattern was observed in a 1995 survey of 15 intertidal sites in Casco Bay, Maine; juvenile lobsters were found at sites exposed to open waters, but not along the shoreline of inner bays and sounds (Cowan 1999a).

This distribution pattern suggests that biotic and abiotic factors may influence the suitability of locations as sites for benthic recruitment (Cowan 1999a). Potentially important physical factors include salinity, temperature, substrate type, wave action, ice scouring, and prevailing wind and water currents. Physical characteristics of settling sites exhibiting good intertidal habitat characteristics further upstream (i.e., inner bays) may be unsuitable as lobster nurseries due to extreme ice scouring, propensity of fresh-water runoff, and the amount of cover available in adjacent subtidal areas (mud bottom). Biological factors influencing suitability of settlement habitat may include proximity to postlarval supply, shelter availability as refuge from predators, and food availability.

If the contrast between inner and outer bays is influenced mainly by postlarval supply, it is possible that either currents carrying postlarvae do not reach inner regions of bays, or that postlarvae settle out of the water column in outer regions before currents reach the inner regions. Work being conducted by Dr. Lew Incze as part of the Penobscot Bay Lobster Collaborative may help to distinguish these competing hypotheses.

In our study, settlers were confined to a band across the mouth of Penobscot Bay (Figure 10). At extreme outer islands (Monhegan and Matinicus islands), no settlers were observed and juvenile densities were very low or zero. There are no reasons to doubt the accuracy of densities observed on Matinicus. In fact, subtidal studies in 1998 showed low settlement at Matinicus, and fairly low densities of early benthic phase lobsters (5-40 mm CL) (Steneck and Wilson in prep.). In the case of Monhegan Island, which showed high subtidal densities of settling and juvenile lobsters in 1998 (Steneck and Wilson in prep.), the lack of appropriate intertidal habitat may have negatively biased our observed intertidal densities. The eastern side of the island is sheer cliffs, while the western, southern, and northern portions do not contain coves with rocks small enough for lobster monitors to overturn. The test monitoring site had some such rocks, but not for a full 20 meters, and thus was not truly appropriate for intertidal monitoring.

Because of this problem at Monhegan, we are limited to using Matinicus as example of an extreme outer island for comparison with sites at the mouth of the bay. The lack of settlers and the comparatively low densities of juveniles at Matinicus suggest a lower postlarval supply on Matinicus than across the mouth of the bay. This may be related to the inward turning of the EMCC in late summer and fall, which would turn it away from Matinicus. Alternatively, it may reflect a behavioral response of postlarvae which may preferentially settle in shallow, warm waters (Steneck and Wilson in prep.). Cues indicating shallow, warm water would be stronger and more abundant nearer the mainland than out to sea.

Along the mouth of the bay, intertidal densities of juvenile lobsters were higher in western regions than eastern regions. In western regions, densities decreased from south to north (i.e., Allen Island > Port Clyde > So. Thomaston). Similar patterns have been observed in the subtidal zone (Steneck and Wilson in prep.). These patterns of juvenile abundance support the hypothesis that regions with greater and earlier exposure to onshore currents will receive greater numbers of settling postlarvae.

Conclusions
The Lobster Conservancy's volunteers have proven to be reliable data collectors in Penobscot Bay. The data they gathered in 1998 and 1999 as part of the Penobscot Bay Lobster Collaborative have been used to determine community structure, habitat use, and timing of settlement and molting in the intertidal zone. Data from Pen Bay are remarkably consistent with data gathered by TLC's volunteers elsewhere in New England, and with long-term data collected by TLC's trained biologist.

The Pen Bay intertidal data have also been used to document patterns of abundance and distribution of juvenile lobsters, including young-of-the year. The observed patterns of settlement and juvenile abundance of lobsters in the intertidal zone support hypotheses that onshore currents, such as the Eastern Maine Coastal Current, deliver proportionally more postlarvae to western, outer regions of the bay. Extreme outer islands appear to receive fewer postlarvae.

In Penobscot Bay, the Intertidal Lobster Monitoring Program has proven to be not only an important form of educational outreach through its involvement of community volunteers, but also a useful scientific research program. The long-term goals of the Pen Bay Lobster Collaborative to use indices of settlement and juvenile abundance to make predictions about lobsters landings will be possible only with long-term monitoring. It is important that the Pen Bay Intertidal Lobster Monitoring Program be continued in upcoming years.

Acknowledgments
This project was supported by the Island Institute. Maine/New Hampshire Sea Grant provided funding to purchase monitoring equipment, and the Maine Community Foundation Community Fisheries Project supported production of a volunteer training handbook. We especially thank the many volunteers who dedicated their time and effort to collecting data.