Improved Analysis Of Shell-Disease Severity In Offshore Populations Of American Lobster     Richard A. Robohm, Diane Kapareiko, John J. Ziskowski, and George R. Sennefelder   NOAA, NMFS, Northeast Fisheries Science Center, Milford, CT     In previous work we established the prevalence of shell-disease and measured the lesion areas as well as   carapace lengths in populations of American lobster inhabiting offshore canyons of the northeast coast of the United States.  Subsequently, we found a mathematical relationship (y = 1.1034 + 1.9677 * log CL) between carapace length (CL) and surface area (y) for lobsters within the size-range of our database (manuscript submitted).  The formula allowed us to determine the percentage of total surface area affected by shell disease on each lobster.  This percentage, multiplied by 103 , constitutes a “disease severity index”(DSI) that will be useful in evaluating intensity of disease for various lobster size-classes in continental-shelf locations.  Regression-tree analysis (binary recursive partitioning) of the DSIs indicated that, unlike earlier prevalence data showing significant sex differences, the most important variable was carapace length.  Kruskal-Wallis analyses showed that DSIs for smaller females (< 8.5-cm CL) were significantly higher than for larger females (p = 0.05).  Smaller males (< 9.5-cm CL) had significantly higher DSIs than larger males (p = 0.025).  No differences were found between small males and small females (p = 0.094) or between large males and large females (p = 0.798).  Additional regression-tree analyses suggested both seasonal differences and site differences.  Complete non-parametric analyses of these variables should clarify any relationships between shell-disease in lobsters and proximity to anthropogenic pollution such as the 106-mile sewage-sludge dumpsite.