Chronology of the Invasion of the Tennessee and Cumberland River Systems by the Mississippi Silverside, Menidia audens, with Analysis of the Subsequent Decline of the Brook Silverside, Labidesthes sicculus
The Mississippi Silverside, Menidia audens, was first encountered in Kentucky Reservoir, the most downstream reservoir on the mainstem Tennessee River, in 1991. Over the next 19 years, it invaded all nine mainstem reservoirs, consisting of 1049 river km, and two tributary reservoirs, at an average rate of 55 km per year. This species was first encountered in Barkley Reservoir, the lowermost reservoir of the Cumberland River, in 1990, and has since established populations throughout at least 402 river km of the mainstem Cumberland River. Following the first record of M. audens in an unimpounded section of a Tennessee River tributary during 1998, 1048 sites were sampled throughout the Tennessee River system and this species was documented in only eight additional unimpounded tributaries, indicating aversion to lotic systems. Following the invasion of Tennessee River reservoirs by M. audens, mean abundance of the Brook Silverside, Labidesthes sicculus, linearly decreased with time. Electrofishing catch per unit effort (CPUE) data indicate that 14 years after sympatry with M. audens, L. sicculus were undetectable or occurred at extremely low densities in reservoirs. Reservoir populations of L. sicculus exhibited cyclic variation in abundance among years prior to the presence of M. audens. The observed range of electrofishing CPUE of L. sicculus before invasion by M. audens (n = 3865) was significantly different (P < 0.001) than the observed range following the presence of M. audens (n = 3975). After detection of M. audens, electrofishing CPUE of L. sicculus decreased and the range of variation from year to year diminished. Data from Cumberland River reservoirs showed similar patterns of displacement.
Tennessee and Cumberland River reservoirs, lock and dam locations, major tributaries and dated locations of first detection of M. audens.
Locations of collections of M. audens in unimpounded tributaries to the Tennessee River during TVA stream sampling. Black circles represent detected occurrences of M. audens; gray diamonds represent all TVA tributary sites sampled since 1998 (1048 sites); black diamonds represent records of L. sicculus at these sites (204 sites); and open squares represent areas where L. sicculus were collected prior to 1998, but were not present in two or more subsequent samples at a particular site (25 sites).
Estimated number of L. sicculus per hectare in Wheeler and Chickamauga reservoirs from cove rotenone samples collected prior to the presence of M. audens.
Range of electrofishing CPUE of L. sicculus within Tennessee River reservoirs, 18 years before occurrence of M. audens (3865 electrofishing runs; ∼1160 shoreline km sampled) and 19 years following presence of M. audens (3975 electrofishing runs; ∼1193 shoreline km sampled). Negative numbers on the x-axis represent years before occurrence of M. audens; positive numbers represent years following the presence of M. audens.
Log10 of the mean electrofishing CPUE of L. sicculus occurring in Tennessee River reservoirs, displayed against time (years) after initial documentation of M. audens (3750 electrofishing runs; ∼1125 shoreline km sampled).
Mean electrofishing CPUE + SE of M. audens within Tennessee River reservoirs, 16 years after detection (3450 electrofishing runs; ∼1035 shoreline km sampled).
Contributor Notes
Associate Editor: D. Buth.