Anthropogenic disturbance causes habitat destruction and thereby presents a serious threat to many forest species. Seasonal variation in diversity and community composition of carabid beetles was investigated in the Uzungwa Scarp Nature Forest Reserve (USNFR). Carabid beetles were collected using three methods, namely pitfall traps, active day search and active night search. Sampling of ground beetles was carried out in three seasons: end of dry season, end of wet season and wet season. Results showed that the diversity of ground beetles was high during the wet season (Shannon-Wiener H' = 2.19) and low at the end of the dry season (H' = 1.83). The difference was statistically significant when comparing the end of dry season against the wet season, likewise wet against end of wet season (t = 5.85 and 4.88, respectively, p < 0.05). Species richness was high during the wet season and the difference was statistically significant among the three seasons (χ² = 13.46, p < 0.05). Kruskal-Wallis test revealed that the difference in abundance among the three seasons was statistically significant (H = 128.3, p < 0.05). In all the seasons, Galeritiola procera was the most abundant species and it was overrepresented during the wet season. Bray-Curtis similarity index showed low similarity in community composition among all three seasons. We conclude that communities of carabid beetles in disturbed areas vary with season, and therefore, to have a comprehensive sampling of carabid beetles, sampling should cover at least one-year involving different months in all seasons of the year.

人为干扰导致栖息地破坏，从而对许多森林物种构成严重威胁。在 Uzungwa Scarp 自然森林保护区 (USNFR) 中研究了步甲甲虫多样性和群落组成的季节变化。使用三种方法收集步甲虫，即陷阱、白天主动搜索和夜间主动搜索。地面甲虫的采样分三个季节进行：旱季结束、雨季结束和雨季。结果表明，地面甲虫的多样性在雨季较高（Shannon-Wiener H' = 2.19），在旱季末较低（H' = 1.83）。当比较旱季结束与雨季结束时，同样比较湿季结束与雨季结束时，差异具有统计学显着性（分别为 t = 5.85 和 4.88，p < 0.05）。雨季物种丰富度较高，三个季节间差异有统计学意义（χ ² = 13.46，p < 0.05）。Kruskal-Wallis 检验显示，三个季节之间的丰度差异具有统计学意义（H = 128.3，p < 0.05）。在所有季节中，Galeritiola procera都是最丰富的物种，并且在雨季出现过多。Bray-Curtis相似指数显示三个季节的群落组成相似性较低。我们的结论是，受干扰地区步甲虫群落随季节变化，因此，为了对步甲虫进行全面采样，采样应至少覆盖一年，涉及一年中各个季节的不同月份。