W artykule zbadano wieloletni i roczny przebieg najwyższych miesięcznych sum opadów atmosferycznych w Polskich Karpatach Zachodnich na podstawie danych z 6 stacji: Nowy Sącz, Sanok, Bielsko-Biała, Maków Podhalański, Krynica i zakopane z lat 1951-2005 oraz wyznaczono ich wartości prawdopodobne. Wykazano, że najwyższe miesięczne sumy opadów mogą występować w ciągu całego roku, z największą częstością w miesiącach letnich. Nie stwierdzono istotnych statystycznie trendów zmian wieloletnich, a jedynie krótkookresowe fluktuacje.

The authors have investigated long-term and annual patterns in maximum monthly precipitation totals in the Western Carpathian Mts. in Poland. Data from six stations (Nowy Sącz, Sanok, Bielsko-Biała, Maków Podhalański, Krynica and Zakopane) spanning the period 1951-2005 were used both in their raw form and as derived probabilities. The study showed that the highest monthly precipitation totals were likely to occur throughout the year, but most frequent were in the summer months. No long-term statistically significant trends were identified, just shorter-term fluctuations. All of the stations recorded high precipitation totals in last years of the 20^t^h c. and in 2001 too, while lower totals were recorded in the mid-1990s. During the first five years of the 21^s^t c., the trend to a decrease was found at most of the stations. At some of them peak monthly totals exceeded the monthly average figures by more than 350%. For precipitation totals with a 1% probability of exceedance, the highest totals were obtained from Gumbel distribution and the lowest from the normal distribution at all of the stations. In high probability ranges (e.g. \mathit{p} = 20%) the highest precipitation was derived from Weibull distribution. The highest maximum monthly totals recorded during the period 1995-2005 corresponded approximately to a 1% probability of exceedance. The only exception was the station at Maków Podhalański, where the highest monthly precipitation of 521.1 mm was at 88.1 mm higher than the figure derived from the theoretical distribution with \mathit{p} = 1%.