Hello, I have a dataframe with e.g. three different curves (columns), which I want to have averaged. I use the row.Means function, which is working good. However as I have a time series, some parts of the three columns have sometimes NA. I want however that if only one or two columns are available, the level of the averaged column value does not suddenly rise or fall, because one column in missing. Is there maybe already a syntax available for it? I would like to calculate an average of all curves, but in those cases where I have only two or less columns, just use those but still try to get a harmonized curve. Here are my three columns. The fourth one is the averaged one. I want that row 7534 continues with 1086.77... and that the difference, due to the reduction of "N" is calculated and removed automatically. Row number 7532 900 1479.00 877.0000 1085.3333333 7533 901 1482.00 877.3375 1086.7791667 7534 903 1482.00 NA 1192.5000000 7535 904 1486.00 NA 1195.0000000 7536 905 1487.00 NA 1196.0000000 Same if third columns is coming back: Row number 8832 1471 2172.00 NA 1821.5000000 8833 1463 2173.00 1242.0000 1626.0000000 Any ideas or help is really appreciated. Regards Loenn -- View this message in context: http://r.789695.n4.nabble.com/Mean-of-three-inconsistent-curves-tp3560788p3560788.html Sent from the R help mailing list archive at Nabble.com.