r - custom functions with dplyr summarise -

i've got data frame looks this:

row     year    rainfall area species density rainfall1 1   46  1993    433.70  br  red 2.9300000   low 2   47  1994    365.65  br  red 8.0000000   low 3   48  1996    545.80  br  red 5.8558559   high 4   49  1999    785.40  br  red 17.0158617  high 5   50  2000    736.30  br  red 8.8778409   high 6   51  2001    370.40  br  red 6.9874901   low 7   52  2002    174.80  br  red 2.0579308   low 8   53  2003    290.50  br  red 7.6328655   low 9   54  2004    424.40  br  red 7.4234908   low 10  55  2005    336.30  br  red 0.7580045   low 11  56  2007    524.40  br  red 0.4500000   high 

this repeats resulting 4 areas , 2 species giving 120 results.

i'd add new column relative density year in area/species (as percentage). i've wrote small function relative density:

relative <- function(x) (x/sum(x)) * 100 

i'm not sure how right set of data work out group_by , summarise functions. need able retrieve densities given year,species,area , rainfall category

any help?

in base r that:

df$sumval <- tapply(df$density, list(df$year, df$area, df$species), sum ) df$perce <- df$density / df$sumval 

but said in comment has 1 result because every year has 1 value.