Playing around with some data on protoshell length taken by Ryann (PSRF intern) from the salinity experiment.

In R:

 > plot(Length~Population, data = sal30df)

Boxplot of protoshell length in 7 day old Olympia oyster larvae from 3 populations in Puget Sound. n = 20 per population

Did an anova:

> aov.sal30 <- aov(Length~Population, data=sal30df)
> summary(aov.sal30)

Df Sum Sq Mean Sq F value Pr(>F)
Population 2 9281 4641 14.81 6.61e-06 ***
Residuals 57 17861 313

Signif. codes:  0 ‘***’ 0.001 ‘**’ 0.01 ‘*’ 0.05 ‘.’ 0.1 ‘ ’ 1

Indicates that the means of protoshell length are significantly different (yay!).

> pop <- sal30$Population
> len <- sal30$Length
> ptest <- pairwise.t.test(len, pop, p.adjust = "bonferroni")

Pairwise comparisons using t tests with pooled SD

data: len and pop

Fidalgo Bay Hood Canal
Hood Canal 0.0913             –
Oyster Bay 0.0069          3.9e-06

P value adjustment method: bonferroni

Very significant p-value between Hood Canal and Oyster Bay.