Date as continuous variable?

We are comparing pesticide amounts in three different species of bees. We collected equal numbers of each bee species on three different dates throughout the season and are testing to see if pesticide amounts vary by bee species. JMP gives me the option to set each variable as continuous, ordinal, or nominal. Should date be set as a continuous variable if we only have 3 discrete collection dates?
The dates themselves are not necessarily important or the focus of the study, but I included them in the model because pesticide load might vary across our three sampling dates. Setting the date as either continuous or nominal changes the results I get from the JMP output.


TS Contributor
I'm not sure of the distinction here because ordinal and continuous measures aren't necessarily different. Dates might be interval scale, but either way, accounting for the date is reasonable for the situation you metnioned in which variablility may exist from different production runs on the pesticide (or other things related to time).