Data Generation - Classification , why do we use a list of means?

In the Data Generation - Classification video at 3:55 a list containing two means is created i am not able to understand why do we need 2 means for creating 1 curve. I am Just confused about this point ?