deletions | additions       diff --git a/discussion.tex b/discussion.tex  index 2e8d01a..3108d0d 100644  --- a/discussion.tex  +++ b/discussion.tex 
...
Rhythmic power is proportional to the magnitude of the changes in activity level and the regularity with which they occur.  Since the increased sleep observed in the DmCa\textsubscript{v}3-null mutants does reduce the change in overall activity level between subjective day and subjective night, the increased sleep must also cause a reduction in rhythmic power.  Although we were able to replicate the increased sleep phenotype of DmCa\textsubscript{v}3-null mutants via pan-neuronal knock-down of DmCa\textsubscript{v}3, we were unable to further narrow the cause of this phenotype to a specific neuronal subpopulation.  This was in spite of making many attempts with a host of neuronal Gal4 driver lines ranging from broadly expressed enhancer traps and neurotransmitter Gal4 drivers to much more narrowly expressed neuropeptide drivers. This suggests DmCa\textsubscript{v}3 may function in novel sleep circuits.  The ``three channel'' compensation hypothesis in mice may yet turn out to be correct, but our results in flies suggest that other factors---isoform-specific differences, differences related to protein--protein interactions, or even something completely unforeseen---may allow mice and flies lacking these broadly expressed and highly conserved ion channels to still function remarkably well.  It will be interesting to see whether future studies focused on the technically demanding study of isoform-specific expression patterns and isoform-specific rescues in both mice and flies will clarify how T-type channels can at various times and in various contexts both enhance and reduce sleep.