# [Taxacom] barcode of life: PS

Jim Croft jim.croft at gmail.com
Sun Jul 4 21:02:10 CDT 2010

```This isn't really true is it?  Aren't notion(s) of short/tall,
dark/light, fat/thin, etc. and other continua are dependent on the
range, not on the distribution of the attribute within that range?
Most distributions of characters in nature follow a bell curve, but
what if they didn't?  Say the distribution was flat - individuals
clustered towards end would still be considered tall, the other short.
What if it was an inverse bell curve?  Don't tell me... different
species at each end, for sure... :)

Even with a bell curve, the notion of short vs tall vs normal is
arbitrary. You can pick your working percentile and I can pick mine.
And it applies a flat distribution as well.  If you are a 25% person
and I am 33% person all it impacts is the number of individuals each
of us would call short of tall.

jim

On Mon, Jul 5, 2010 at 10:11 AM, Stephen Thorpe
<stephen_thorpe at yahoo.co.nz> wrote:
> Analogy: if there were an equal proportion of people for every possible height, then there would be no useful concept of "tall people" or "short people". Of course you could arbitrarily set a particular height such that anybody above it was a "tall person", but this would not be a "real" distinction. A person who is 7ft tall really is a tall person simply because there is a distinct gap in nature between that and the normal range of heights (which is NOT to suggest that there aren't a few people with each "in between" height). Note that a 7ft tall person would still be a tall person if lots of people were 7ft. tall, provided only that there were relatively few people of heights somewhat less and then more again below that, so it isn't just about the rarity of 7ft. tall people, but about the whole frequency distribution of heights.
>
> It is all about finding (describing) patterns in nature - patterns which "really are out there" (and need not have been if things were different)

--
_________________
Jim Croft ~ jim.croft at gmail.com ~ +61-2-62509499 ~