<p>
	Each panel is one measure, titled with its column name in the snapshot
	history. Each color is one membership rule; the dashed line is snapshots
	rebuilt from studbook dates.
</p>

<p>
	<b>Which way each panel moves</b>
</p>
<ul>
	<li>
		<b>fe, fg and neGD</b> (founder equivalents, founder genome equivalents
		and gene diversity): higher means more of the founders' diversity is
		still in the colony. fe falls when the founders' contributions grow more
		unequal; fg and neGD also fall when founder alleles are lost by chance.
		neGD is a share between 0 and 1, not a number of animals.
	</li>
	<li>
		<b>nFounders, nMaleFounders and nFemaleFounders</b>: the known founders
		(animals with no recorded parents) among the analyzed animals and their
		ancestors. They rise when such an animal joins, an import for example;
		in a series of living animals they also fall when a founder and all its
		descendants have left.
	</li>
	<li>
		<b>neSexRatio and neVariance</b> (the sex-ratio and variance effective
		sizes, counted over the living breeders): higher means less diversity is
		lost each generation. neSexRatio falls when breeding relies on few
		animals of one sex; neVariance falls when a few breeders produce most of
		the offspring. Read each as an index, not a head count.
	</li>
	<li>
		<b>meanIndivMeanKin and medianIndivMeanKin</b>: the colony's average and
		median of each animal's mean kinship. A rise means the animals are, on
		average, more closely related to the rest of the colony.
	</li>
	<li>
		<b>meanGu and medianGu</b> (genome uniqueness): neither direction is
		better. They fall when a rare founder allele is lost, and also when
		breeding spreads it into more animals, so a fall alone is not a loss; fg
		and neGD show loss.
	</li>
	<li>
		<b>The skewness and kurtosis panels</b>: the shape of the spread across
		animals, not a measure of health. Positive skewness means a few animals
		with unusually high values stretch the upper tail.
	</li>
	<li>
		<b>fgSE and meanGuSE</b>: how precise fg and meanGu are, not how healthy
		the colony is. They shrink when more gene-drop iterations are run.
	</li>
</ul>

<p>
	<b>When a change is real</b>
</p>
<ul>
	<li>
		The grey band around fg and meanGu spans one standard error on each side
		of each point. A change smaller than the bands can come from the random
		sampling alone.
	</li>
	<li>
		fe, the founder counts, the effective sizes and the mean-kinship panels
		have no band because they are calculated without random sampling: a
		change in them is a change in the records. The other genome-uniqueness
		panels have no band but come from the same sampling as meanGu.
	</li>
	<li>
		A point drawn with a different shape was taken with different settings
		(the number of gene-drop iterations, the rare-allele threshold or the
		package version) than the point before it, and the caption names the
		change; a jump there may come from the settings, not the colony.
	</li>
</ul>

<p>
	<b>What a past trend can answer</b>
</p>
<ul>
	<li>
		How fast founder diversity is being lost: the slope of fg and neGD
		across the years.
	</li>
	<li>
		Whether an import or a change in breeding shows: an unrelated import
		with no recorded parents lowers mean kinship and raises the founder
		counts as soon as it is in the population, and raises fe and fg once it
		has offspring there. A change in breeding practice shows as a change of
		slope from that year on, for example a fall in neSexRatio when fewer
		sires are used.
	</li>
	<li>
		A baseline: rebuilt years let today's snapshot of the living animals be
		read against the colony's past instead of alone.
	</li>
</ul>