Some of this you'll read and say, "Well, DUH! Of course ... blah blah" But it's always best to make sure those "of course" assumptions are true. Sometimes they are not; and sometimes they are only partially true.

---

If we look at the Visible Maximum Size over a large population, we notice we have lots of mid-sized and only a few very large or very small. This means Maximum Size is almost certainly a combination of two or more values and is a trait we can breed for.

Next, if look at Maximum Size in combination with other Visible Traits, we see no clear relationship. This means Maximum Size is not only a trait we can breed for, but we don't need to worry about other traits effecting things, and we don't need to worry about Maximum Size effecting our breeding for other traits at the same time.

The designer's public statements claim a random factor applies AFTER determining the genetics for the new Nest. Grouping parents by their Visible Maximum Size and examining the sizes of their offspring produces approximately equal numbers for each size. This means the claim is correct, a random factor does apply. At the same time we notice that some parent sizes never produce offspring outside a range. This means the randomness is bounded by the parent's size. Obviously, it is also bounded by the fact that there are only 9 sizes.

The effect of this randomization means it's *possible* Maximum Size has some relationship to another trait. But, if so, it is a weak relationship, and any evidence of it was lost by the randomization. Careful breeding over several generations may expose a relationship but, at this time, insufficient samples exist with enough history *and* cross-line breeding to definitively state there is, or is not, any relationship to another trait.

---

Given this, I propose the follow model for Maximum Size genetics.

A) Each parent has two Maximum Size values. One value was received from its mother and the other from its father.

B) Upon procreation, each parent randomly selects one of its Maximum Size values to pass to the offspring.

C) The offspring determines the next-smaller size from the smaller size passed to it, and the next-larger size from the larger size passed to it. This is the range of possible Visible Maximum Sizes for *this* offspring. For example, if we received Toy(3.5) and Large(10.5), the range we'll use is Teacup(1.75) through Giant(12.25).

D) The offspring randomly chooses a Visible Maximum Size within this range.

At this point we are done. We have two values to be used for further procreation and we have the Visible Maximum Size.

Here are some additional observations from comparing this model to real values observed in the wild.

1) In almost every case, we do not need to examine beyond the parents to make a good guess at the range of sizes possible for their offspring. This is due to bounding limits for the random number. Yes, it's *possible* to get a size at, or even beyond, the edges, but it's far more *probable* the Visible Maximum Size of the child falls between its parents' Visible Maximum Sizes.

2) In those few cases where the child produced appears to lie outside the range, we look to the grand-parents. The child's Visible Maximum Size will probably be within their limits. If this fails, we can look back another generation to explain an apparently abnormal value. At some point we reach Levio and Dawnara and punt because at that point all things are possible.

The number of times we have to do this decreases significantly with each generation we're removed from Levio and Dawnara. Basically, the odds strongly favor the Visible Maximum Size being between the parent's Visible Maximum Sizes.

3) If, as breeders, we immediately cull any offspring which lies outside its parents' range, we're eliminating that wild gene which was quietly passed down from Levio and Dawnara to finally appear in the child. We can re-classify that wild gene into another size group (see below), or release it if we don't want it. This will occur most frequently in Generation 2, and will occur significantly less frequently in each succeeding generation. By Generation 4 or 5 the odds of a wild gene remaining in the family line are extremely small.

4) If, as breeders, we separate out Visible Maximum Size into three groups of three sizes:

a) the largest Size group (10.5 to 14.0) almost certainly eliminates the smallest half-range (0.0 to 7.0) from consideration.

b) the middle Size group can produce anything

c) the smallest Size (0.0 to 3.5) almost certainly eliminates the largest half-range (8.0 to 14.0) from consideration.

5) The limits imposed by the design of Maximum Size, taken with the limits imposed by the randomizing scheme, mean the smallest (0.0) and largest (14.0) sizes, when bred to another of the same size, have the best odds (1 in 2) of producing another of the same size (the other will be the next-best size: 1.75 or 12.25, and should be considered for culling if it presents no other desirable traits). All other sizes, when bred to another of the same size only have best-case odds of 1 in 3 of producing another of the same size.

6) For sizes not at the extreme limits, we can slowly walk toward the limits by choosing to keep Visible Maximum Sizes toward the size we desire, and culling Visible Maximum Sizes which work away from it. So, if you find yourself with nothing but Average and Medium examples, you can still obtain Teacup(0.0) and Giant(14.0); it will just take you a little longer (on average: three times longer) than if you had some Toys or Larges to begin with.