This stage upon stage suggest a sense of order. If the process of evolution had been purely random, human pre-develop should appear much more unique, instead of a composite of previous layers tweaked in the form of another layer. Once a layered is mastered, another is added, which is gaining potential, sort of like moving to higher and higher energy levels.
It is precisely the product of random events. That is what embryology teaches. We discover the hidden layers of every lower life form within the first 6 to 8 weeks. Hidden to primitive people, that is. Today we see these played out in 3d and in Technicolor, as plain as the nose on your face.
Take the blastocyst: it is found in all the metazoa. Why? It is inherited. From what? Most likely, from colonial forms in which cells cluster in spheres. I mentioned Volvox, the spherical algae. Can you see the commonality in form (a hollow sphere of identical cells, which signal and specialize)? The process for developing ecto-, endo- and mesoderm "germ layers" is found in all species from the flatworm to humans. The gills in the human embryo (evolving in the larynx), the shrimp-like body plan, the tail, i.e, the vertebral column, the division into banded stripes that will define specific body regions where limbs or organs will articulate...all of these features are just the most basic of evidence that evolution proceeded in a manner completely driven by changing stress and adaptation.
The appearance of order as a violation of entropy, without considering the vast energy sources these creatures used to gain a foothold in a particular niche, is NOT an increase in entropy. You are forgetting that they BROKE DOWN the materials they absorbed, and released waste products in more disordered states. This is why you must account for TOTAL ENERGY GAIN OR LOSS ACROSS A BOUNDARY before you can invoke the 2nd law of thermodynamics as you are hoping to do.
Beyond these fairly superficial comparisons I described, in embryonic development, are the deeply significant interactions involving gene expression, only to unpack, almost step by step, the sequence of evolutionary stages that our ancestors experienced in sequential stages of evolution. These processes, which rely entirely on the inordinate coding capability of DNA, through its cooperating mechanisms (transcription factors, mRNA, ribosomes, etc.) reach into the germ layer to build the molecular framework, beginning with protein diffusion as a local enabler for a particular localized metamorphoses to take place.
Against the background that this process is coded and predictable lies the huge influence that random processes play. We already talked about crossover (merging of grandmother and grandfather alleles in the meiosis of he mother. Crossover is entirely random. This should give you a clue that there is no entropic cause at play during crossover. From the available set of genes, you will get a random selection from your mother and a random set from your father. And THAT PARTICULAR SET, if the offspring survives, is passed down to subsequent generations, merged with a RANDOMLY CHOSEN partner's egg or sperm. There is no energy gained or lost, because all of the energy absorbed into the life processes of an individual do not correlate with the work done by the offspring. They are not even concurrent, and there is no way to correlate entropy back and forth through time.
Most importantly, you should reach out for an understanding of random processes. They are part of nature, and in fact the normal distribution (bell curve) is called "normal" because it normally occurs in nature.
The alternative view, that evolution accomplishes what it needs, is bogus:
Natural selection has no intentions or senses; it cannot sense what a species or an individual "needs." Natural selection acts on the genetic variation in a population, and this genetic variation is generated by random mutation — a process that is unaffected by what organisms in the population need. If a population happens to have genetic variation that allows some individuals to survive a challenge better than others or reproduce more than others, then those individuals will have more offspring in the next generation, and the population will evolve. If that genetic variation is not in the population, the population may survive anyway (but not evolve via natural selection) or it may die out. But it will not be granted what it "needs" by natural selection.
http://evolution.berkeley.edu/evolibrary/misconceptions_faq.php
If you are not satisfied that the randomization in human reproduction is sufficient, consider the general case in nature in which the process is entirely random. If you're not satisfied that animals breed randomly, consider plants. Surely you won't deny that fertilization in plants can simply be a matter of which way the wind blows.
