Certainly the question "Did prokaryotes stop evolving into novel organisms which, if we knew about then, would have been declared different taxa?" is both legit and valid. The question "Why did bacteria stop evolving into novel organisms which, if we knew about then, would have been declared different taxa?" is not, as it presupposes that that is the case, and needs to be preceded by a thorough analysis of the data before it can be asked. The significance of the second question -- which is the one you ask -- is precisely nil, until the first question has unequivocally been answered, "Yes".
Randman is, of course, correct. That is why we only see one kind of flying organism, one kind of swimming organism, and one kind of walking organism.
These quotes are from 1973 (1), 1929, and 1951, respectively. They are therefore not necessarily relevant to a discussion on modern evolutionary theory, as they are from before much of the genetic, paleontological, and theoretical work that today forms much of the basis of said theory was carried out.
You have made this statement before, and it is not less of a wrong conclusion this time. Grassé and the other pre-1973 authors you and Davison cite are not necessarily idiots, but they certainly -- indeed necessarily -- based their arguments and conclusions on the data available to them at the time, and on how biology and evolutionary theory was understood at the time. This is not necessarily the same as how these matters are understood today, and thus it is irrelevant what they wrote or believed, as they didn't have access to the same data that we have today. I believe you cannot show that Grassé, Bateson, Berg, Broom, Goldschmidt and Schindewolf -- the authors on whom Davison allegedly bases his claims, which is more or less by definition the same as the ones on which randman bases his -- would have drawn the same conclusions had they lived today, but feel free to prove me wrong.
Of course they do. They do it all the time, likely, especially as the industrial sector keeps finding new compounds that they can dump in nature. However, finding and occupying a niche does in no way imply evolving multicellularity, let alone evolving into a creature like a mammal or a reptile. There are several reasons for this.
First of all, as several people have told you, evolutionary theory does not oblige a given organism to evolve into a resemblance of any other organism. This is true even on such a basic level as multicellularity. If a certain niche could be filled by either a multicellular organism or a unicellular organism, and both organisms would initially (that is, prior to specialization for that particular niche) fill this niche equally well, the unicellular organism is no more obliged to evolve into a multicellular one than the multicellular one is to evolve into a unicellular one. Certainly there is a possibility that it may happen, but it is in no way a requirement, and it is certainly not expected.
On a less fundamental level, we of course find parallelism and convergence all over the animal kingdom (and possibly the other kingdoms as well, but I am less familiar with these, and will leave them out unless prompted to include them). However, almost invariably these examples are either spatially or temporarily isolated from each other, or have been over evolutionary time. We expect that a given taxon in the general case will over time become more specialized for the niche it occupies, or for a part of it. Therefore, upon discovering a novel niche for a given taxon, it is not necessarily its privilege to occupy it, as it may already be occupied by members of a different taxon. This taxon is likely to be better adapted to this particular niche than your taxon is, why an additional factor of some sort of required for your taxon to succeed in occupying that niche.
Bluntly, one might say that it is typically a case of "first come, first serve", where any subsequent colonizer of a given niche (absent any extinction of its previous tenant) would need an additional factor of some sort in order to make it prevail. This could be as simple as a resistance to a toxin or a parasite, an ability to switch between the novel niche and another one, or a behavioural, reproductive, or morphological difference between the two taxa that allows one to supplant the other. We expect that the difficulty for an organism to supplant another in a given niche typically would (again absent extinction) increase over time, provided the niche is sufficiently stable.
However, there is at least one factor that may prove this expectation to be flawed, and that is the fact that evolution, while capable of producing a high amount of variation, nonetheless is typically obliged to operate within certain limits. The most obvious, and most relevant, of these limits, is that evolution is expected to operate only on extant material. Evolution may modify existing material -- sometimes substantially, as in the transition from fin to foot to hand to wing -- but more sensational changes in morphology and other traits would typically be expected to require more sensational changes in genetics. For instance, it has been noted (e.g. by Christensen, 1980) that there is a clear correlation in Annelids between a transition from sexual to asexual reproduction, dissolution or severe modification of sexual organs, and polyploidization.
That is not to say that evolution is bound to change a taxon within narrow limits as in "horses changing into other horses". It is more a case of evolution being bound to change a taxon based on what is already present, but these already present features can be changed a lot, or even made to disappear, as in sexual organs of many Annelids, legs in snakes and whales, and teeth in birds. This means that evolution could change a present-day bird into something that would not be classified as a bird any longer by a lay-person (but still technically be a bird, because of the way taxonomy works), but it could not change into something which would require it to have 19 legs, external gills, or unicellularity, without a correspondingly sensational change in the genome.
This is relevant in that evolution in bacteria would be expected to be limited by the same factors as evolution in metazoans: it is forced to operate on what already exists. It is certainly possible for bacteria to evolve multicellularity, but it is not at all expected, as long as the problems connected with filling any given novel niche that bacteria are exposed to, and successfully out-competing any other taxa seeking to do the same, can be solved within the framework of unicellularity. This is what we expect would happen, as modifications of the existing framework is easier (and more likely) than the adoption of a radically novel framework.
This also implies that if a novel taxon in a certain niche has less constraint in the forms it may take -- or even just shorter generation times for evolution to operate on -- it may under some circumstances outcompete an established taxon in the same niche, if that taxon has more evolutionary constraint or longer generation times. Such constraint could be, for instance, that of the size required to fit all the organs needed for the established taxon's particular mode of utilization of the niche.
While considering this, you should still bear in mind that there is no point in multicellularity for the sake of multicellularity. There is no innate drive towards multicellularity in nature, and neither multicellularity nor unicellularity are by definition "better", "higher" or "more desirable" than the other. Multicellularity has occurred at least once, perhaps many times, but there is no single piece of evidence that multicellularity is a necessary result of evolution, regardless of how much time you allow to pass.
On top of this is the fact that the probability of two taxa to evolve into a sufficiently similar morphology decreases with the phylogenetic distance between these taxa. Apart from on a very basic, and primitively defined, bauplan-level (where "worms" are "worms"), we expect to find parallelism and convergence (2) only in relatively closely related taxa. "Mammals", for instance, is a sufficiently small and homogenous taxon for us to expect that convergence between different and relatively distantly related subtaxa could be common, and that is what we find. We might expect the same within, for instance, Lumbriculida, Charadriiformes, and other smaller taxa. We don't expect it to be very common to find convergence between a mollusc and a money, as the last common ancestor is too removed, and if we find convergence between such distantly related taxa, we expect the similarity to be on a more fundamental level (e.g. having a shell, having antenna, having egg cocoons) and be dissimilar in the details.
The most amusing part is, of course, that if Davison's ideas were correct, we would expect to find these things. As all organisms would contain all the information required to form all other organisms, a simple rearrangement would be all that was necessary for bacteria to change into non-bacteria. This I believe is a necessary consequence of his "Ontogeny" paper. You then use the lack of evidence for this process as evidence against evolutionary theory, which claims the process is unlikely, and, implicitly, as evidence for Davison's alternative theory, under which the process is an inevitable conclusion. That is almost as mind-twisting as your claims on phylogenies.
The question certainly gives an informative insight into how little biology you know.
---
(1) Note that it says 1977 on Grassé's work. This, however, refers to the English translation. Until it has been demonstrated that this translation differs widely from the original version in that it takes into account facts and theories from the intervening years, it is safest to assume that 1973 is the accurate date for determining what data Grassé incorporated in that particular work.
(2) I here follow some people whose name escape me at the moment in making "parallelism" and "convergence" distinctive terms, with the former meaning, approximately, evolution of a homologous feature in two sister taxa or almost-sister taxa in the same direction after isolation. For instance, if a population of warblers is isolated on an island without the sexual selection pressures being changed, these could develop patterns and plumages similar to those on the mainland (their sister population) within the same time frame, without this being convergence. I believe this distinction, though perhaps not very common, is nonetheless useful.
