So to be clear - species A has several mutations over a long timespan, creating A1, A2, A3, A4. Over another timespan, each of these have mutations (and some of the original species die out); we now have: A11, A12, A13, A14 from A1, and also have A31, A32, A33, and A34 from A3. A2 and A4 die out; they were viable mutations, but selected out due to whatever factor. Over more time, more mutations occur. We now have A121 A122 A123 A124 (A11 died out), A131 A132 A133 A134 (from A13, and A14 dies), and similar for A331, A332, A333, A334. Over more time... etc, you get the idea. So yes, given my naming convention, for some currently existing species A331 there were prior splits at A33, A3, and A, each building off the prior in some manner. But you say this accumulation is false - now to why. Generation is very much not the same as species. A single new viable species might occur after millions of generations, including many false starts where after (4) generations all (13) carriers of that mutation die off (small numbers for example). It's only after that this does manage to happen successfully - and we do have evidence for this happening in fast breeding animals such as fruit flies - and a new mutation (or several) establishes themselves as independent over the course of tens to hundreds to thousands of generations do we define the creation of a new species. This may be one source of misunderstanding on evolution; it is slow as fudge. Generation is very much not the same as mutation. It is in fact probable that an individual animal with a mutation would make that animal less likely to reproduce, or even make it unable to do so. In only very rare cases does a failure of a *species*, once it dies out, fall back to "step zero". When A11 and A14 die out, A12 and A13 still exist. Certainly if all of A11 A12 A13 and A14 die, then yes, the A1 line is extinct. Even this happens all the time in nature, we agree. But what about A33? That's far beyond 'step zero' of A. What about H7, or L2315 that also show promise? We agree, a single species failing is common; but you seem to think that species cannot branch out, if I understand correctly? We've gone back and forth a little bit on 'accumulative'. The end result of a specific species or family thereof sharing some trait ("flight") is accumulative, yes. However, this is only from the standpoint of looking at the end result, and working backward. Evolution, as I've described here, does not aim or attempt to get to the end result. It simply shotguns changes, constantly, and sees what sticks. Over time, collections of features arise that humans like to put into boxes and call 'flight', 'sight', 'hearing', etc. But there exists incredible variety in any feature you can name; even intelligence is a sliding scale. Your choices of "otter" and "dolphin" are odd considering both are aquatic but from different routes as evolution predicts; otters share genetics with weasels, land carnivores ( https://en.wikipedia.org/wiki/Mustelidae ) and dolphins with whales, which remain unchanged for a very long time ( https://en.wikipedia.org/wiki/Cetacea ). Their common ancestor would be way, way, way back. I'm going to assume you mean a generic aquatic / land transition. In such a case, what precisely would require such a species to have a lower survival rate than one that was purely either/or? This is a claim with no proof, but stated with what feels like considerable belief behind it. In the general case, partial adaptation takes many steps. One species has better lung capacity, another better eyesight underwater (one mutation present in humans by the way! https://www.ncbi.nlm.nih.gov/pubmed/12747831 This is also where I can point out I'm also not quite using species correctly, or even the concept of purely mutations as able to create new species. Periodic isolation is also required; see https://en.wikipedia.org/wiki/Speciation ), another mutation increases mobility underwater, etc, until the transition completes, or doesn't. There are a great many species that are semi-aquatic, you own example of otter technically one of them! Therefore the conclusion drawn - requiring "a huge lot of magical luck" is drawn from a false premise, given the sheer variety we do in fact see for these partial transformations, and that such middle species are in fact viable. The given description of fish into amphibian into etc. is simply not what the theory of evolution predicts or suggests, but is an often quoted strawman. Please briefly examine a diagram of mapped branches ( https://evogeneao.s3.amazonaws.com/images/tree_of_life/tree-of-life_2000.png ) ; this should hopefully not seem linear as you say. This mapping is done by a mix of both examining genetic markers for similarities, as well as yes some guesswork for species that have died out (of which extremely few are shown on that diagram!) and ordering of which branches appeared first, as time goes further and further back and mapping becomes more difficult. So yes, there are indeed several branches even "by the Fishes" : https://en.wikipedia.org/wiki/Agnatha , https://en.wikipedia.org/wiki/Chondrichthyes , https://en.wikipedia.org/wiki/Osteichthyes for three at random. Trying to fit everything so tightly into either "fish" or "amphibian" or "reptile" or "bird" etc, loses so much definition to be not worthwhile to discuss. Sharks and dolphins and whales do not, in fact, strictly share food sources, but perhaps this was a bad example on your part. Further, this is again entering into an ecology / evolution confusion, not to mention more oversimplification. The very concept of a species differentiation implies there exists some difference, for which species F might be slightly better at some things and worst and others than species G. Ergo, if F begins to dominate, not only does that change the ecosystem, but G may be able to survive by capitalizing on what it can do better than F. If it can't then yes, G may perish in the local region. This happens. But it does not imply a civ-style winner-takes all approach, where F slowly pushes everyone else out of the ecosystem! Bacteria not devouring the world is a good point! There are a few comments I'd leave on this, partly because I'm not certain of it myself! One thing is that the bacterial world is incredibly, incredibly, incredibly competitive compared to what we'd consider the macro biological world; animals we can see and stuff. If one bacteria does very well and starts to eat all the other ones, then whichever bacteria that can find a weakness in the previously good one is suddenly eating like a king and does well themselves, leading to pressure to eat the second, etc. Additionally, optimization is absolutely critical here. If a bacterium develops some means to do X, that comes at the cost of resources to create that, which necessarily slows down growth rate unless what it gains can improve growth rate more than the cost; you may repeat this many times, but eventually, you end up with bigger and bigger bacterium, which suffer from problems that smaller ones don't. These are my thoughts at any rate... but I'm certain there are better, evidence-backed reasons. My training is in physics and microscopy, but @Kazavon might know more? Alternatively - you can always search the internet for reputable sources yourself ^.^ If you do not believe that rabbits and wolves can exist in a balanced ecosystem by virtue of balance being impossible without magic or external control, then does this external control also govern your refrigerator, which is also a self governing system? What about you - does it tell you when to eat and when you're full, or is there some chemical system that does it? I'm being a little exaggerating here and perhaps not so kind, I do apologize. I should probably wrap this up and sleep soon... Consider: Is there anything magical about a ball resting in the bottom of a bowl? If you disturb the ball by poking it, or placing it partway up the sloped sides of a bowl, it is no longer stable; the sloped sides of the bowl guide it toward the center, revealing a stable equilibrium at the bottom. Even if the ball is bumpy or rough, or the bowl is lightly moving in irregular motions, the average location would still be the bottom of the bowl. Ecosystems occur under the same general principle, even if more complex. A very simplified version of it - wolves eat too many rabbits -> not enough rabbits -> wolves starve or have fewer pups -> more rabbits live, or at least, the rate of rabbit decline decreases, and a new equilibrium is attained. It is certainly possible for ecosystems to be unbalanced - invasive species are often this - but the presence of an unbalanced system does by no means prove balanced systems cannot exist. If you wish to, by all means. I've some time, and find it useful to talk to people with different viewpoints; questioning your own ideas, seeing evidence based discussion and deciding thereon is, imo, one of the better things a person can do.