Around 1895, a German embryologist named Hans Driesch split early starfish embryos in two halves, at the two cell stage, by strongly shaking a container of sea water. To his great surprise, each of the two cells developed into a normally-proportioned, "scale-model", half-sized pluteus larva (phenomenon A).
This also works with sea urchin embryos; and you can also separate at the 4 cell stage, and they will develop into normally-proportioned, "scale model", quarter-sized pluteus larvae. If you push together two one-cell stage embryos, they will develop into a double-sized larva. (phenomenon B).

Mammal embryos can be split into halves, quarters, or even eighths; with each fragment developing into a normal (and normal-sized) embryo and normal animal. (phenomenon C). Sometimes this is called "cloning", which I think is very misleading, and is meant to be misleading. It does not include any nuclear transplantation, in the sense of "cloning" Dolly the sheep. [In my opinion, "cloning" has become an over-used word: cloning of transplanted genes; cloning by nuclear transplantation in frogs and mammals; monoclonal antibodies produced by cell-cell fusion; cloning of genes etc. are not the same as each other].

Multiple mammal embryos, even as many as ten, can be merged together and develop into a single animal. If embryos of different species are mixed together (mouse and rat early embryos; sheep and goat embryos), then chimeric animals can develop, with normal size and proportions. (phenomenon D)

Dictyostelium slime mold slugs can be split into halves, or tenths, or hundredths, and each slug can "fruit" to form a normally proportioned stalk and lemon-shaped mass of spores. (phenomenon E) Normally-proportioned slime mold fruiting bodies can be formed by aggregations of as few as 15 cells, or as many as 150,000 or 250,000 cells. (phenomenon F) Starting with fewer cells is not the same as cutting up.

I (tentatively) regard all these phenomena as results of one underlying cause; but please notice their differences as well as their similarities. Nobody, including me, has previously used these "(phenomenon A,B,C,D,E,F)" terms, which I just now invented to emphasize differences between cases of "regulation".

Conversely, if you do Driesch's experiment, separating cells at the 2 or 4 cell stage, using embryos of clams, snails, sea squirts, and nematodes, then regulation will not occur. Instead, the separated cells will develop into shapes that look like halves or quarters of embryos of that species. This is called "mosaic development".

2) REVERSIBILITY OF CELL FATE DECISION: Maybe the difference is the irreversibility of determination of cell fate: regulative development reflects abilities of embryonic cells to change what they develop into, even up to the 4 or 8 cell stage; but mosaic development reflects irreversibility of cell fate, even at these early stages of development.

3) >METHOD OF CELL FATE DECISION: Maybe mosaic development results from cell fates being decided by "ooplasmic segregation", in the sense of differentiation being controlled by different special cytoplasmic materials being put into different early embryonic cells ("yellow crescent" in sea squirts; "gray crescent" in salamanders & frogs). This would be in contrast to control by cell-to-cell signals (or maybe extracellular signals), which embryologists call "embryonic induction". However, please note that Prof. Goldstein of this department proved that an example of induction does occur in nematode.

4) DISTANCE / "RANGE" OF CELL FATE DECISION: Maybe regulative development results when the signals that control cell fate can act at either long range, medium or short range, in contrast to mosaic development resulting from the use of signals that can only act over some particular distance? {Please try to invent more different possible mechanisms, causes, explanations}

His advocacy of this theory used an example of regeneration of an imagined flatworm that is made of three differentiated cell types, arranged in the pattern of the French flag, 1/3 red at one end, 1/3 white in the middle, and 1/3 blue at the other end. In the hypothesis, differentiation is controlled by local concentrations of an imagined diffusible chemical signal, which causes cells to differentiate into the blue cell type where the concentration is below some certain low threshold; likewise, wherever this chemical is more concentrated than a certain higher threshold, it causes the cells to differentiate into the red cell type.

Concentrations of signal molecule between these two thresholds are postulated to cause cells to differentiate into the white cell type. In addition, this theory postulates that cutting a worm in half will cause each gradient to adjust so that it has either a maximum or a minimum concentration at the new cut edge. By responding to the new pattern of concentrations of signal molecules, two scale-model French flag patterns will be formed. This is called the Theory of Positional Information.

Please note:

1) P.I. is a theory of regeneration in multicellular animals, more than regulation in embryos.
2) What regenerates is/are the diffusion gradients, with cell differentiation being subsequent.
3) It doesn't even try to explain adjustment of dimensions, to restore normal a-p/d-v proportions.
4) It is much more specific that just postulating diffusible chemicals.
5) The key point is that differentiation into all cell types is controlled by the same 3 chemicals.
6) Diffusion gradients had already been major parts of embryological theories for 50+ years;
7) What was new in this theory was the claim that regulation is impossible in any other way.
8) This theory is very attractive to molecular biologists & drew many to the study of embryos;
9) If true, it would mean that all development amounts to gene control by transcription factors.
10) The effects of the bicoid gene in fly development are much like this theory;
11) Fly development is mosaic, positional information was invented to explain regulation.
12) Actual regeneration is accomplished more by cell rearrangement than by re-differentiation.
13) Research about any embryological event can be reanalyzed as a search for signal molecules.
14) Identification of any diffusible molecule that alters development can seem to confirm P.I.
15) P.I. makes development 1-directional: chemical patterns develop first, anatomy is response
16) Supporters of P.I. think that they were the first to consider chemical gradients.
17) The complexity of actual embryos causes people to cling to the simplest possible explanations
18) It's hard to beat something with nothing! (Hypothetical specifics beat vague alternatives)
19) Most biologists in related fields think P.I. means any kind of chemical gradients or signals.
20) P.I. is a rationalization for transferring embryological grant funding to molecular genetics.
21) Anatomical differences between species would result entirely from differences in responses;
22) All kinds of animals would use the same three perpendicular, monotonic diffusion gradients;
23) Evolution of new anatomical patterns would result only from mutations changing responses.
24) Finding very similar hox genes in vertebrates as in insects is thus viewed as a confirmation.
25) Differences and changes in any other properties are regarded as being effects, not causes.
26) That's even true for monotonic gradients of non-diffusible proteins, like Ephrins.
27) Most embryologists once regarded P.I. as too absurd to waste time arguing against;
28) But now P.I. dominates the best-selling embryology textbooks.
29) The "wave-front" in the "Clock & Wave-Front" theory serves the function of a PI gradient.
30) Reaction-Diffusion theories are in opposition to P.I., but are related to the "Clock".