Hubble discovered galaxies in 1922. He found elliptically shaped objects and disky ones showing spiral arms sometimes with a bar at the center. He also thought that because of Jean’s law, elliptical galaxies would flatten into a disk, so the elliptical galaxies are the most ancestral ones.
We can code Hubble’s initial observation with to characters:
- the arms
- the bar
with in both cases “0” meaning absence and “1” presence. According to the Jean’s law, then “0” for the arms is the ancestral state. There is no obvious polarization for the character bar. The matrix describing galaxies is thus:
| arms | bar | |
| E (elliptical) | 0 | 0 |
| S (spiral) | 1 | 0 |
| Sb (barred spiral) | 1 | 1 |
Humm, easy matrix for cladistics, no?
The corresponding unrooted cladogram is (the tick marks indicate a change of the state of the character, from 0 to 1 or conversely).:
It looks exactly like the extremely famous Hubble diagram also known as the tuning fork diagram!
I sometimes wonder whether Hubble discovered cladistics in the 1930s, well before Hennig! But I also know that linguists have long used such a methodology without formalizing it like Hennig did.
A rooted representation, provided that we assume that all galaxies have a common ancestor (careful, not an individual, but a species or a type), is more convenient to me because it removes the apparent direction of evolution imposed by the Hubble diagram. This gives:
Of course, as cladists know, this representation is more open to discoveries of new objects and new characters. However, we know today that the arms and the bars in galaxies are non persistent and probably do not say anything about the evolutionary state of a galaxy. In addition, the morphology may be acquired through different assembly processes. But there are so many other observables…
Astrocladistics 