June 1, 1996
The data of R&D extend this finding to suggest that the phrase "birthdates of neural structures" in F&D's conclusion can be replaced by the broader phrase "dates of events in neurological development". In particular, the one event in neurological development that R&D considered most extensively is eye-opening. R&D showed that the date of eye-opening can be used to estimate with good accuracy the dates of numerous other events in neurological development. It follows from that conclusion that those dates can be estimated accurately from each other. Thus consistent with F&D's conclusion, it follows that the date of any one event can be used to estimate the dates of other neurological events.
The importance of the R&D finding is increased by the ease with which eye-opening can be observed, relative to the considerable difficulty of measuring birthdates of neural structures. It follows from their work that the dates of a whole series of events in neurological development can be estimated accurately from the date of one easily observed event.
Birth is of course observed even more easily than eye-opening. However, the R&D data suggest that birth itself is not usefully considered to be a "neurological event" in the sense used here. That is, the dates of various neurological events are not predicted nearly as well from the length of gestation as from the date of eye-opening.
R&D used a simple model in which the dates of other events were expressed as proportions of the date of eye-opening. That model is equivalent to a model of the form
log(date of other event) = log(date of eye-opening) + a
where a is constant across species but varies according to the event being predicted. That model differs subtly from the model used by F&D, which essentially replaced log(event date) by log(event date - 7). Let k denote the amount subtracted from dates. When we applied the F&D approach to the R&D data, k came out to be 4 rather than 7. Though we are not sure that 7 is the correct value, for the moment we have more confidence in it than in 4, for two reasons. First, the value of 7 was derived in a larger data set than the R&D set. Second, the R&D data set emphasized events late in neurological development, and k is not well estimated from such events. (Of course, estimating k was not one of the purposes of R&D; they assumed k to be 0.) Thus, at least for the moment, we continue to support a model with k = 7, over either the value of 4 suggested by the R&D data or the value of 0 tacitly assumed by R&D.
F&D concluded that a different scaling factor must be used for marsupials than for eutherian mammals. Essentially, development appears to be more protracted for marsupials than for eutherians. That conclusion might appear to be contradicted by R&D's title ("The visual pathways of eutherian mammals and marsupials develop according to a common timetable".) However, a closer reading of R&D shows that they intended no such contradiction. On page 184 they say, "Nevertheless, it is quite apparent that developmental events in the first half of the CP [caecal period--the period before eye-opening] generally occur much earlier [relative to eye-opening] in marsupials than they do in eutherian mammals....By contrast, most of the events that occur during the second half of the CP do so at about the same stage [i.e., the same proportion of the CP] in both marsupials and mammals." This difference between marsupials and eutherians is exactly the difference captured by the use of different coefficients for marsupials and eutherians in the F&D model. Thus the R&D data support that model in this respect as well.
In summary, the data and work of R&D suggest the following points about the F&D model:
Finlay, Barbara and Richard Darlington (1995), "Linked regularities in the development and evolution of mammalian brains" Science 268:1578-1584
Robinson, S. R. and B. Dreher (1990), "The visual pathways of eutherian mammals and marsupials develop according to a common timetable", Brain Behavior and Evolution, 36: 177-195
Go to Darlington home page