21 October 1976
Dear Francis,
Thank you for all the new letters, particularly the correspondence with Bak.  It certainly is an exciting time and I very much like your idea of the supersolenoid.  I have called a chromatin meeting for this Friday, at which Ashburner and Sydney will be present.  I have already had a word with Sydney, and he thinks your point (in your letter of 14 October) about translocations is a good one.  I will write again after the meeting at which the genetic implications will have been discussed.
On the supersolenoid itself, I have now found my notes on the E.M. Meeting in Jerusalem, and Hans Ris definitely said that mitotic chromosomes were formed by folding the 250 A fibre into one of diameter about 1000-2000 A.  However he did not present any pictures (his slides were lost when his baggage was stolen at Paris Airport!) but I suspect they would show the same sort of murky pictures as before, and the idea of coiling into a thicker fibre is just a general one.  I doubt if he has any hard evidence, but he may have some estimates of diameters from sections.  This value may be a tricky one since I would imagine that it could change fairly easily if the coils of the supersolenoid were not held by ties.  Your secondary idea of the ties being preserved is a very neat one, but I wonder if it isn't all oversimplified.  Sidney thought it unlikely, but I do see the point of the possible correlation between the smallness of Drosophila metaphase chromosome and that smaller size of the genes.
I have been looking at papers on the electron microscopy of sectioned mitotic chromosomes and there is just no trace of a hollow centre.  There are many examples, but I enclose a reprint of an old paper of Hugh's, see figure 14.  Richard Skaer has also brought in some much more recent photographs, and what one sees in transverse section are fairly dense bodies about a micron across.  Even allowing for the possible double helical coiling at the next level, I think a hollow tube would have been evident even if it had been considerably flattened.  Perhaps the lumen is filled with other proteins and possibly some looser DNA.  This last notion is distasteful as I think you are right that the metaphase chromosome should have a compact, and therefore regular, structure.
affects the way in which Bak has done his calculation of the contraction ratio.  One minor point: perhaps you could change the reference to Bradbury in such a way that it is not cited as equal evidence with the paper by John and myself.  I don't know if you have read the paper itself (Carpenter et al., July 1976, Nucleic Acids Research) but the level of significance of the "split peak" is barely above noise level.  I personally don't think anyone could have deduced the existence of the solenoid, let alone the number of units per turn, from such pictures unless one knew what answer to expect.  (Moreover, the key pattern was taken at 32% relative humidity, which corresponds to a concentration of over 60%.)
By the way, one other point which you don't mention explicitly, but must be obvious to you.  If the interphase structure is to go continuously into the metaphase structure preserving the solenoids in the inactive genes, then I think your idea of DNA down the middle of a solenoid can't hold.  I append a sketch illustrating the point.  I have always favoured the situation (a) because of the fact that, in the original experiments in which we extracted (rather than reconstituted) solenoids in the presence of magnesium, they always looked curved.
Thank you also for your comments on the packing of the crystals.  I can't think for what reason you suggested, in your letter of 11 October, that you liked a model in which the dyads were tilted in the AC plane.  Indeed, I had something very like this in mind when I wrote in my letter that the ideal packing of columns, that I had proposed, would be perturbed.  I draw a picture showing the general type of perturbation which indeed does account for the difference in the X-ray patterns looking down the 110 A axis and the 192 A axis.  How did you derive it, unless you remembered the pictures themselves?!  I am still not sure, but I think I have evidence for the approximate 120 degrees rotation by considering the distribution of intensities, but more on that later.  In any case, I am not sure how your particular detailed model would relate to the packing in a solenoid.  Why would tilts of 30 degrees be required for this?
More soon,
Yours ever,
A. Klug