rwg>So far I only have primes through 11, Foo, I just remembered how I used to get these on a timeshared, 1 meg Macsyma over the Arpanet. For etas, you want to avoid nonconstant coefficients in the reduction process, and all you really need is the ECHELON function. So 0=-16777216*h13^2*h2^72-196608*h1^24*h13^2*h2^48+302875106592253*h1^22*h13^28*h2^24+605750213184506*h1^24*h13^26*h2^24+582452128062025*h1^26*h13^24*h2^24+351263437231252*h1^28*h13^22*h2^24+146681435327336*h1^30*h13^20*h2^24+44326807379140*h1^32*h13^18*h2^24+9858921494006*h1^34*h13^16*h2^24+1610126946220*h1^36*h13^14*h2^24+189124030238*h1^38*h13^12*h2^24+15274994020*h1^40*h13^10*h2^24+778915592*h1^42*h13^8*h2^24+21099988*h1^44*h13^6*h2^24+196885*h1^46*h13^4*h2^24-22*h1^48*h13^2*h2^24+h1^50*h2^24-h1^72*h13^2 2 72 24 2 48 0 = - 16777216 h13 h2 - 196608 h1 h13 h2 22 28 24 24 26 24 + 302875106592253 h1 h13 h2 + 605750213184506 h1 h13 h2 26 24 24 28 22 24 + 582452128062025 h1 h13 h2 + 351263437231252 h1 h13 h2 30 20 24 32 18 24 + 146681435327336 h1 h13 h2 + 44326807379140 h1 h13 h2 34 16 24 36 14 24 + 9858921494006 h1 h13 h2 + 1610126946220 h1 h13 h2 38 12 24 40 10 24 + 189124030238 h1 h13 h2 + 15274994020 h1 h13 h2 42 8 24 44 6 24 + 778915592 h1 h13 h2 + 21099988 h1 h13 h2 46 4 24 48 2 24 50 24 72 2 + 196885 h1 h13 h2 - 22 h1 h13 h2 + h1 h2 - h1 h13 This is, of course, a completely nonrigorous and nearly mindless process, and somewhere there's probably a well developed theory of these things. --rwg