“…In another direction, Riemann's result was generalized in 21 by Laszlo, where it was shown that if d = r ( g − 1), and [ E ] ∈ B 1 r , r ( g − 1) , then \documentclass{article}\usepackage{amssymb,mathrsfs}\begin{document}\pagestyle{empty}$\rm {mult}_{[E]} B^1_{r,r(g-1)} =h^0(C,E)$\end{document}. We point out that the Zariski tangent spaces to the B k r , d are well understood: at a point E ∈ B k r , d , if E ∉ B k + 1 r , d , the tangent space to B k r , d at E is where is the cup product mapping, and if E ∈ B k + 1 r , d , then \documentclass{article}\usepackage{amssymb,mathrsfs}\begin{document}\pagestyle{empty}$T_EB^k_{r,d}=T_E\mathscr{U}(r,d)$\end{document} (see 12 for more details). Using work of Kempf 16 as motivation, the following extends the results above by giving a description of C E B k r , d , the tangent cone to B k r , d at E .…”