Comparison of ξ_GG(r) in PPS2 for the loose group sample TB96₂ and other samples identified as in previous analyses. (a) As in JZ88/GH83 (D₀ = 0.52, V₀ = 600; here we adopt φ(L) as in RGH89), for both PPS1 and PPS2. (b) As in RGH90/RGH89 (D₀ = 0.27, V₀ = 350; φ(L) as in RGH89), for PPS2. (The parameters of TB96₂ are D₀ = 0.27, V₀ = 350, and φ(L) as in TB96a, TB96b.) The high noise in DGH83₁ is partially due to the low number of groups (N_G = 60), but the same degree of noise is also present in DGH83₂ (N_G = 185). The error bars of RGH89₂ are similar to (but sometimes larger than) those of TB96₂, though N_G = 192 in both cases. (c) TB96₂ with its high‐ and low‐density counterparts. (d) TB96₂ with its high and low velocity dispersion counterparts (hot₂ and cold₂, respectively). For the group sample TB96NW₂, also shown, both links (normalized as for TB96₂) were scaled with the recipe of NW87. Error bars (omitted for clarity) are always similar to those in Fig.3. Note how ξ_GG is only moderately sensitive to variations of D_L, while it is almost independent of V_L; the effect of φ(L) is small. In some cases amplitudes vary, but slopes are almost always unchanged.