In most muscle cells, [ATP] is 5-8mM. Total [Mg] exceeds this so ATP exists mainly as MgATP^2-and is the effective substrate for cellular ATPases, including the actomyosin crossbridge (XB). However, ATP also binds Ca²⁺ .Since [Ca²⁺] rises during cardiac activation, so too will [CaATP]. Others have suggested that CaATP can be a substrate. We investigated this hypothesis for cardiac muscle. Rigor develops when ATP is removed in skinned preparations, but it can be reversed when (Mg)ATP is restored. One test is whether CaATP, like MgATP, can relax cardiac muscle from rigor; another is whether CaATP can fuel Ca-activated contraction in the absence of MgATP.Fine ventricular trabeculae from rats (killed humanely) were fully chemically skinned and isometric force recorded (see Miller et al, 2004 for methods). We have tested various [CaATP], and also various [ATP] (divalent cation-free). Rigor force is very sensitive to the conditions prevailing as it develops (Smith & Steele, 1994). We evoked rigor at pCa >7.5 by simultaneously removing Mg and ATP (zero creatine phosphate). Studies with CaATP may be confused since a few µM contaminating [Mg] may permit XB detachment and cycling. Adding 10µM Mg (muscle pre-equilibrated with 0.5-20mM ATP) produces some relaxation, whereas 50µM evokes near full relaxation of rigor (n=6). 1mM EDTA (sufficient to lower both [Ca²⁺ ] and [Mg²⁺] to <10^-9M) fails to affect rigor tension versus control (nominally Mg-free). We conclude that contaminating Mg is well below 10µM and thus not significant here.At high concentrations (>20mM), ATP alone (pCa>8, pMg>6) can almost entirely relax muscle from rigor. Under similar circumstances, CaATP (5 to >200µM) can produce biphasic responses (n=6). An initial relaxation takes several seconds, followed by slower force redevelopment to a new steady-state. Redeveloped force is a function of both [CaATP] (Km 1.65±0.4µM (s.e.m.) at pCa 6.3) and [ATP], as well as [Ca²⁺].Preliminary tests reveal that muscle stiffness in CaATP-supported force is frequency-dependent. This suggests that CaATP-supported tension involves active rather than rigor XBs. However, their characteristic frequency (fmin) is very much lower (<0.03Hz) that that of MgATP-fuelled XBs (c.2Hz at 22^oC). This contrasts with Ca vs MgATPase rates for (skeletal) myosin (Lymn & Taylor, 1970, Mineheardt et al, 2002). It remains to be established how far contraction is influenced by CaATP formed during normal, Ca-dependent activation. However, even if just a small fraction of XBs is CaATP-fuelled, their slowness could be significant for contraction kinetics because of cooperativity.