The pathogenesis of Theiler's murine encephalomyelitis virus (TMEV)-induced demyelinating disease is still controversial. Our hypothesis is that primary infection of oligodendrocytes (OLGs) is not a crucial event in the pathogenesis of demyelination in this model. In fact, it has been proposed that myelin may be destroyed, as an innocent bystander, following an antiviral delayed-type hypersensitivity (DTH) response. This hypothesis would not need widespread oligodendroglial infection, because virus present in other cells would be sufficient to trigger a DTH response. The present study demonstrates that cultured OLGs and astrocytes from susceptible strains of mice (SJL and DBA) and immortalized OLGs can be infected with TMEV in vitro. Infection of OLGs, however, is at very low levels and does not result in overt cytolytic effect. In contrast, infection of immortalized OLGs is very efficient and results in clear cytolysis. Because an important characteristic of DTH responses is the liberation of potentially injurious cytokines into adjacent tissues, we also examined the effects of mouse recombinant tumor necrosis factor-alpha (TNF-alpha), interleukin-1 alpha (IL-1 alpha), and interferon-gamma (IFN-gamma) on cultured OLGs and immortalized OLGs. We found that TNF-alpha caused immortalized OLG cytotoxicity in a time- and dose-dependent manner. In contrast, no cytotoxicity was observed on primary OLGs with any of the above cytokines. To determine whether functional effects could be demonstrated on primary OLGs by either virus or cytokines, we measured mRNA expression of different myelin proteins in primary and immortalized OLGs exposed to virus or TNF-alpha. Neither the BeAn strain or the GDVII strain of TMEV interfered with myelin protein mRNA expression in primary OLGs, whereas GDVII virus dramatically reduced myelin OLG glycoprotein (MOG) mRNA in immortalized OLGs. Interestingly, although even high concentrations of TNF-alpha (10,000 U/ml) did not produce primary OLG cytotoxicity, they resulted in a significant reduction in mRNA for both myelin basic protein (MBP) and MOG in these cells. TNF-alpha (at 500 U/ml) also specifically reduced MOG mRNA in immortalized OLGs. Because immortalized OLGs are considered to be arrested at an early stage of maturation, our results suggest that immature OLGs are susceptible to both virus- and cytokine-dependent cytotoxicity, whereas mature OLGs are resistant to cytolysis by either TMEV or cytokines. TNF-alpha, however, is capable of reducing mRNA expression of myelin proteins in primary OLGs; therefore, it may participate in the induction of demyelination, as suggested by the DTH-mediated hypothesis.

Bergstrom has shown that becker's “Rotten Kid theorem” holds in a world of two commodities if their utilities are transferable. The present paper identifies a further circumstance in which the theorem is valid. We show that it also holds in the absence of transferable utility if the externalities are assumed to take the form of a single pure public good.

In a recent article in this Journal, Cummings al. (1997) offer experimental evidence rejecting the hypothesis that a hypothetical referendum is an incentive‐compatible mechanism for obtaining the social value of, or willingness to pay for, changes in the allocation of environmental goods and services. While cautions in offering a blanket rejection of hypothetical referenda as a means of measuring economic values on the basis of their findings, the authors do suggest that at least in experimental settings respondents appear to behave differently when real money is changing hands. Wheather hypothetical situations adequately mimic real‐market situations has been debated in the contigent valuation literature since its inception (see Cummings, Brookshire, and Schulze 1986). Cummings et al.'s finding of incentive incompatibility appears to offer strong evidence for the critics of hypothetical referenda.

Analyses of common property extraction under free access follow two distinct paths, traditional and game‐theoretic, giving rise to two standard methodologies. One methodology avoids game‐theoretic analysis by assuming that aggregate extraction in each period induces fel rent dissipation. The second methodology solves for the Marko‐perfect equilibrium of an n‐player extraction game investigating aggregate behavior over time as n → ∞. We show by example that these coexisting standard methodologies can yield conflicting predictions. We then provide conditions, relatively easy to satisfy, sufficient for the two approaches to yield the same predictions.