Dehydrated lucerne is used as a protein source in dairy cow rations, but little is known about the effects of lucerne on greenhouse gas production by animals. Eight Holstein dairy cows (average weight: 582 kg) were used in a replicated 4×4 Latin square design. They received diets based on either maize silage (M) or grass silage (G) (45% of diet on dry matter (DM) basis), with either soya bean meal (15% of diet DM) completed with beet pulp (15% of diet DM) (SP) or dehydrated lucerne (L) (30% of diet DM) as protein sources; MSP, ML, GSP and GL diets were calculated to meet energy requirements for milk production by dairy cows and degradable protein for rumen microbes. Dry matter intake (DMI) did not differ among diets (18.0 kg/day DMI); milk production was higher with SP diets than with L diets (26.0 v. 24.1 kg/day), but milk production did not vary with forage type. Milk fatty-acid (FA) composition was modified by both forage and protein sources: L and G diets resulted in less saturated FA, less linoleic acid, more trans-monounsaturated FA, and more linolenic acid than SP and M diets, respectively. Enteric methane (CH4) production, measured by the SF6 tracer method, was higher for G diets than for M diets, but did not differ with protein source. The same effects were observed when CH4 was expressed per kg milk. Minor effects of diets on rumen fermentation pattern were observed. Manure CH4 emissions estimated from faecal organic matter were negatively related to diet digestibility and were thus higher for L than SP diets, and higher for M than G diets; the resulting difference in total CH4 production was small. Owing to diet formulation constraints, N intake was higher for SP than for L diets; interaction between forage type and protein source was significant for N intake. The same statistical effects were found for N in milk. Faecal and urinary N losses were determined from total faeces and urine collection. Faecal N output was lower for M than for G diets but did not differ between protein sources. Urinary N output did not differ between forage types, but was lower for cows fed L diets than for cows fed SP diets, potentially resulting in lower ammonia emissions with L diets. Replacing soya bean meal plus beet pulp with dehydrated lucerne did not change CH4 production, but resulted in more N in faeces and less N in urine.