This thesis investigated the relationships between the feeding behaviour and use of stored fat in overwintering Atlantic salmon parr remaining destined to remain resident in freshwater the following year. Experiments investigated the responses to periods of food shortage at different times of the year to assess the influence of season. Investigations were carried out to examine how feeding motivation and fat storage were influenced by changes in those environmental cues that indicate the change of season. The effect of the normal winter behavioural pattern upon feeding and fat was also investigated.

No description available.

Two trials were conducted to evaluate effects of grain adaptation protocol on subacute acidosis, feeding behaviour and ADG of cattle. In trial one, 12 crossbred heifers (384 ¡À 25 kg) were stepped from 40% to 90% dietary concentrate by either rapid adaptation (RA; one step-up diet fed for 3 d) or by gradual adaptation (GA; five step-up diets fed for 3 d each). Mean daily ruminal pH variables did not differ (P > 0.10) between treatments but variances of a number of ruminal pH variables were greater (P < 0.05) for RA than GA during adaptation to 65% and 90% concentrate. Mean hourly pH did not differ over the first 24 h of adaptation to 65% concentrate, but variance of hourly pH tended (P < 0.10) to be greater for RA than GA for eight of the first 24 h of adaptation to 90% concentrate. Increased variance in ruminal pH parameters was associated with detection of acidosis in certain individuals. On d 1 of 90% concentrate, ruminal pH tended (P = 0.07) to be lower at 11 and 12 h post-feeding with RA than with GA. Ruminal volatile fatty acids (VFA) and osmolality were similar between treatments. In trial two, 120 crossbred heifers (366 ¡À 23 kg) were adapted from 40% to 90% concentrate. A protocol identical to trial one was used with the addition of moderate adaptation (MA; three step-up diets fed for 3 d each). The increase to 65% concentrate caused reduced daily bunk attendance and increased maximum intermeal interval for RA compared to MA and GA cattle but the increase to 90% did not. ADG was reduced for RA compared to MA or GA during adaptation but over day 1 to 69 ADG did not differ between treatments (P ¡Ý 0.41). Current management strategies for preventing acidosis in pens of cattle are based on responses of the most susceptible individuals. Improved understanding of individual responses to acidotic challenge may allow development of more effective acidosis prevention practices.

Les noves directives de la UE sobre la protecció dels porcs (Sus scrofa) fan necessari el desenvolupament de sistemes d'allotjament en grup apropiats des del punt de vista del benestar i de la productivitat. L'objectiu general d'aquest estudi fou la comparació entre dos sistemes comercials d'allotjament en grup i les gàbies convencionals per a truges gestants, mitjançant mesures de comportament, fisiològiques i productives. Cent vuitanta truges gestants, de primer a novè part, se seleccionaren d'una granja comercial i foren emprades en tres rèpliques diferents (60 truges per rèplica). Les truges foren allotjades del dia 29 de gestació a una setmana abans del part en gàbies convencionals (STALL), en grups de 10 amb un sistema d'alimentació de caiguda lenta (TRICK) i en grups de 20 amb un dispensador de pinso electrònic sense protecció (FITMIX; 20 truges per sistema d'alimentació i rèplica). Totes les truges havien estat prèviament allotjades en gàbies i foren alimentades de forma igualment restringida. Les truges s'observaren 11 dies no consecutius durant 4 h al dia. En tots els sistemes, els nivells d'activitat general i d'estereotípies es mesuraren mitjançant la tècnica d'observació scan-sampling a intervals de 10 minuts. S'obtingueren dues mostres de sang per rèplica i es determinaren els nivells d'haptoglobina (Hp) i pig-MAP per valorar el grau d'estrès físic. El gruix de greix dorsal, el pes viu i els resultats reproductius també foren considerats. En els dos sistemes d'allotjament en grup, s'enregistraren les agressions mitjançant observació contínua i posteriorment es calculà l'ordre de dominància (RI). Al sistema FITMIX, s'analitzaren les dades de comportament alimentari enregistrades automàticament durant 25 cicles alimentaris no consecutius. El número de visites al dispensador, la duració de les visites i l'ocupació del dispensador s'analitzaren per dia i per hora per determinar els patrons de comportament alimentari i la seva evolució circadiària i al llarg del temps. També es determinà l'establiment de l'ordre d'accés al dispensador. En general, l'allotjament de les truges en grup amb FITMIX incrementà el descans i disminuí les estereotípies oronasofacials en major mesura que l'allotjament en grup amb TRICK. Ambdós sistemes d'allotjament en grup reduïren considerablement la masticació en buit. Les primípares mostraren uns nivells d'activitat general i d'estereotípies menors que les truges adultes en tots els sistemes. Els nivells d'Hp i de pig-MAP, i les mesures productives no diferiren entre sistemes. En canvi, les incidències com coixeses o ferides de vulva foren més freqüents en els sistemes d'allotjament en grup. La freqüència, la intensitat i la proporció de resolució de les agressions foren més altes al sistema FITMIX, tot i que el contacte físic intens fou inusual en ambdós sistemes. Els conflictes al sistema FITMIX foren principalment per menjar, mentre que al sistema TRICK es produïren principalment a la zona de descans. Tanmateix, la proporció d'agressions per beure fou major al sistema TRICK. La correlació entre l'RI i el pes viu fou major al sistema FITMIX, tot i que la correlació entre l'RI i el número de part fou similar en ambdós sistemes. Les primípares participaren menys en agressions, tot i que en reberen més que n'iniciaren en ambdós sistemes. El 46% de les truges del sistema FITMIX, majoritàriament primípares i subordinades, necessitaren ajuda per adaptar-se al sistema d'alimentació. Finalment, el 8,3% de les truges del sistema FITMIX no s'adaptaren i foren eliminades.Les truges del sistema FITMIX feren diverses visites al dispensador per obtenir la seva ració diària, tot i que l'ocupació diària del dispensador fou menor que la citada a la literatura referent als sistemes d'alimentació electrònica amb protecció (ESF). L'ocupació diària disminuí al llarg del temps. Dins de cada cicle alimentari, la màxima activitat al voltant del dispensador s'observà a les hores immediates al seu inici, de manera que l'activitat fou molt baixa en la segona meitat del cicle. Un ordre d'accés al dispensador relativament estable i correlacionat amb l'RI s'establí ràpidament i es mantingué durant l'experiment. Les truges dominants accediren al dispensador més aviat i feren un número de visites similar a les truges subordinades però de major duració i per tant, ocuparen el dispensador durant més temps. En conclusió, els sistemes d'allotjament en grup augmentaren el descans i disminuïren les estereotípies mantenint uns nivells de productivitat i de dany tissular similars als de les truges en gàbies. El sistema FITMIX presentà major competitivitat que el sistema TRICK a causa de l'alimentació seqüencial. El sistema FITMIX fou eficient en grups estables de mida mitjana, tot i que l'optimització de l'eficiència no s'assolí fins al cap d'uns dies. D'altra banda, l'establiment d'un ordre d'accés al dispensador correlacionat amb l'ordre de dominància podria disminuir el conflicte per menjar i ajudar al personal a detectar problemes. Tenint en compte tots els factors, l'allotjament en grup podria ser una bona alternativa a les gàbies sempre que el maneig sigui l'adequat. Un personal més qualificat podria ser necessari per la detecció precoç i la solució dels problemes, sobretot per prevenir l'agressivitat excessiva i assegurar l'adaptació en sistemes d'alimentació seqüencial. Malgrat tot, els efectes a llarg termini dels sistemes d'allotjament en grup sobre el benestar haurien de ser avaluats prèviament a llur recomanació. / Las nuevas directivas de la UE sobre la protección de los cerdos (Sus scrofa) hacen necesario el desarrollo de sistemas de alojamiento en grupo apropiados desde el punto de vista del bienestar y la productividad. El objetivo general de este estudio fue la comparación entre dos sistemas comerciales de alojamiento en grupo y las jaulas convencionales para cerdas gestantes, mediante medidas de comportamiento, fisiológicas y productivas. Ciento ochenta cerdas gestantes, de primer a noveno parto, se seleccionaron en una granja comercial y se repartieron en tres réplicas diferentes (60 cerdas por réplica). Las cerdas se alojaron desde el día 29 de gestación a una semana antes del parto en jaulas convencionales (STALL), en grupos de 10 con un sistema de alimentación de caída lenta (TRICK) y en grupos de 20 con un dispensador de pienso electrónico sin protección (FITMIX; 20 cerdas por sistema de alimentación y réplica). Todas las cerdas habían estado previamente alojadas en jaulas y se alimentaron de forma igualmente restringida. Las cerdas se observaron 11 días no consecutivos durante 4 h al día. En todos los sistemas, los niveles de actividad general y de estereotipias se midieron mediante la técnica de observación scan-sampling a intervalos de 10 minutos. Se obtuvieron dos muestras de sangre por réplica y se determinaron los niveles de haptoglobina (Hp) y pig-MAP para valorar el grado de estrés físico. El espesor de tocino dorsal, el peso vivo y los resultados reproductivos también fueron considerados. En los dos sistemas de alojamiento en grupo, se registraron las agresiones mediante observación continua y posteriormente se calculó el orden de dominancia (RI). En el sistema FITMIX, se analizaron los datos de comportamiento alimentario registrados automáticamente durante 25 ciclos alimentarios no consecutivos. El número de visitas al dispensador, la duración de las visitas y la ocupación del dispensador se analizaron por día y por hora para determinar los patrones de comportamiento alimentario y su evolución circadiana y a lo largo del tiempo. También se determinó el establecimiento del orden de acceso al dispensador. En general, el alojamiento de las cerdas en grupo con FITMIX incrementó el descanso y disminuyó las estereotipias oronasofaciales en mayor medida que el alojamiento en grupo con TRICK. Ambos sistemas de alojamiento en grupo redujeron considerablemente la masticación en vacío. Las primíparas mostraron unos niveles de actividad general y de estereotipias menores que las cerdas adultas en todos los sistemas. Los niveles de Hp y de pig-MAP, y las medidas productivas no difirieron entre sistemas. En cambio, las incidencias como cojeras o heridas de vulva fueron más frecuentes en los sistemas de alojamiento en grupo.La frecuencia, la intensidad y la proporción de resolución de las agresiones fueron mayores en el sistema FITMIX, aunque el contacto físico intenso fue inusual en ambos sistemas. Los conflictos en el sistema FITMIX fueron principalmente por comer, mientras que en el sistema TRICK se produjeron principalmente en la zona de descanso. Asimismo, la proporción de agresiones en el bebedero fue mayor en el sistema TRICK. La correlación entre el RI y el peso vivo fue mayor en el sistema FITMIX, a pesar de que la correlación entre el RI y el número de parto fue similar en ambos sistemas. Las primíparas participaron menos en agresiones, aunque recibieron más que las que iniciaron en ambos sistemas. El 46% de las cerdas del sistema FITMIX, mayoritariamente primíparas y subordinadas, necesitaron ayuda para adaptarse al sistema de alimentación. Finalmente, el 8,3% de las cerdas del sistema FITMIX no se adaptaron y fueron eliminadas.Las cerdas del sistema FITMIX hicieron diversas visitas al dispensador para obtener su ración diaria, aunque la ocupación diaria del dispensador fue menor que la citada en la literatura referente a los sistemas de alimentación electrónica con protección (ESF). La ocupación diaria disminuyó a lo largo del tiempo. En cada ciclo alimentario, la máxima actividad alrededor del dispensador se observó en las horas inmediatas a su inicio, de manera que la actividad fue muy baja en la segunda mitad del ciclo. Un orden de acceso al dispensador relativamente estable y correlacionado con el RI se estableció rápidamente y se mantuvo durante el experimento. Las cerdas dominantes accedieron al dispensador más temprano e hicieron un número de visitas similar a las cerdas subordinadas pero de mayor duración, por lo que ocuparon el dispensador durante más tiempo.En conclusión, los sistemas de alojamiento en grupo aumentaron el descanso y disminuyeron las estereotipias manteniendo unos niveles de productividad y de lesión tisular similares a los de las cerdas en jaulas. El sistema FITMIX presentó mayor competitividad que el sistema TRICK a causa de la alimentación secuencial. El sistema FITMIX se mostró eficiente en grupos estables de tamaño medio, aunque la optimización de la eficiencia no se alcanzó hasta al cabo de unos días. Por otra parte, el establecimiento de un orden de acceso al dispensador correlacionado con el RI podría disminuir el conflicto por comer y ayudar al personal a detectar problemas. Teniendo en cuenta todos los factores, el alojamiento en grupo podría ser una buena alternativa a las jaulas siempre que el manejo sea el adecuado. Un personal más cualificado podría ser necesario para la detección precoz y la solución de los problemas, sobre todo para prevenir la agresividad excesiva y asegurar la adaptación en sistemas de alimentación secuencial. No obstante, los efectos a largo plazo de los sistemas de alojamiento en grupo sobre el bienestar deberían ser evaluados previamente a su recomendación. / The new EU directives on the protection of pigs (Sus scrofa) make necessary the development of appropriate group housing systems on a welfare and productivity basis. The general objective of this study was the comparison among two different commercial group housing systems and conventional stalls for pregnant sows by assessing behavioural, physiological and performance measures. One hundred and eighty pregnant sows, from first to ninth parity, were selected on a commercial farm and used in three different replicas (60 sows per replica). Sows were housed from day 29 of pregnancy to 1 week before parturition in conventional stalls (STALL), in groups of 10 with trickle feeding (TRICK) and in groups of 20 with an unprotected electronic sow feeder (FITMIX; 20 sows per housing system per replica). All the sows had been previously stall-housed and were equally feed-restricted. Sows were observed on 11 non-consecutive days during 4 h a day. General activity and stereotypies were measured by scan-sampling observation (10-min intervals) in all the systems. Blood samples were obtained twice a replica and levels of haptoglobin (Hp) and pig-MAP were determined to assess physical stress. Backfat depth, liveweight and reproductive performance were also considered. In the two group housing systems, aggressions were recorded by continuous behaviour sampling and a dominance rank (RI) was thereafter calculated. In FITMIX, data on feeding behaviour recorded by the system from 25 non-consecutive feeding cycles were analyzed. Number of visits to the feeder, duration of the visits and feeder occupation was analyzed on a daily and hourly basis to determine over-time and circadian feeding patterns. The establishment of a feeding order was also assessed. Overall, group housing sows with FITMIX appeared to increase resting behaviour and decrease oronasofacial stereotypies to a higher extent than with TRICK. Both group housing systems reduced considerably the performance of sham-chewing. Gilts showed a lower general activity and performance of stereotypies than older sows in all the systems. In general, Hp and pig-MAP levels and productivity measures did not differ among systems. However, incidences such as lameness or vulva injuries were more often detected in group housing systems. Frequency of aggressions, intensity and proportion of resolution were higher in FITMIX, yet intense physical contact was unusual in both systems. Conflicts in FITMIX were largely for feeding whereas in TRICK, they occurred mainly in the resting area. However, proportion of aggressions for drinking was higher in TRICK. Correlation between RI and liveweight was stronger in FITMIX, although correlation between RI and parity was similar in the two systems. Gilts got less involved in aggressions, although they received more aggressions than they initiated in both group housing systems. Forty-six per cent of the FITMIX sows, mostly gilts and subordinates, needed assistance to adapt to the feeding system. Eventually, 8.3% of the FITMIX sows failed to adapt and had to be removed.FITMIX sows made several visits to the feeder to get their daily ration, although daily feeder occupation seemed to be shorter compared to literature on protected ESF. Daily feeder occupation per sow appeared to decrease over time. Within each feeding cycle, maximum activity around the feeder was observed in the hours following the start of the cycle, so that activity was very low in the latter half of the cycle. A relatively stable feeding order was quickly established and maintained over the experiment and it highly correlated with RI. High-ranking sows fed earlier and made as many visits as low-ranking sows but longer and thus, they occupied the feeder for longer.In conclusion, well-managed group-housed sows seemed to increase resting behaviour and decrease stereotypies at similar productivity and tissue damage levels than conventional stalls. Sequential feeding appeared to make FITMIX a more competitive feeding system than TRICK. In terms of efficiency, FITMIX appeared to be a good system for medium-sized stable groups of sow, although optimization of the feeder efficiency may take several weeks. However, the establishment of a rank-related feeding order may ameliorate disturbance around the feeder and help stockmanship to detect problems. Taking everything into account, it can be suggested that group housing systems can be a good alternative to stalls provided that management is appropriate. Higher quality stockmanpeople may be required for early detection and solution of problems, particularly to prevent excessive aggression and ensure adaptation in sequential feeding systems. Nevertheless, long-term effects of group housing systems on welfare should be further assessed before recommendations are made.

Obesity is a growing public health issue with few effective treatments. Many of the human genes involved in weight regulation have homologues in other species. In Drosophila melanogaster, 11-cis-vaccenyl acetate (cVA) is a pheromone shown to affect both aggression and feeding behaviour. The aim of this study was to investigate the effect of cVA exposure on gene expression in wild type flies and to confirm crosses with genetic knockouts in the cVA pathway. The genes studied in the wild type flies were Akh (homologue of glucagon), Dilp2,-3 and -5 (homologues of insulin), Dsk (homologue of cholecystokinin), sNPF (homologue of neuropeptide Y) and TβH. The knockout genes studied were Or65a, Or67d and TβH. RNA was extracted from whole heads, cDNA synthesis was performed and the cDNA was then used in SYBR Green RT-qPCR. The knockout genes were not confirmed, due to methodological problems. The expression of Dilp3, Dilp5 and Dsk were significantly lower in the experimental flies, the expression of the other genes where not affected. The results indicate that exposure to cVA affects the expression of some of the neuropeptides involved in weight regulation. Due to the methodological problems experienced, a recommendation is to confirm the data using different reagents.

Two trials were conducted to evaluate effects of grain adaptation protocol on subacute acidosis, feeding behaviour and ADG of cattle. In trial one, 12 crossbred heifers (384 ¡À 25 kg) were stepped from 40% to 90% dietary concentrate by either rapid adaptation (RA; one step-up diet fed for 3 d) or by gradual adaptation (GA; five step-up diets fed for 3 d each). Mean daily ruminal pH variables did not differ (P > 0.10) between treatments but variances of a number of ruminal pH variables were greater (P < 0.05) for RA than GA during adaptation to 65% and 90% concentrate. Mean hourly pH did not differ over the first 24 h of adaptation to 65% concentrate, but variance of hourly pH tended (P < 0.10) to be greater for RA than GA for eight of the first 24 h of adaptation to 90% concentrate. Increased variance in ruminal pH parameters was associated with detection of acidosis in certain individuals. On d 1 of 90% concentrate, ruminal pH tended (P = 0.07) to be lower at 11 and 12 h post-feeding with RA than with GA. Ruminal volatile fatty acids (VFA) and osmolality were similar between treatments. In trial two, 120 crossbred heifers (366 ¡À 23 kg) were adapted from 40% to 90% concentrate. A protocol identical to trial one was used with the addition of moderate adaptation (MA; three step-up diets fed for 3 d each). The increase to 65% concentrate caused reduced daily bunk attendance and increased maximum intermeal interval for RA compared to MA and GA cattle but the increase to 90% did not. ADG was reduced for RA compared to MA or GA during adaptation but over day 1 to 69 ADG did not differ between treatments (P ¡Ý 0.41). Current management strategies for preventing acidosis in pens of cattle are based on responses of the most susceptible individuals. Improved understanding of individual responses to acidotic challenge may allow development of more effective acidosis prevention practices.

The objective of this dissertation was to determine the feeding motivation of dairy heifers and to determine management factors that contribute to satiety in limit-fed dairy heifers. The first study of this dissertation determined the difference in behavioural patterns between limit-fed and ad libitum-fed dairy heifers so that we may understand how limit feeding diverges from normal behaviour. Limit-fed heifers were found to decrease time spent feeding, consume less DM, and increase their feeding rate across the day in comparison to ad libitum-fed heifers. Additionally, the diurnal patterns of ad libitum-fed heifers were much closer to normal behaviour as they consisted of several short, small meals distributed throughout the day. The next three studies examined the management factors of frequency of feed provision and feed bunk space and how these may impact the behaviour of heifers. Increased frequency of feed provision did result in more distribution of feeding activity throughout the day but also imposed feeding amounts that were likely insufficient to achieve satiety. Heifers fed once daily likely experienced satiety in the short term and did spend the most time feeding throughout the day but the diurnal patterns of these heifers did not allow for expression of normal foraging behaviour. Providing increased feed bunk space did not allow limit-fed heifers to spend more time feeding or impact competition within a pen whereas providing straw alongside of a nutrient-dense TMR did result in more normal diurnal patterns, albeit with an increase in competitive behaviour. There was no interaction found between feed bunk space and frequency of feed provision, indicating that limit-fed heifers must be provided with sufficient bunk space to feed simultaneously. Heifers provided with un-restricted bunk space did gain more and were more feed efficient and less variable in feeding time than heifers given restricted bunk space. Feeding once daily resulted in an increase in competition but also enabled heifers to gain well and to spend more time feeding each day. The remaining studies sought to quantify whether a preference for supplementary feed exists in limit-fed heifers and whether heifers will work, and to what extent, for this extra feed. The results indicated that heifers will consume similar amounts of supplementary long or short straw if provided to them alongside of a limit-fed TMR. The limit-fed heifers do, however, show a clear preference for long straw when offered the choice, suggesting that they find long straw to be more satisfactory for achieving rumen fill and/or meeting their behavioural foraging needs. Heifers were also found to work harder (push more weight), spend less time feeding and ruminating, and consume feed faster when provided a high-concentrate, limit-fed ration than when provided a high-forage, ad libitum-fed ration, suggesting that these animals are experiencing feelings of hunger and may not be physically or behaviourally satisfied. These findings indicate that there are clear behavioural differences between limit-fed and ad libitum-fed dairy heifers. Management factors, such as frequency of feed provision and feed bunk space, may be altered in such a way as to provide some benefit to the animals but are limited in their ability to normalize feeding behaviour and diurnal patterns. Provision of supplemental, long particle low-nutritive roughage (i.e. straw) aids much more in allowing limit-fed heifers to achieve satiety and is considered a desirable resource by the animals, as evidenced through their willingness to work for this supplemental feed. / Ontario Ministry of Agriculture, Food and Rural Affairs (OMAFRA)/University of Guelph Production Systems research grant and a Campbell Centre for the Study of Animal Welfare research grant. This project was also supported through contributions from the Canadian Foundation for Innovation (CFI) and the Ontario Research Fund. Angela Greter was supported through NSERC Alexander Graham Bell Canadian Graduate Student scholarship

The objective of this thesis was to determine the effects of frequency of milking and feed delivery on the behavioural patterns and productivity of lactating dairy cows. In two independent experiments, twelve free-stall housed, lactating Holstein dairy cows were exposed to either varying milking frequency or varying feed delivery frequency over 21-d periods. In the first study, cows milked three times per day and multiparous cows produced more milk than those milked twice per day and primiparous cows. Milking three times per day altered the distribution of feeding activity throughout the day. Multiparous cows had longer, and larger meals, while primiparous cows had smaller, more frequent meals throughout the day when milked three times per day. In the second study, cows delivered feed three times per day consumed more DM than those fed once per day and twice per day. Feed delivery frequency had little effect on feeding behaviour and the distribution of feeding activity, but altered the magnitude of DMI following feed delivery. Cows delivered feed more frequently achieve greater daily DMI by consuming more feed following the return from milking and the delivery of feed. / This project was financially supported by a Dairy Farmers of Ontario (Mississauga, Ontario, Canada) research grant and an Ontario Ministry of Agriculture, Food and Rural Affairs (OMAFRA; Guelph, Ontario, Canada)/University of Guelph Production Systems research grant. Additional project support was received from the Canadian Foundation for Innovation (CFI; Ottawa, Ontario, Canada) and the Ontario Research Fund (Toronto, Ontario, Canada).

ABSTRACT Collembola are known to feed on soil fungi, mycorrhizae and plant derived products. A recent study revealed that one species of Collembola, Protaphorura fimata, completely switched from decomposer to herbivore when live roots were present. The current study investigated the occurrence of diet-switching in Folsomia candida Willem. from plant detritus to live by examining its dietary preferences using stable isotope techniques. They were offering with live maize roots (C4 plant) in C3 soil, along with 15N enriched ryegrass litter and. Results demonstrated the presence of a partial diet-switch from detritus to live maize roots. Additional tests suggested that the diet-switch towards maize roots was a response to both improved food quality and greater food availability. The presence of live roots improved the body growth of F. candida and the incorporation of C from live roots into Collembola tissue suggesting

Animal survival depends on the brain’s ability to detect the energetic state of the body and to alter behaviour in order to maintain homeostasis. Current research in the control of food consumption stresses the importance of identifying and establishing the specific roles of homeostatic neurons, which sense the body’s energetic state and elicit complex and flexible food seeking behaviours. Recent developments in optogenetics, molecular genetics, and anatomical techniques have made these investigations possible at the resolution of specific cell types and circuits. These neurons are of particular interest because they serve as key entry points to the identification of downstream circuits and reinforcement mechanisms that control feeding behaviour. This dissertation probes the role of two kinds of homeostatic neurons— agouti-related peptide (AGRP) in the arcuate nucleus (ARC) and leptin receptor (LepRb) neurons in the lateral hypothalamic area (LHA)—in the control of food intake. First, I examined the role of LepRb neurons in the LHA in feeding. Results from electrophysiological studies indicate that these neurons consist of a subpopulation of homeostatic sensing LHA γ-aminobutyric acid (GABA) expressing neurons. In addition to their response to leptin, these neurons are capable of modulating their activity in response to changes in glucose levels, further substantiating their role as homeostatic sensing neurons. Behavioural studies using optogenetic activation of these neurons show that their elevated activity is capable of reducing body weight, although their role in modulating feeding remains unclear. Second, I investigated the reinforcement mechanisms employed by AGRP neurons to elicit voracious food consumption and increased willingness to work for food. Conditioned place avoidance studies under optogenetic activation of AGRP neurons reveal that their increased activity has negative valence and is avoided. In addition, imposition of elevated AGRP neuron activity in an operant task reduced instrumental food seeking with particular sensitivity under high effort requirements. Taken together, these results suggest that AGRP neurons employ a negative reinforcement teaching signal to direct action selection during food seeking and consumption. Third, I systematically analyzed the contribution of specific AGRP neuron projection subpopulations in AGRP neuron mediated evoked-feeding behaviour. Optogenetic activation studies of AGRP neuron axons in downstream projection regions indicate that several, but not all, subpopulations are capable of independently evoke food consumption. This work reveals a parallel and redundant functional circuit organization for AGRP neurons in the control of food intake. Interestingly, all AGRP neuron subpopulations examined displayed similar modulation by states of energy deficit and signals of starvation, despite their apparent divergence in function. As a whole, this dissertation extends our understanding of the role of homeostatic neurons in food consumption and uncovers previously unappreciated functional organization and reinforcement mechanisms employed by neuronal circuits that control feeding behaviour.