Mobile Ad hoc Network (MANET) is a collection of mobile nodes that are connected wirelessly forming random topology through decentralised administration. In MANETs, multicasting is an important mechanism which can increase the network efficiency and reliability by sending multiple copies to a group of nodes in a single transmission without using several unicast transmissions. Multicast routing can be classified into tree based and mesh based multicasting. Mesh based protocols offer path redundancy for packets to move from senders to receivers. Thus, they offer greater resilience to link breakages than tree based protocols which offer only a single path from senders to receivers. Mesh based protocols have higher packet delivery ratios on the cost of higher overhead due to maintaining redundant paths. Receiver initiated mesh based multicast routing is the improved version having reliability and reduced overhead because the receivers only report the missing packets to the sender as compared to the sender initiated approach where each successful packet is reported to the sender. Receiver initiated mesh based multicast routing strongly relies on proper selection of a core node. The existing schemes suffer from various problems. First, the core selection process is not sophisticated that usually selects core in a manner that may decrease core lifetime and deteriorate network performance in the form of frequent core failures. Second, the existing schemes cause too much delay for core re-selection(s) process. The performance becomes worse in situations where frequent core failures occur and hence, the protocols may become unsuitable for delay sensitive applications. Finally, a malicious node may try to illegitimately become selected as core node (for some malicious purposes) or selfishly evade the core election process to save its resources. To solve the above issues, we propose an efficient, reliable and secure core assisted multicast Adhoc routing protocol (ERASCA) in which an efficient/stable core is selected based on parameters like battery capacity and location in a group. The selection of a stable core within the group minimizes the core failure scenario, thereby minimizes the flooding for finding another core and will decrease the overhead. To minimize delay and data collection process during core election after failures, we introduce the concept of mirror core in a group; hence, after the failure of the core the mirror core will take the responsibility as the main core without data collection process and delay and hence increases the overall reliability. To prevent the malicious/selfish nodes from illegitimately become core or evade the process of being a core node for saving resources, a iv malicious/selfish receiver is detected and discarded within the mesh by estimating their battery capacity. To collect the data for the estimation of battery capacity, we propose an overhearing based approach. To further protect the data from malicious/selfish receiver and to preserve the integrity of data, a packet authentication process is used. The proposed protocol is evaluated in NS-2 and compared against PUMA and MAODV which are the state of the art protocols in mesh based and tree based multicasting respectively. We compare these protocols under different metrics, such as mobility, number of senders, number of receivers, interface queue length and simulation area. According to the simulation results, ERASCA and ERASCA-MC attains higher packet delivery ratios and throughput than PUMA and MAODV, while incurring far less overhead, delay and energy consumption because of the efficient core election and with the introduction of the mirror core. To secure the core election process from malicious and selfish receivers, multiple simulations in NS-2 are performed in the presence and absence of detection technique. Simulation proves that in the presence of detection technique, core election will be more secured with improved performance in PDF, overhead, throughput and energy utilization.