One of the most important but least-studied aspects of Internet policy is the emergence of a trading market for previously allocated Internet number blocks. Without unique Internet protocol numbers for the networks and devices attached, the Internet simply doesn’t work. The original Internet Protocol standard, known as IPv4, specified a 32-bit numbering space, which provided slightly less than 4 billion unique numbers that could be used as addresses (Postel, 1981). A large part of that number space has already been handed out to organizations. The available supply is dwindling, and the Asia-Pacific region is already reduced to rationing its last /8 in tiny, 1024-address chunks, one to an organization, while the European region is only a few months from that status. Recent market developments and policy changes by Internet number registries now allow organizations with more numbers than they want to sell them to another organization. In other words, a market for IPv4 numbers is now possible. This is a major change in the political economy of Internet governance. It is likely that the commercial forces unleashed will have far-reaching consequences for Internet businesses, users and governance institutions. In a related area, domain name registration, the emergence of a commercial market led to the growth of a domain name industry and major changes in policies and institutions, such as the formation of ICANN, the separation of registries and registrars and the new gTLD program. After a highly publicized deal in which Microsoft bought Nortel’s number assets in a bankruptcy proceeding, the reality of an IP number market can no longer be denied. But many in the Internet technical community still feel uncomfortable about it. One reporter with attitudes typical of the technical community, has predicted that “a functioning market won't form at all, or will break down very quickly after it forms" (van Beijnum, 2011). This is partly due to an ideological resistance to the commodification of a critical Internet resource, and partly due to their fears that an IPv4 market might delay or even prevent a migration to a new Internet protocol (IPv6). Either way, few wish to openly acknowledge the market’s existence. Thus the topic of number markets brings to mind the phrase “the elephant in the room.” Moreover, the information we have about this elephant is fragmented and unsystematic. IP number allocation is controlled on a contractual basis by five separate regional Internet registries (RIRs). Each has different policies toward transfers, different registry databases, and different disclosure practices. It is difficult, therefore, to obtain a comprehensive picture of the emerging market for IPv4 number blocks. This recalls the old fable about the five blind men and the elephant, with each one having access to a small part of the body and none of them quite grasping the nature of the beast as a whole. Based on this double-elephant metaphor, this paper tries to make a less than elephantine but much-needed empirical contribution to the literature on the economics and institutions of IP addressing. While several papers already discuss IP number markets in theoretical or policy terms, no one has actually compiled and analyzed the transactions themselves. This paper draws on the records of the RIRs to compile as much information about traded IP number blocks as possible, and then conducts some very basic analysis of stocks, flows and proportions to assess the nature of this emerging market and explore some of its implications for Internet governance. Although it is early days for IP number markets, the data we have gathered already seem to have interesting implications for certain policy issues currently being debated.