The internet has been used for data transfer in radio astronomy ever since its inception; however it is only recently that network bandwidth capability means that the internet becomes competitive with traditional forms of data storage. Very Long Baseline Interferometry (VLBI) uses widely separated telescopes between which high bandwidth direct connections have not been feasible until recently. The academic networks now allow us to connect at high data rates (~1Gbps) in "eVLBI". The ESLEA project (Exploitation of Switched Lightpaths for E-science Applications) has played a major role in the development of eVLBI. We outline this development in this paper.

We present results from the first successful open call e-VLBI science run, observing the X-ray binary GRS 1915+105. e-VLBI science allows the rapid production of VLBI radio maps, within hours of an observation rather than weeks, facilitating a decision for follow-up observations. A total of 6 telescopes observing at 5 GHz across the European VLBI Network (EVN) were correlated in real time at the Joint Institute for VLBI in Europe (JIVE). Constant data rates of 128 Mbps were transferred from each telescope, giving 4 TB of raw sampled data over the 12 hours of the whole experiment. Throughout this, GRS 1915+105 was observed for a total of 5.5 hours, producing 2.8 GB of visibilities of correlated data. A weak flare occurred during our observations, and we detected a slightly resolved component of 2.7 x 1.2 milliarcsecond with a position angle of 140 (+/-2) degrees. The peak brightness was 10.2 mJy per beam, with a total integrated radio flux of 11.1 mJy.

As part of an on-going radio supernova monitoring program, we have discovered a variable, compact steep spectrum radio source ~65 arcsec (~4 kpc) from the centre of the starburst galaxy NGC 3310. If the source is at the distance of NGC 3310, then its 5 GHz luminosity is ~3 x 10^{19} WHz^-1. The source luminosity, together with its variability characteristics, compact structure (<17 mas) and its association with a group of HII regions, leads us to propose that it is a previously uncatalogued type II radio supernova. A search of archival data also shows an associated X-ray source with a luminosity similar to known radio supernova.