The cable connection between Banjoewangi and Port Darwin was always tenuous at best. Breakdowns were too frequent for a connection which was so critically important. Given the scientific understanding of the time, the nature of the breakdowns could not be explained reliably. Indeed it would not be until the late 1950s that the implications of plate tectonics would explain the real causes.
It was therefore thought that moving the cable a few hundred miles would overcome the problems of the breakages. The Eastern Extension Telegraph Company therefore decided to lay a third cable from Java to Australia - with a landing at Roebuck Bay in Western Australia (now Broome). That cable could then link with the Western Australian lines from the far north to Perth and then with the inter-colonial lines via Eucla to Adelaide and from there to the other Colonies.
After only four years of operation, the Sydney Evening News of 4 November 1876 reported:
Mr. Cracknell and the Duplicate Cable.
From Melbourne we have particulars of the contents of Mr. Cracknell's letter to the New South Wales Government, enclosing an offer for a duplicate cable from a company, and strongly recommending its immediate acceptance. The Eastern Extension Company is offering to duplicate the cable from Singapore to Banjoewangi for £21,780 per annum and from Banjowangie to North-west Cape, Western Australia, for £23,220.
Mr. Cracknell recommends that the cable from Singapore to Banjoewangie be arranged for at once, the cost of which to the colonies would be £14,520 for interest and £7,260 for renewal .
The cable from Banjoewangie to North-west Cape could follow if necessary.
It was not improbable, in Mr. Cracknell's opinion, that the Imperial Government would assist in the subsidies for the new cabIe from Singapore to Java, and the reduction of the present telegraphic charges to the colonies. If this were carried out, there would be a duplicate line the whole way from London to Banjoewangie, as a new cable was almost completed from Penang and Rangoon and duplicate cables were being shipped for the Red Sea and Bombay sections.
The Eastern Extension announcement was made in August 1888. On 5 September 1888, the Perth Inquirer published the following note from their London Correspondent:
"In my last letter I mentioned that the origin of the simultaneous rupture of the telegraphic cables was probably due to volcanic activity and, in consequence of the conversation I had with the Secretary of the Eastern Extension Telegraph Company, I called at their offices yesterday to learn what further information he might have for me as to this point.
I learned that after the Banjoewangie end of the broken cable had been safely grappled at a depth of 900 fathoms and buoyed for reattachment, the steamer proceeded to take up the Port Darwin portion. This, however, broke from the tension in raising and was lost so that a new cable had to be supplied and the broken fragment recovered. When the two ends are compared, and not till then, experts will be able to form their opinions as to the cause of the accident.
While on this subject I may inform you that the negotiations between the Telegraph Company, the Colonial Office and your Executive are rapidly proceeding. The idea at first appears to have been to ask for the landing rights at Beagle Bay, but inasmuch as the land line seems only to have been carried as far north as Roebuck Bay, the point arises that either the 150 miles between that point and Beagle Bay would have to be completed at once, or the landing take place at the nearest available spot to Roebuck Bay. The point is, of course, not one of great importance, as certainly it is to be hoped that no long time will be allowed to elapse before the entire circuit is completed as far as Port Darwin; so that, in case of accident, an alternative route of intercolonial communication may remain open.
The proposal of the Eastern Extension Telegraph Company has been so well received by the authorities that they have felt themselves justified in placing the order for the new cable to be laid to your coast, and the same, which will be nearly one thousand miles in length — about 950 miles I believe to be particular — is already in course of manufacture".
The boat appointed for that purpose will, immediately after recovering the lost portion of the now re-united cable and repairing the second, proceed to take soundings for the third, which, it is anticipated, will be about 2000 fathoms.
I must here interpolate a remark with respect to the projected Indo-Australian cable of Messrs. Millar Brothers to the effect that the recent annexation of Christmas Island in the Indian Ocean by Captain May of H.M.S. Impérieuse, was made in view of its being carried into effect. Now, however, I suppose the matter will probably lapse".
In October 1888, a survey vessel was taking soundings along the proposed route which kept to the east of a direct line so as to avoid a very deep depression between Java and Australia.
A suitable landing place on the Western Australian coast was difficult to identify. The first preference of the Western Australian Government was for the cable to be brought ashore at Derby in King Sound. As noted elsewhere, it was intended to complete construction of the the land line to Derby in 1889. There were however insuperable difficulties at the entrance to King Sound and so the constructors selected Beagle Bay as the most suitable spot at which to take the cable ashore. Unfortunately the nature of the country inland from Beagle Bay proved almost impassable - even the line to Derby was to pass south of that point - so Roebuck Bay, an inlet still further south, was selected.
Telegraph lines had been constructed from Perth to Broome so there was a direct telegraphic connection to the south and then across the new wires to Adelaide and the Eastern States.
The Telegraph & Maintenance Company's S.S. Seine left London on 31 December 1888 to the Java with the Western Australian cable. The cable was of the most improved design. and was 970 nautical miles long. It was expected to be installed by the end of February 1889 (Ballarat Star 4 Jan 18 p2).
The cost of the cable was about £120,000.