Copyright © 1997, Neal McEwen
In the early days for landline telegraphy, the oceans presented a barrier
to linking the continents. Before there were cables that went under the
oceans, a telegraph could still be sent from Berlin to Chicago for example.
From Berlin it went over landline wires to a telegraph 
office close to a port. The telegram would be given to a ship heading for
New York City for example. When the ship reached New York, the telegram
was given to a local telegraph office and the telegram was relayed to Chicago
for delivery. Each leg of the trip was quite speedy except for the long
sea voyage.
The first submarine cable between Europe and North American was laid by Cyrus Field in 1858. The cable was in service only 23 days and only 21,000 characters were sent between Newfoundland and Valentia, Ireland, before it failed. It was not until after the American Civil War that another attempt was made. In 1865 a second cable broke while being laid in deep water. In 1866 a third cable was successfully laid and put into service.
What was the technology like that supported instantaneous news and commerce
between continents? Landline telegraph lines during this period were only
a few hundred miles long. To transmit a message further required manual
or electromechanical relay of the message. This was not possible in spanning
the thousand plus miles between continents. The current at the far end
of a cable was very, very weak. The capacitance effect between the cable
and the earth prevented normal keying speeds. Telegraph engineers and physicists
of the mid 19th century called this the "retardation effect." Only very
slow sending could be accommodated because of the long rise and fall time
due to the capacitance. 
The rigorous environment imposed by extremely long distances required new technology. There was not enough current in the early cables to pull the electromagnets of a telegraph sounder or relay. To copy the weak signals a mirror galvonometer was used on the receiving end. Galvonometers are very sensitive to weak currents and hence the movement of a galvo could be detected. Mirrors were added to the movement of the galvo and a light source placed in front. In a dark room, the operator could follow the swing of the mirror by watching the light projected by the mirror on a wall or a scale a few feet away. The pieces of a mirror galvonometer system are shown in the top image. The galvo itself is on the right, and a scale on the left. A lamp and shade, for focusing a beam is in the middle.
On the sending end, keys with two pole were used to make the dots and dashes. The keys had two levers. Closing one lever put a positive voltage on the wire and closing the other lever put a negative voltage on the wire. The receiving operator watched the galvo swing left and right.
The mirror system had two flaws. First it took two men to operate it,
one to watch the beam of light and call out the letters while a second
operator wrote the letters down. Second, there was no permanent record
of the message received. These limitations were overcome by the invention
of the siphon recorder. This device uses a siphon of ink which is moved
transversely across a thin paper tape which is also moving. The movement
of the siphon is controlled by 
currents in coils, upon which the siphon is attached to. A sample of siphon
recorder tape is shown. Note the letter "t" is a single dip down, followed
by three dips down for "o." Notice how "h" swings up for four dips and
how "p" is one up, two down and one up. Observe how the "retardation" has
smoothed the trace. On a short telegraph wire, the dots and dashes look
like a square wave. The Continental code is used. American Morse with the
dot and space letters would have made reading the galvo or the tape more
difficult that it needed to be. A siphon recorder is shown below. 
By the 1890s, speeds of 20 words per minute could be obtained, still not nearly as fast as landline circuits. Over the years, the technology was improved and speeds got higher and higher. To improve the throughput of cable stations automated sending and receiving equipment was used. Paper tapes were punched ahead of time, then feed to a device that read the tape and sent the message. On the receiving end, the signal was recorded to paper tape for later transcription by an operator.
As the end of the 19th century, there were 12 cables spanning the North Atlantic and a 160,000 nautical miles of submarine cable throughout the world. There were no cables across the Pacific. Links to Australian and New Zealand came via Asia.
The successful laying of the first tans-Atlantic cable was probably the most significant engineering feat of the 19th century. Because of this and the effect it had on inter-continental communications, the laying of the cable by Field is well documented. There were many detailed accounts published and these books are quite collectable.
Thought not as plentiful as landline instruments, cable instruments can be found. The most popular artifact among collectors is the double lever sending keys. Quite available are pieces of various cables that have been taken up and certified; souvenir pieces of cable have been popular since the late 19th century. There also are many wall maps available showing the routes of the worldwide cables. Some collectors have pieces of siphon recorder tape and punched tape from the automatic sending equipment.
I have only scratched the surface of cable history and technology. It is a fascinating subject. You local library will no doubt have more on the subject.
If you have any questions about any aspect of telegraphy don't hesitate to ask. I check my E-mail at least once a day and will be happy to answer any question or steer you to another source.