Engines : H and K-Series prototypes

The A-Series may have been almost immortal by the early 1970s, but it did not stop the company developing alternatives. Here we describe the H and K-Series prototypes which were designed for the ill-fated Austin ADO74 supermini.


Transmission-in-sump could have had a future

Left: H-Series engine. Right: K-Series engine.

Any AROnline readers who have read the Metro development story might recall that, back in the mid-1970s, the decision was made to build a car around a developed version of the A-Series engine and transmission. This was instead of designing a new power unit.

Both the ill-fated BMC 9X and ADO74 had been destined to have an all-new engine. Eminent Engineer Roy Brocklehurst described how the company was ‘tickled pink with its NVH performance’, once the upgrades had been made which turned it into the A-Plus.

A new decade, a new engine

However, what about the engine that had been passed over when the ADO74 was scrapped? Obviously, at the time of the Metro launch in 1980, the company’s Directors tended to describe the K-Series engine as not being a great enough step over the A-Series to justify the cost (and, in essence, they were correct to do so, as Bill Appleby and his team knew what they were doing when they designed the A-Series), but that is to undermine what Austin-Morris Engineers achieved with the H and K-Series engines.

In 1972, when the ADO74 programme began, the engine which was then intended to be used was called the H-Series and that employed the transmission-in-sump layout. As the ADO74 was initially devised as a straight replacement for the Mini, the capacity was limited to a maximum of 1000cc.

Transmission-in-sump, like its predecessor

The integrated engine/gearbox was designed for ease of servicing, and access was achieved by removing the single sump – the clutch was also similarly easily accessible through its cover. In one fell swoop, this clever design overcame the criticism of the A, B and E-Series engines that they were difficult to service. This design also did not sacrifice the undeniable packaging advantage of the transmission-in-sump system.

Ray Battersby, who was one of the H and K-Series Designers, said: ‘Harry Webster was the engineering dictator and control freak of all we did in the engineering block. His design strictures were designed to prevent any future boring and stroking of the cylinder block.

‘The very narrow cylinder bridge width would also cause much anguish by Coopers and Klingers, who were each tasked with producing a cylinder head gasket design able to withstand the heat, the fretting and general hostility in this bridge area without leaking between adjacent cylinders.’

So, what did the H-Series engine look like?

Battersby recounts that ‘the first H-Series engine ran on schedule on the test bed in December 1971.’ It was a 1000cc ‘dedicated’ engine that had been laid-out by Harry Webster in a planning document, released on 24 September 1970. It was an extremely compact unit which featured Siamesed bores (just like the E-Series engine) and an OHC cylinder head with bucket tappets (see Battersby’s comments, above). The distributor was mounted directly on the end of the belt-driven camshaft and the water pump and alternator were driven from the crankshaft.

The oil-pump was also crankshaft mounted. A two-shaft transmission layout was designed, and this was driven by helical spur gearing. This was carried in a housing integral with the rear transverse face of the cylinder block. The clutch was of conventional design, and this had a separate cover.

According to Ray Battersby, it was ready to go. ‘One engine, after running in, was subjected to a 400-hour Endurance Test, which started on 11 January 1972 and was completed on 11 February 1972,’ he said. ‘This gruelling test was the equivalent of running on the road for 26,000 miles with the engine running at full throttle and full load for over half of the time. The full test comprised of 69 cycles (each of 5:50 hours).’

Why was the H-Series dropped?

In total, four or five H-Series prototypes were built, and were run for between 200 and 800 hours on the test bed and 25,000 miles on the road in ADO16 mules. But the H evolved into the K-Series engine, when the ADO74 grew, and the limitations of the 1000cc upper limit became too much for the larger car.

In essence, the K-Series differed from the H in three areas: it had a larger engine capacity, the gearbox featured five ratios and more of it was inclined steeply to the rear to improve packaging.

The new two-shaft, five-speed gearbox was carried in a hefty housing, cast integrally with the rear face of the cylinder block (see diagram above) and, like its predecessor, the engine and gearbox were closed off underneath with a single pressed steel oil pan. As before, servicing was made relatively easy by this layout – since the gearbox internals and crankshaft could be released by the removal of simple half-bearing caps.

So, what did the K-Series engine look like?

K-Series engine was designed to be canted back, rather like Peugeot's 'suitcase' engine, as used in the 104
K-Series engine was designed to be canted back, rather like Peugeot’s ‘suitcase’ engine, as used in the 104

The K-Series evolution was to have been the definitive ADO74 power unit, and was designed in two forms: the 900/1100 version, and later 1100/1300. This change in policy was down to Harry Webster deciding to hedge his bets about the new supermini’s engine range make-up. It was laid down in late 1972.

Like the H-Series, the 900/1100cc K-Series was canted back but, according to Ray Battersby, ‘its cylinder bank was also canted back relative to the sump joint-face. Its doubly laid back cylinders allowed the unit to fit beneath the very low bonnet-line of ADO74.’

However, as Battersby put it, ‘something gave Harry Webster the jitters, because no sooner had we started designing the K-Series 900/1100 than he issued another edict, changing the capacity range yet again. This led to the birth of the K-Series 1100/1300cc project. Due to the continuing stipulation in Webster’s ‘Design Considerations’, neither the bore nor the stroke of the K-Series 1300cc engine could be increased.’

The H and K-Series prototypes are canned

The ADO74 project was then dropped and its engine died with it. The legacy of these engines was their inspiration within the design of the A-Series OHC and O-Series OHC engines, whose cylinder heads and valvegear bear the hallmarks of their ancestry.

According to Battersby, good progress was made on the H-Series before it grew up to become the K. ‘I believe that within this ‘family’ of engines (H-1000, K-900/1100, K-1100/1300), only the H-Series was ever built and only the H-Series was ever run on a testbed and fitted to a vehicle for road testing. I seem to recall that many of the K-1100/1300 parts were manufactured, but not assembled by the time the project was scrapped in 1974.’

Would the story of the Metro have been any different with this engine in place? Probably not, as the car was a success, and a long-lived one at that, with its A-Series engine.

Austin Ant/Firefly ADO74 prototype

Keith Adams
Latest posts by Keith Adams (see all)

35 Comments

  1. I dare say the A+ did perform well. But the Metro/Maestro where being launched just as Ford and GM/Vauxhall where ditching their Iron headed push rod lumps for the CVH and Family 1/2 respectively. Alloy heads, Overhead Cams, hydraulic tappets and power outputs for the standard engines that exceeded those for their predecessor “GT” versions. Slotting a 30 year old pushrod Engine into ARG’s new cars sent out all the wrong messages.

    • It never did the Renault 5 any harm using an old pushrod design. Theirs was alloy-headed and 5-bearing but a similar lifespan to the A Series.

      • I meant to specify the 1984 Super 5 there (being a similar ‘new car, old engine’ to the Metro), not the original 5.

  2. the metro and maestro still had problems with the transmission of one sort or another.

    Could the A+ not have use the gearbox and configuration shown above, (or the E series in the case of the maestro).

    It could have saved AR money wasted importing VAG units and looks like it could also have had positive benifits in terms of packaging and weight distribution.

    I know both metro and maesto handled well, but both could have benifited.

    • Drae,

      Better late than never…

      The E-Series transmission with its remote cable-operated gear-shift was berated by the press from its launch for indistinct changes that felt like pushing the gear stick through a box of loose ball-bearings. Remember that at that time, the mass-produced Ford gearbox fitted to the Escort, Cortina etc was absolutely superb; slick, short movements and strong with large synchromesh hubs. It was used by many track racing cars. A brilliant gearbox design.

      No, no no. The E-Series box was a gonner for that reason alone.

      Then there was the crippling effect of designing an engine around an alien component. The H and K-Series engines were crippled enough by board members who specified its minutiae even before a designer had sharpened his first pencil.

      We engine designers at that time worked literally arm in arm with our colleague transmission designers who were all extremely proficient. The problem for both teams (about 25 designers in each team) was a company ham-strung for cash and a management team who were way out of their depth.

      • Ray Battersby

        Given your involvement as a designer of the H/K-Series engine, was wondering whether you can clarify some things regarding this project?

        Did the H/K-Series share anything with the A-Series as one or two books have suggested a connection with the A-OHC?

        How did the H/K-Series compare with the A-Plus and A-OHC (assuming the latter is completely unconnected with the H/K unit)?

        What was the minimum capacity of the H/K-Series units and had the project been given the green light, how much further could the later K-Series engine capacity have been increased by beyond 1300cc upper limit?

        Thanks in advance

  3. Harry Webster was the Engineering dictator and control freak of all we did in the engineering block. His deputy, Ray Bates was a poodle by comparison (though he had much bigger feet).

    Regarding the capacity of the H/K-Series engines they were chronologically (from memory):

    H-SERIES – 1000cc ‘dedicated’ engine. In his single-sided, single sheet document ‘DESIGN CONSIDERATIONS FOR NEW 1000 CC UNIT’ dated 24 Sep 1970, Harry Webster stipulated…

    1. Fully siamesed bores

    7. 3 Main Bearing Crankshaft*

    17. Cylinder bridge width 7/32″ **

    21. Crankcase to have approx. 1/8″ clearance for conn rod path.

    * See later
    ** 0.218″/5.5mm

    Collectively, these strictures were designed to prevent any future boring and stroking of the cylinder block. The very narrow cylinder bridge width would also cause much anguish by Coopers and Klingers who were each tasked with producing a cylinder head gasket design able to withstand the heat, the fretting and general hostility in this bridge area without leaking between adjacent cylinders.

    Two further designers joined the team between January and March 1971. The first H-Series engine ran on schedule on the test bed in December 1971.

    K-SERIES 900/1100 – Despite Webster’s steely determination to have a ‘dedicated’, ham-strung 1000cc engine, early in 1972, he decided to hedge his bets. He now wanted it to be a dual-capacity design encompassing 900 and 1100ccs.

    This 900/1100 engine became the basis for the K-Series whose entire power unit – like the H-Series – was canted back on installation. However, its cylinder bank was also canted back relative to the sump joint-face. Its doubly laid back cylinders allowed the unit to fit beneath the very low bonnet-line of ADO74.

    K-SERIES 1100/1300 – But…but something had clearly given Harry Webster the jitters because no sooner had we started designing the K-Series 900/1100 than he issued another edict, changing the capacity range yet again. This led to the birth of the K-Series 1100/1300ccs project. Due to the continuing stipulation in Webster’s ‘Design Considerations’, neither the bore nor the stroke of the K-Series 1300cc engine could be increased.

    But it didn’t matter anyway. The entire ADO74 project was dropped including work on its engine. I believe that within this ‘family’ of engines (H-1000, K-900/1100, K-1100/1300) only the H-Series was ever built and only the H-Series was ever run on a test-bed and fitted to a vehicle for road testing. I seem to recall that many of the K-1100/1300 parts were manufactured but not assembled by the time the project was scrapped in 1974.

    The legacy of these engines was their inspiration within the design of the A-Series OHC and O-Series OHC engines, whose cylinder heads and valve-gear bear the hallmarks of their ancestry.

    In his excellent book ‘The A-Series Engine’, Graham Robson says that the H-Series ran for 200 hours on the test bed and completed 25,000 miles on the road. Actually one engine, after running in, was subjected to a ‘400 hour Endurance Test’ which started on 11 Jan 1972 and was completed on 11 Feb 1972. This gruelling test was the equivalent of running on the road for 26,000 miles with the engine running at full throttle and full load for over half of the time. The full test comprised of 69 cycles (each of 5 hours 50 minutes duration). During each of these cycles the engine speed and load was varied as follows:

    RPM LOAD TIME
    1000 None 10 mins
    6250 Full 10 mins
    4500 ½ 10 mins
    4800 ⅓ 60 mins
    2400 Full 60 mins
    2400 ¼ 10 mins
    5400 Full 60 mins
    4500 ¼ 10 mins
    6000 Full 60 mins
    6000 ¼ 60 mins

    A-SERIES CARRY-OVER PARTS (Ignoring fasteners)
    Oil filler cap
    Mainshaft Bearing Caps and Screws
    Water pump bearing and seal

    • Why was it so important to the Harry Webster – The leather patch king of BL – that the engine should not be capable of being bored or stroked? Deliberately removing any future proofing from this critical and expensive component almost sounds like some sort of industrial sabotage!

      • The following is admittingly my very limited conjecture though perhaps it is related to unrealised plans to develop 4-cylinder versions of the PE166 6-cylinder to replace the Triumph Slant-4, a 4-cylinder version of the 2.3 SD1 Six for example would displace around 1566cc.

        However while the H/K-Series prototype units were all-new designs, given the PE166 was loosely derived from the Triumph I6 OHV. It brings up the question of to what degree were the H/K-Series engines loosely derived from the 803-1493cc Triumph 4-cylinder OHV engines or at least influenced by Triumph thinking (apart from retaining a form of BMC’s in-sump transmission layout)?

        Would definitely agree regarding the brief of preventing further enlargement beyond 1300cc as being a mistake with the K-Series prototype engine, particularly since the Supermini opposition were moving towards 1400-1600cc engines (and had it reached production there is little certainty a PE166-based 1566cc 4-cylinder unit would have even fitted into ADO74).

    • 1000 RPM, No Load 10 minutes.
      then
      6250 RPM Full Load 10 minutes

      Rather harsh a treatment for an engine which has not reached full working temperature

  4. This basic essence of this design of course DID make it into production – as the Peugeot/Citroen “suitcase” engine – as found in the 104/Citroen LN/Talbot Samba and the early 205s. Almost identical in fact, with the engine canted so far back at an angle, and the gearbox sort of on the side. Proof that in Britain we’re good at coming up with the initial idea, but lousy at following it through.

    • It does give a bit of insight as to how the H/K-Series engines would have fared in the Metro using the PSA X/TU engines as a rough guide, especially given the limitations of the Metro’s end-on gearbox that did much to hamper the output of the MG Metro 1275 Turbo from 120-130 hp down to 94 hp and 85 lb⋅ft (115 N⋅m) in order to preserve the life of the gearbox.

      It would seem the closest naturally aspirated sporting equivalent to an 1.3-litre H/K-Series powered MG Metro at best would be the 94-102 hp 1.3 PSA TU2 in the Peugeot 205 Rallye, Citroen AX Sport and Peugeot 106 Rallye as well as the possibly 84-99 hp 1.4 PSA TU3 in the Peugeot 205 XS/Citroen AX GT and Peugeot 106 XSi/CItroen AX GTi.

      However if 85 lb⋅ft (115 N⋅m) is indeed the maximum reliable torque limit for the Metro’s end-on gearbox (unless other gearbox factors are involved) than cannot really see a 1.3 H/K-Series MG Metro being significantly different to the existing A-Series MG Metro 1.3 Turbo in terms output and torque figures, let alone to the 94-99 hp Peugeot 106 1.3 Rallye/Citroen AX 1.3 Sport or later catalyst 94 hp versions of the Peugeot 106 1.4 XSi/Citroen AX 1.4 GTi.

  5. I felt the A-Plus did a good job in the Maestro, let down by the gearbox and in one model silly ratio’s for economy. If my memory serves me correctly wasn’t the E-series designed to be 1300cc?

    • Think it may have been, but obviously 1.5 was as small as it ever got. I think there has been few cases where manufacturers have succeeded in building a truly “scalable” engine that can be both very small and very large. Trouble is at lower cubic capacities, you still have a big heavy engine displacing a small capacity so they end up offering little or no economy advantage over an engine that’s designed to be small in the first place. Ford for instance tried it twice and failed….

      – Ford CVH; designed to go as low as 1.1 litres but this capacity was dropped when they found out that it wasn’t any more economical than the old 1.1 Kent/Valencia engine yet was more expensive to build.

      – Ford Zeta/Zetec – again it was designed to be as small as 1.4 litres, but never got this low.

      But in essence most European manufacturers of that era seemed to use two engine families in a ‘small’ and a a ‘large’ format to cover the entire 1.0-2.0 size range – some examples:

      Peugeot-Citroen = TU (1.0-1.4 litres) XU (1.6-2.0 litres)
      GM/Opel-Vauxhall = Family I (1.2-1.4 litres) Family II (1.6-2.0 litre)
      VW/Audi = EA111 (0.9-1.4 litres); EA127 (1.5-2.0 litres)
      Renault = C-type/E-type (0.8-1.4 litres) F-Type (1.7-2.0 litres)

      sure people can think of some more!

      • Interesting, do any specs exist on the unproduced 1.4 Ford Zeta engine? Would have certainly been an improvement over the 1.4 Ford CVH.

      • Remember the old Alfa DOHC four from the Sixties?
        Originally designed as a 1.6 it was downsized to a 1.3 which was never as lively or well loved as its predecessor that was designed as a truly phenomenal 1.3.
        Later it was bored and stroked to 1.75 which was a very racy engine with some durability and reliablity problems. Taking strange measures like asymmetrical conrods to move cylinder bores further apart it was increased to 2.0 which was the least pleasing iteration, unwilling to rev and with NVH problems when compared to its smaller siblings.

        • Not sure if it was exactly the same engine or a semi-related design though the Alfa Romeo Tipo 103 prototype featured a 51 hp 896cc Alfa Romeo Twin-Cam.

      • The smallest Ecoboost engine is 1 litre, the largest is about 4.6 litres (in the new Mustang). Are they basically the same engine scaled up/down or different designs under the same name?

  6. I was never a fan of the engine and the gearbox sharing their oil. The two require different lubrication properties. Probably why no-one does it today.

    Strikes me that Harry Webster was a bit of a ditherer. As any engineer will tell you, there is nothing worse than moving the goal posts half way through a project for increasing the cost.

    I have nothing against the A series which was a reliable and economical unit, but it was always handicapped by the 5 port cylinder head.

    • Motorcycles used shared oils, engine, transmission and a wet clutch, and typically operate at 100 bhp / litre, the top selling vehicle in the world, the Honda Supercub (nifty fifty), 120 million and still counting, has always used a shared system for the horizontal-cylinder air-cooled engine. The Honda also lacks a disposable oil cartridge filter,the filter is a metal gauze sieve.

  7. When was the A series OHC built? About 25 years ago I bought an O series 16v twin cam which I ran in an MG Maestro, the chap I bought it from not only had a few of these but also quite a few A series OHC engines. Were these originally intended for the metro?

  8. @Kevin Steele.

    The Peugeot engine laid flat was a fantastic pice of packaging, but it was a pain to work on.My brother’s 104 was pretty long in the tooth by the time we had it so had it’s issues. We became well versed with it while he had it, but it lasted longer than the rotbox Metro he replaced it with.

    Minor services were easy enough, but anything major was an engine out. With one working on top and one below,we could have the entire engine and gearbox sitting in front of the car in 40 minutes using just our basic facilities at the roadside.

    As was the case with a lot of French cars of the time, the spare wheel sat on top of the engine was well. That couldn’t have done the rubber any good!

    • I did wonder how well Peugeot sorted the oil flow in the suitcase engines.

      In F1 BMW canted over their turbo engine for the Brabham BT55 & it had all sorts of problems with the oil flow being disrupted.

  9. The H and K engines would have probably gone over budget, when British Leyland was at its most cash strapped, so keeping the venerable A engine and upgrading it to the A plus made sense at the time. I do remember the engine being premiered in the Ital, and the improvements in economy, performance and refinement were considerable over the A series Marina. Also in the Metro, the 1275cc A plus endowed even non MG versions with a top speed of nearly 100 mph, excellent for the time, and economy over 40 mpg.

  10. It is amusing to note the Metro/100 later featured a 1.4/1.5 PSA TUD diesel whose linage can be traced back to Peugeot’s ‘suitcase’ engine, just a pity the H/K-Series never merited to evolve into a full British analogue of the Peugeot “suitcase” / PSA TU engine (including dieselized/turbodiesel variants) due to cost as well as Webster’s continuing stipulation for neither the bore nor stroke of the 1300cc engine could be further increased.

    What the company really needed was an engine family ranging from sub-1000cc up to 1600cc, the most disappointing thing would be that such an engine should have appeared under BMC at the early/mid-60s at the very earliest akin to a British Nissan A OHV / E OHC with a more viable 9X engine ideally being a 3/4-cylidner precursor to the Volkswagen EA111 instead of being an unnecessary 4/6-cylinder engine. Aside from the H/K-Series engine, there was also some potential with the 4-cylinder version of the PE146/PE166 as a sort of successor to the 803-1493cc SC engine.

    While the H/K-Series was to be canted back in ADO74, could the engine have easily slotted into the space of the existing Mini and Metro?

    If so could it have also slotted into the space of the two aforementioned cars had both utilized an end-on gearbox layout like on the R6 Metro/100 with the altogether different production K-Series?

    Know that the production K-Series was apparently slightly longer than the A-Series being the equivalent of a 3 and a half cylinder A-Series in length, which initially necessitated the Mini featuring a 3-cylinder K-Series in the first Minki prototype before the later Minki-II was both widened and lengthened by 2-inches in order to easily slot in a 4-cylinder K-Series engine.

    Just curious to know how the H/K-Series dimensions compare with both the A-Series and K-Series.

    • Meant to say the A-Series was roughly equivalent to a 3 and a half cylinder K-Series engine in length as opposed to the reverse.

  11. Interesting that these engines employed just two shafts, specifically the crankshaft and camshaft, with the former also carrying the oil pump and the latter the distributor. Neat idea because it eliminates the need for a separate shaft and set of skew gears to drive those items. And a toothed belt for the camshaft too – all very advanced. This arrangement did see the light of day in the Vauxhall/Opel OHC engines in the Astra and Mk2 Cavalier though.

    Second point, must also have reduced cost of manufacture – compare with the amount of shafts and skew gears in the Triumph slant-4 for example.

    • Fascinating. Are you referring to the GM Family I engine? Apparently the latter was originally conceived by Opel during the 1960s as a 850-1600cc OHC unit before being utilized by GM Brazil in the Brazilian Chevrolet Chevette 1.6 and later forming the basis for the GM Family I engine.

  12. Nate – yes I was. Maybe the GM engine came first then. The GM version used in mk2 Cavaliers loses points for having the cam follower ‘finger’ components made from chocolate tho!

    • I heard they could suffer from lazy tappets, piston slap & camshafts that needed replacing after 80k miles.

      Luckily GM made the next generation of engines a lot stronger, supposedly able to last 300k miles with enough TLC.

  13. Had the Metro received the 0.9-1.3-litre H/K-Series engines in place of the A-Plus units, can still see its end-on gearbox being the limiting factor as was the case with the 94 hp MG Metro Turbo.

    In which case its output both power / torque in naturally aspirated MG Metro form would probably be little different to the 94-102 hp 1.3 PSA TU2 in the Citroen AX 1.3 Sport and Peugeot 106 1.3 Rallye, along with possibly the catalyst 94 hp 1.4 PSA TU3 in the Citroen AX 1.4 GTi and Peugeot 106 XSi.

  14. PAUL’s Query of 2018 – Like good wine, you have to wait for maturity…

    Why did Harry Webster insist on physical design features that restricted future increases in engine capacity? By this I mean preventing the cylinders from being bored out and preventing the use of a longer stroke by giving the inner profile of the crankcase just 4-5mm clearance from the con rod as it ran around inside.

    Why? I recall the reason given was, “Look what happened to the A-Series. It started as 803cc and now its over 50% bigger. Look at all the problems that’s caused!” By this Webster meant (for example) the 1275 unit still running a 3 bearing crankshaft. He meant the 1275’s cylinder head gasket sealing problems between cylinders 1+2 and 3+4. The gap between 2+3 was greater due to the central main bearing spreading things apart. He meant cooling problems caused by completely removing water circulation between the cylinders inside the crankcase. No water between the bores = very inadequate cooling. Poor cooling leads to cylinder distortion that leads to piston scuffing and ultimately, seizure. He meant problems with inadequate breathing capacity. If the cylinder diameter is limited, so too are the valve sizes which reduces volumentric efficiency, hence low performance, poor fuel consumption and excessive emissions.

    So Harry Webster was right in principle but the reason for the later A-Series problems was actually weak management.

    Things would have been completely different had the bosses, back in the day, said “No. We won’t increase the A-Series’ capacity. We won’t compromise its cooling, its reliability its efficiency. We can design you a bigger engine instead.”

  15. Was it down to the Machine Tools the Factory inherited when Austin and Morris merged?
    The bore to bore centres of the engine block had to be the same for all capacities of the A ana A+ engines

  16. In the very early 80s I was working in South Oxfordshire and one of the people who was part of the project had a son who was based up the road at Cowley, he had a Metro with a three cylinder engine. OHC and with fuel injection. It was apparently a vehicle being used to get experience of this engine in everyday use. There was a data logging unit in the back, which apparently recorded stuff like the number of warm and cold starts, the gear it was driven in, and throttle openings / engine revs.

    Was probably a version of the K series engine, but I don’t recall that there was ever any production attempts on it.

Leave a Reply

Your email address will not be published.


*


This site uses Akismet to reduce spam. Learn how your comment data is processed.