Tuesday, 30 April 2013

The West Lothian Answer

A few weeks ago saw the publication of the Report of the Commission on the Consequences of Devolution for the House of Commons.

I need to explain my own background here. I am mainly English but with Irish ancestry. When I have holidayed in the Republic of Ireland I don't feel that I am abroad. I have loads of Northern Irish friends. I have lived in Scotland. When I was 12 we nearly moved to Wales.

I am politically a Unionist.

One thing that grates is when people consider the United Kingdom of Great Britain and Northern Ireland to be just a Greater England. Attitudes like "Well, if the Northern Ireland Protestants want to be English they can move to England and leave Ireland to the Irish", "Give the IRA a united Ireland so they'll leave England alone" or "Well, the Scots have their own Parliament so they shouldn't sit in England's" are both ignorant and annoying. I have not yet met anyone from Northern Ireland who desires to be English. The Parliament of England hasn't met for 306 years.

Indeed a few years back I was catching up with a friend from Northern Ireland who was living in London. Ireland had beaten England at rugby a few days earlier, and we passed a couple of guys wearing England rugby shirts. He started clapping his hands above his head and chanting "Eye-er-land". We all laughed. Precisely - just because he's a staunch Unionist doesn't mean he wants to be English.

The last "Queen of England" was Anne.

One lady, mentioning Andy Murray win the Gold at the 2012 Olympics, expressed her gladness that an Englishman had won. I pointed out that Murray was British, not English. "Same thing" she snapped.

By the way, why is Murray referred to as British when he's winning and Scottish when he isn't?

The West Lothian Question is often a variant of "Is it fair that Scottish MPs can vote on English-only matters post-devolution when English MPs have no say on Scotland-only matters?"

Now, one way to respond to a question "Is it fair that..?" is to ask "Is it true that...?"

The Scotlsnd Act 1998 created the Scottish Parliament, giving it legislative powers over devolved matters.

One thing has to be noted here - Scottish devolution did not begin in May 1999. There was effectively a Scottish Executive (namely the Scottish Office ministers appointed by the Prime Minister) already in existence. Scotland's education was a matter for the Scottish Secretary, not the Education Secretary. Its health service was a matter for the Scottish Secretary, not the Health Secretary. Its police were a matter for the Scottish Secretary, not the Home Secretary.

There was a bit of a fuss when Labour's John Reid served as Health Secretary (from June 2003 to May 2005) and later as Home Secretary (from May 2006 to June 2007). Didn't it show the problems with the devolution settlement, that he could make decisions that affected the people of England but couldn't make decisions that affected his constituents in Hamilton North & Bellshill (when Health Secretary) or Airdrie & Shotts (when Home Secretary)?

Firstly, there are reserved matters - issues which the Scottish Executive cannot take action on, nor the Scottish Parliament legislate on - within the Health Secretary's and Home Secretary's remit.

Secondly, even if the Scottish Parliament had not been created, there would be large areas in health and home affairs that would have been dealt with on a day-to-day basis by a minister in the Scottish Office and would not have come anywhere near Reid's desk.

Next up is a problem in the way that the West Lothian Question is framed, for it is based on a flawed premise. The Scotland Act makes clear that sovereignty remains at Westminster.

What do the Regulation of Care (Scotland) Act 2001, Protection of Children (Scotland) Act 2003 and Criminal Justice (Scotland) Act 2003 have in common? They were all amended by the Serious Organised Crime & Police Act 2005 - so there is an example of a piece of legislation passed by Westminster that amends legislation on devolved matters passed by the Scottish Parliament.

The Scotland Act made clear that the Scottish Parliament's authority was limited to matters where it had "legislative competence" and legislation passed by it that was ultra vires was invalid. Firstly, before allowing any Bill to be presented, the Presiding Officer had to ensure that it was within the Scottish Parliament's legislative competence. In addition, after the Scottish Parliament passed a Bill, there is a 28 day period in which the Lord Advocate (a devolved position), Attorney-General (a reserved position) or the Advocate-General (a reserved position) could refer the Bill to the Judicial Committee of the Privy Council for a ruling on whether it was ultra vires or not. In addition, in that 28 day period, the Scottish Secretary could rule that the Bill not become law if it has an impact on reserved matters, or interfered with the United Kingdom's international obligations, or our defence or security.

The Judicial Committee of the Privy Council's powers relating to the Scottish Parliament were transferred to the new Supreme Court under the Constitutional Reform Act 2005.

So, the Scottish Parliament is not sovereign even when it comes to devolved matters in Scotland. There are courts and Westminster-based politicians who can stop legislation it passes becoming law. And Westminster can - and does - legislate on devolved Scottish matters, even amending legislation passed by the Scottish Parliament.

One suggestion to answer the inaccurately-framed West Lothian Question is "English votes for English laws" - namely that there is "England-only legislation" which only MPs from England should vote on. And there are two problems with this.

The first problem can be considered by looking at the May 2010 general election result in England only:

  • Conservatives - 297 (including 1 Deputy Speaker)
  • Labour - 191 (including 2 Deputy Speakers)
  • Liberal Democrats - 43
  • Greens - 1
  • The Speaker - 1

This gives the Conservatives an overall majority of 63.

However, across the United Kingdom the result was different:

  • Conservatives - 306 (including 1 Deputy Speaker)
  • Labour - 258 (including 2 Deputy Speakers)
  • Liberal Democrats - 57
  • Democratic Unionist Party - 8
  • Scottish National Party - 6
  • Sinn Féin - 5
  • Plaid Cymru - 3
  • Social Democratic & Labour Party - 3
  • Greens - 1
  • Alliance Party of Northern Ireland - 1
  • Independent Unionist - 1
  • The Speaker - 1

The story is familiar - the Conservatives and Liberal Democrats formed a coalition Government. But things could have been different. The Speaker and Deputy Speakers don't vote (by convention) and the Sinn Féin MPs don't vote (as a matter of principle), leaving 641 voting MPs. Between them Labour and the Liberal Democrats had 313 voting MPs - and hence just needed 8 more to have a majority. The Democratic Unionist Party had previously shown that it doesn't take much taxpayers' money to win them over, so 8 votes could come from that direction. Or Labour could look to Plaid Cymru - with whom it was forming a coalition in the Welsh Assembly - and other minor parties to give it the votes it needed.

So, we could have ended up with a Labour/Liberal Democrat coalition Government in charge now. And if we had "English votes for English laws", then on some divisions they would only have 232 of the 529 English voting MPs. Hence, any England-only legislation would run into trouble in the House of Commons.

There would be Cabinet ministers - Communities & Local Government Secretary, Home Secretary, Education Secretary, Health Secretary etc. - who would find themselves with a problem, introducing Bills that represent a Labour/Liberal Democrat agreement, only to find them chucked out by what is, for brief periods, a House of Commons in which the Conservatives have an overall majority. Now, Governments have existed briefly when another party has held an overall majority (the Liberal Government between December 1905 and February 1906 was the last example), but this is normally only for a short period while a general election is prepared.

This would be a situation existing for the entirety of the Parliament. It would be unstable.

The second problem is that we don't have a unicameral Parliament. There is the House of Lords, and members represent the whole of the United Kingdom. Even under the aborted plans for a majority-elected House of Lords, there would still be members without a territorial affiliation. How then do you define who in the House of Lords can or cannot vote on an "England-only" matter?

There is a third problem, which is more fundamental. What is an "England-only" piece of legislation anyway? It is not as obvious as it seems, and here we need to look at a bit of history.

The American Revolution was based on the idea of "No taxation without representation". The problem with "English votes for English laws" is that there will be people in the United Kingdom who are taxed without being represented.

In July 1886 there was a general election held in the shadow of devolution - this being the issue of Irish Home Rule:

  • Conservatives & Liberal Unionists* - 393
  • Liberals - 192
  • Irish Parliamentary Party - 85

[*In British Electoral Facts, Colin Rallings and Michael Thrasher state that approximately 77 of these were Liberal Unionists - which leaves aproximately 316 Conservatives]

A minority Conservative Government was formed, with the Liberal Unionists supporting it from the backbenches. In December 1886 Randolph Churchill resigned as Chancellor of the Exchequer, and he was replaced by George Goschen, who was the first Liberal Unionist in the Cabinet.

Goschen made an important decision at the Treasury, one that has implications today.

That decision was to allocate grants to local authorities in the following proportions - 80% to England & Wales, 11% to Scotland and 9% to Ireland - which actually gave Scotland less per capita than England & Wales. The Education (Scotland) Act 1918 was the first piece of legislation to formally use this formula in giving Scotland 11/80th of the spending allocated to England & Wales.

The October 1974 general election saw the sitting Labour Government gain a small majority, and proceed with its plans for devolution to Scotland and Wales, leading to the Scotland Act 1978 and the Wales Act 1978. Joel Barnett, the Chief Secretary to the Treasury at the time, used the Goschen formula as the template for the issue he was looking at - how much money should go from the Treasury to the devolved assemblies being planned in Edinburgh and Cardiff. And he based this on what we can call "equivalent devolved expenditure" - i.e. how much was being spent in England on things that would be devolved in Scotland and Wales.

The formula that Barnett came up with was that the amount given to the proposed Scottish and Welsh Assemblies would be based on the amount of equivalent devolved expenditure in England, but that increases/decreases would depend on increases/decreases in England. So, if equivalent devolved expenditure went up/down by £85, then the amount given to the Scottish Assembly would go up/down by £10 and that given to the Welsh Assembly would go up/down by £5.

In March 1979 the people of Wales rejected the Assembly in a referendum, and the people of Scotland voted in favour of an Assembly - albeit not enough to get over the "40% of the electorate" threshhold. The Conservatives won the May 1979 general election, and the devolution proposals were shelved. However, the Barnett formula lived on in determining the Scottish block grant given to the Scottish Office.

And the block grant given to the Scottish Parliament is descended from this. Herein lies the problem.

Yes, it may seem that health and education are "England-only" matters. But most legislation has a financial implication. Suppose health policies lead to a reduction or an increase in the spending on health in England. Under the Barnett formula, this leads to a reduction or an increase in the Scottish block grant. And under "English votes for England laws" there would be no Scottish MP able to have a say. Decisions which have an impact on Scottish finances would be made solely by English MPs. No taxation without representation.

Something else needs to be mentioned about financing. Some of the more emotive arguments for an English Parliament, or for "English votes for English laws", or against Scottish devolution concentrate on the way the block grant is spent. We hear about NHS prescriptions for example. However, when the Scottish Secretary had control of the block grant, it was up to him to decide how to spend it. So, there is no reason why a decrease in spending on English schools should be matched by a decrease in spending on Scottish schools - yes, there would have to be a decrease on some item of spending but it could be elsewhere in the Scottish budget.

If you were to ensure that there was a common policy on health spending, or schools spending, then you would have to go further than abolishing just the Scottish Parliament. You would have to abolish the post of Scottish Secretary itself, and have a Health Secretary determining a common United Kingdom health policy, an Education Secretary determining a common United Kingdom education policy etc.. You would not be turning the clock back to before 1999 but to before the Conservatives' Charles Gordon-Lennox became the first Scottish Secretary in August 1885. It appeals to those who believe that when you cross a bridge over the River Sark you are simply crossing the border between Cumberland and Dumfriesshire - nothing more than that.

There is another thing connected to this - a slight tangent. We should be careful of defending/opposing any organisation having powers based on what it does with them. At some point we should have a referendum on the European Union.

Now, tabloid objections to the European Union, and also the unconnected European Court of Human Rights (or the "EU Court of Human Rights" as some tabloids use to confuse us), rest on things like "Brussels states..." or "Strasbourg ruled...".

The problem is that this ignores the fundamental question - should the EU and/or ECHR have these powers to begin with, regardless of whether they use them in ways we approve or not? For example, Parliament has passed legislation I disapprove of - but I don't deny they have a right to. Judges have come up with barmy rulings - but I agree they have a right to pass barmy rulings.

So, just because the Scottish Parliament uses its powers to do something we don't agree with, this is not an argument that it should not have those powers. England is not meant to be Scotland's smother-mother, saying "Naughty Jock, introducing free personal care for the elderly. I'm going to have to ground you. You cannot pass any legislation for a month".

"English votes for English laws" sounds fair on first hearing, but falls apart when you look at it more carefully.

Another suggestion to the West Lothian Question is the idea of a devolved English Parliament. There is even a Campaign for an English Parliament.

Is this the solution? Well, as the wise philosopher Victoria Pollard would say "Yeah, but no, but yeah, but no..."

It sounds simple - have a devolved English Parliament on the Scottish model. But there is one flaw, which is connected with the error in the way the West Lothian Question is phrased.

The Campaign for an English Parliament uses the example of tuition fees.

In July 1997, Ron Dearing presented the Labour Government with his report on the funding of higher education, which had been commissioned by the previous Conservative Government. One of the suggestions was that students pay around 25% of their tuition fees.

The Teaching & Higher Education Act 1998 introduced the principle of students contributing up to £1,000pa to their university education. However, under the Scotland Act 1998, this was a devolved matter.

The May 1999 election to the Scottish Parliament gave the result:

  • Labour - 56
  • Scottish National Party - 35
  • Conservatives - 18
  • Liberal Democrats - 17 (including the Presiding Officer)
  • Greens - 1
  • Scottish Socialist Party - 1
  • Independent - 1

With 72 voting Members of the Scottish Parliament between them, Labour and the Liberal Democrats were able to form an administration with an overall majority of 16. But one area of difference was tuition fees.

Hence, a commission was set up under Andrew Cubie, which reported in December 1999. The main recommendation was abolishing up-front tuition fees but making students pay £3,000 of them when their salary reached £25,000pa - courses in England, Wales and Northern Ireland tended to be 3 years (in Scotland they are 4 years, as pupils sit their Highers exams a year before pupils elsewhere in the United Kingdom sit A-levels. This is reflected in the university offers, e.g. I applied to do Maths at Edinburgh University and the offer stated I could go straight into the second year if my A-level grades were good enough) so the £3,000 was the same as students in England, Wales and Northern Ireland would expect to pay overall.

The Scottish Executive introduced the Education (Graduate Endowment and Student Support) (Scotland) Act 2001 which created the "graduate endowment" of £2,000 to be paid by students from within the EU (apart from England, Wales and Northern Ireland) who did their first degree at a Scottish university once their salary reached £10,000pa.

Students from England, Wales and Northern Ireland studying at Scottish universities would have to pay tuition fees.

At the June 2001 general election, Labour had a manifesto commitment not to increase tuition fees. However, they decided to introduce plans to allow universities to charge variable top-up fees and this was debated in the House of Commons on 27 January 2004, with the Bill that became the Higher Education Act 2004 passed by 316 votes to 311 - a majority of only 5.

Part of the controversy was that Scottish MPs voted on it, with 46 Labour MPs voting in favour and 16 Scottish MPs (10 Liberal Democrat, 5 Scottish National Party, 1 Respect) voting against.

So, would an English Parliament similar to the Scottish Parliament have been the solution?

Well, no it wouldn't - and the reason it wouldn't goes back to the fundamental misunderstanding that underlies the West Lothian Question.

Suppose a Conservative-dominated English Parliament voted not to introduce tuition fees. What legal provisions would stop a Labour-dominated United Kingdom Parliament voting to introduce them in England (and relying on Scottish MPs for that)? Westminster has simply amended legislation on devolved matters passed by the Scottish Parliament, with there being no calls that only MPs representing Scottish constituencies vote on devolved Scottish matters.

Hmm, so it seems to get what campaigners for an English Parliament want, and to avoid the scenarios they use, we would need both an English Parliament and "English votes for English laws".

As the Scotland Act 1998 makes clear, the Scottish Parliament is not sovereign. It is a devolved legislature, not a federal one.

There is a difference. A devolved legislature owes its existence to a higher legislature - Westminster created the Scottish Parliament, can increase or decrease its powers, or can abolish it, simply by passing legislation.

Devolution is top-down.

A federal legislature is one where the higher legislature owes its existence to the lower ones. This is where it gets interesting.

After it was first elected, the first devolved Scottish Parliament had to hold a meeting with the normal business, the swearing-in of Members of the Scottish Parliament and election of the Presiding Officer.

And chosen to preside over this was the oldest MSP, the Scottish National Party's Winnie Ewing, elected as an MSP for Highlands & Islands (and also at the time the Member of the European Parliament for Highlands & Islands), who used the words:

The Scottish Parliament, adjourned on the 25th day of March, in the year 1707, is hereby reconvened.

The Scottish Parliament was not adjourned and then reconvened 292 years later. It was abolished and a new one created 292 years later.

Which leads to the question - who abolished the Scottish Parliament?

The Parliament of the United Kingdom has its powers because in 1800 the Parliament of Ireland and the Parliament of Great Britain passed legislation abolishing themselves and tranferring their powers to a Parliament of the United Kingdom. And the Parliament of Great Britain had its powers because in 1707 the Parliament of Scotland and the Parliament of England abolished themselves and passed legislation transferring their powers to the Parliament of Great Britain.

So, we seem to be in an odd situation. There is a legislature based in Westminster which owes its existence to legislatures each covering a smaller geographical area transferring their powers to it - but we are not a federal state. If only these Parliaments had transferred some of their powers then we would be living in a federal nation and would not be having this debate.

Is federalism the answer? Is it British?

When the American colonies rebelled in the 1770s, they saw federalism as the solution - a group of colonies coming together as one nation with some powers at federal level and some at state level, with the Tenth Amendment making clear that the new nation would use the principle of subsidiarity:

The powers not delegated to the United States by the Constitution, nor prohibited by it to the states, are reserved to the states respectively, or to the people.

Note what it does not say:

The powers not delegated to the states by the Constitution, nor prohibited by it to the United States, are reserved to the United States.

That is the complete opposite to devolution. Even the Scotland Act 1998 uses "reserved" to mean the powers which are not transferred to the Scottish Parliament. There is a default setting in both devolution and federalism - in devolution that default setting is power residing at the higher legislature, while in federalism that default setting is power residing at the lower legislatures.

And to answer the question whether it is British, we need to ask ourselves which country is culturally the closest to the United Kingdom. I would say Canada.

By 1840 there were the North American colonies of Lower Canada and Upper Canada. The British North America Act 1840 combined these into the Province of Canada. The British North America Act 1847 divided the Province of Canada into Ontario and Québec, and combined them with two other colonies - Nove Scotia and New Brunswick - to form a federal nation, Canada.

Hence, by the 1860s the British Empire was experimenting with federalism - so it can hardly be described as unBritish.

This seems to be the solution now. It would create something more than a devolved Parliament - an English Parliament which is itself sovereign and not dependant on another Parliament. Which could not have its powers altered or be abolished by Westminster.

And this would create something new. At the moment, the Supreme Court can rule that the Scottish Parliament has acted ultra vires, but cannot rule that Westminster has acted ultra vires. If there were an English Parliament in a federal system that would have to change.

The Speaker of the House of Commons would have to ensure that any legislation brought before the House of Commons did not infringe on the English Parliament's powers.

The Law Officers of the English Government would have the power to ask the Supreme Court to rule that a Bill just passed by Westminster was ultra vires and have it struck down ensuring it never gets Royal Assent.

Answering the West Lothian Question puts England - and the United Kingdom - into new constitutional situations.

Saturday, 27 April 2013

As Jupiter Approaches Saturn

I got home about 11 last night from work, and it was clear (the sky is taken from Heavens Above, as are all charts in this):

Jupiter was sinking in the west, while Saturn was getting higher in the east.

And, as the decade unfolds, they will be getting closer.

A bit of maths. If one planet takes A time units (days, years, centuries) to go round the Sun, and another one takes B time units, then the number of time units (C) that it takes for them to line up compared to the Sun is given by:

1/C = |(1/A - 1/B)|

The vertical lines are there to denote the absolute value of something, e.g |1| and |-1| are both 1.

Jupiter takes 4,332.59 days to go round the Sun, and Saturn takes 10 759.22 days. So, the values we have are:

1/A = 0.000 230 809

1/B = 0.000 092 944

This gives us:

1/C = 0.000 137 865

And hence:

C = 7,253.47 days

Using an average year of 365.25 days, this works out as about 52 days short of 20 years.

This isn't going to be exact as the planets do not orbit the Sun in nice, regular circles, but in ellipses, travelling faster when nearest the Sun (perihelion) and slowest when furthest away (aphelion). In addition, the Earth is itself in motion around the Sun, but 19 to 20 years should be a ball park figure.

And if you're a cosmologist, getting the correct date a billion years or so out is accurate enough for you.

The last time that Jupiter and Saturn lined up was 31 May 2000, when they were low down before dawn, but they were close for several weeks either side. The day that I really remember was 6 April 2000, when I was living in St Andrews (sky at 9pm):

Mars and Jupiter were so close together it is hard to separate them on the chart. Add in the Moon and Saturn and you have the second most impressive sight in the sky that night - at the time I had no idea that only a couple of hours later I would witness the most impressive display of aurora borealis I have ever seen.

Nature puts on a free show and most people are walking along with heads slightly down not noticing it.

So, we should look for around 2020 for the next time Jupiter and Saturn are close again.

One way we can consider this is to look at the constellations the planets are in - and I mean the real constellations in the sky, not the "starsigns" you see in magazines.

Date Jupiter Saturn
currently Taurus Libra
13 May 2013 Taurus Virgo
27 June 2013 Gemini Virgo
1 September 2013 Gemini Libra
7 July 2014 Cancer Libra
14 October 2014 Leo Libra
17 January 2015 Leo Scorpius
4 February 2015 Cancer Scorpius
12 May 2015 Cancer Libra
10 June 2015 Leo Libra
16 October 2015 Leo Scorpius
1 December 2015 Leo Ophiuchus
9 August 2016 Virgo Ophiuchus
23 February 2017 Virgo Sagittarius
18 May 2017 Virgo Ophiuchus
15 November 2017 Libra Ophiuchus
19 November 2017 Libra Sagittarius
20 November 2018 Scorpius Sagittarius
13 December 2018 Ophiuchus Sagittarius
16 November 2019 Sagittarius Sagittarius
22 March 2020 Sagittarius Capricornus
3 July 2020 Sagittarius Sagittarius
15 December 2020 Sagittarius Capricornus
18 December 2020 Capricornus Capricornus
25 April 2021 Aquarius Capricornus

Saturn remains in Capricornus until it enters Aquarius on 13 February 2023 - by which time Jupiter will be in Cetus.

Another way is to look at the dates of opposition - when the Sun, Earth and planet line up:

Year Jupiter Saturn Difference (days)
2013 None April 28 146*
2014 January 5 May 10 125
2015 February 6 May 23 106
2016 March 8 June 3 87
2017 April 7 June 15 69
2018 May 9 June 27 49
2019 June 10 July 9 29
2020 July 14 July 20 6

[* Jupiter was last at opposition on 3 December 2012]

This tells us that Jupiter and Saturn will be close through 2020. Now, they will be at their closest on 21 December. This is the sky at 4.30pm that day:

Indeed, Jupiter and Saturn will be so close that the chart shows a single object!

Notice Mars over in Pisces and the Moon. If we go back to 16 December (again at 4.30pm) we find the Moon - only about 48 hours after New Moon - below and to the right:

While on 17 December (again at 4.30pm) we find the Moon now to the left of Jupiter and Saturn:

Now, this is all the action happening low down in the south to south-west. However, let's look at midnight on 20 July - the day Saturn is at opposition, and only 6 days after Jupiter is:

(Meanwhile, over in Cetus, Mars is rising).

As I noted in a previous post, the late summer and autumn evening sky is stable - you only lose stars and constellations slowly into the evening twilight. So, as we might expect, Jupiter and Saturn hang around in the evening sky through summer and autumn, being the sight people might notice for several months - if they look up.

And people might notice something else, planet related, which I have hinted at. On 13 October 2020, Mars is at opposition. It will be closer than most oppositions (and brighter) - although the July 2018 opposition is closer and brighter, Mars will be low down in Capricornus, while at the 2020 opposition it will be higher up, in Pisces. At 9pm on opposition day, the sky will be:

Mars will be the brightest object in the night sky - with these close oppositions it will actually be brighter than Jupiter (although only for a few weeks).

It will be later March/early April 2020 that Mars is close to Jupiter and Saturn, with the sky at 5.30am on 1 April, as it is starting to get light, giving us:

and a close-up on the planets shows:

Worth getting up early for.

What about Venus - does it play a part in this gathering?

Venus starts 2020 as an evening object. In 2019 it is at superior conjunction - when the Earth, Sun and Venus form a straight line in that order - on 14 August. At superior conjunction, Venus cannot be seen, but it marks the point when it slowly reappears in the evening sky. In 2020 it is at greatest eastern elongation (when it is at its furthest from the Sun in the evening sky) on 24 March, then inferior conjunction (when the Earth, Venus and the Sun form a straight line in that order) on 3 June, reappearing in the morning sky and reaching greatest western elongation on 13 August, and then in 2021 it is back at superior conjunction on 26 March.

For a close approach, really we are looking at late 2019 - the sky at 4.30pm on 1 December is:

So you need a very clear south-western horizon to get this gathering. As Jupiter and Saturn move into the evening twilight, Venus moves away from them.

So 2020 will be a year of interesting planetary action.

Tuesday, 23 April 2013

MEPs for England

A fairly common objection to the European Union is that somehow it is trying to "abolish England". The objection seems to rest on the idea that since June 1999 England has been split into regions for European Parliament elections, and that these must be a plan to subdivide England for dark purposes.

There has been a lengthy history of English regions, with the Conservative government creating 10 Government Offices of the Regions in April 1994. At the May 1997 general election, Labour won and was experimenting with further regional devolution, which led to the Regional Development Agencies Act 1998 and defined the boundaries of 9 English regions.

The next significant piece of legislation was the European Parliamentary Elections Act 1999, which moved (in Great Britain) from electing Members of the European Parliament by First Past The Post in one-member constituencies to a form of Proportional Representation (d'Hondt using party lists) in multi-member constituencies. And rather than draw up new constituencies, the existing English regions were used. Hence, the first time many people became aware of the English regions was when they were used to elect MEPs.

Currently, England returns 60 MEPs. If England were a single region (and Germany, with 99 MEPs, doesn't subdivide), then the result for the June 2009 election would have been:

  • Conservatives - 19 (down 5)
  • UK Independence Party - 11 (down 1)
  • Labour - 10
  • Liberal Democrats - 9 (down 1)
  • Greens - 5 (up 3)
  • British National Party - 4 (up 2)
  • English Democrats - 1 (up 1)
  • Christian Party/Christian People's Alliance - 1 (up 1)

There is a middle path. When the Jenkins Commission produced its report into elections for the House of Commons, it recommended a version of the Additional Members System called AV+ - whereas, in elections to the Scottish Parliament, Welsh Assembly and Greater London Assembly, the constituencies use FPTP, the Commission suggested the use of the Alternative Vote for constituencies in the House of Commons.

But what I want to focus on is the report mentioning:

Our investigations (see Annex A) suggest that a Top-up of between 15% and 20% of MPs would do sufficient justice to the three competing criteria discussed above to be acceptable

So, could we switch to a two-tier election - where there are some English regional MEPs, but also a group of English national MEPs to provide national proportionality? If we take the lower limit of 15%, this gives us 51 regional MEPs and 9 national ones.

The number of MEPs each region has is determined using d'Hondt's close cousin, Sainte-Laguë, and if we allocated 51 MEPs across the 9 English regions it would give:

  • South East England - 8 (down 2)
  • London - 7 (down 1)
  • North West England - 7 (down 1)
  • Eastern England - 6 (down 1)
  • West Midlands - 6 (down 1)
  • South West England - 5 (down 1)
  • Yorkshire & Humberside - 5 (down 1)
  • East Midlands - 4 (down 1)
  • North East England - 3 (unchanged)

Electing 51 regional MEPs this way gives us:

  • Conservatives - 21
  • UK Independence Party - 11
  • Labour - 9
  • Liberal Democrats - 8
  • Greens - 2

One thing is obvious - the Conservatives get more seats than they would if England was one electoral region, and the UK Independence Party gets exactly the same number of seats as it would if England were one electoral regions. Hence, it is reasonable to assume neither of these parties would get any of the additional national MEPs.

Denmark uses this method to elect its Parliament, or Folketing, with 135 MPs elected using d'Hondt on a regional basis, and a further 40 elected using Sainte-Laguë on a national basis.

If we stuck with d'Hondt for the 9 additional national MEPs, then these would be 4 from the British National Party, 3 from the Greens and 1 from each of the English Democrats and the Christian Party/Christian People's Alliance, to give us:

  • Conservatives - 21
  • UK Independence Party - 11
  • Labour - 9
  • Liberal Democrats - 8
  • Greens - 5
  • British National Party - 4
  • English Democrats - 1
  • Christian Party/Christian People's Alliance - 1

Alternatively, we could follow the Danish model and use Sainte-Laguë instead, in which case the 9 additional national MEPs would be 3 from the British National Party, 2 from the Greens and 1 from each of the English Democrats, the Christian Party/Christian People's Alliance, Socialist Labour and NO2EU.

Whether we would want some of these to claim to "speak for England" and have an "English mandate" is open to debate. But the major parties could, legitimately, but sneakingly, work around this. Imagine you are a typical Labour voter, in the polling booth, trying to find Labour on the national ballot paper, when you remember something in a leaflet that came through your door advising you to vote Labour in the regional section and for the Co-operative Party in the national section. Conservative Central Office could end up thinking about the old National Liberals and whether it's now time to bring the name back and have some National Liberal MEPs. And your Liberal Democrat activist wondering what if the 1988 merger had never happened and that there were now disagreements between the Liberals and the Social Democrats over which party contests the regional section and which the national section.

An alternative method would be for the election of national MEPs to be parallel to, rather than additional to, the election of regional MEPs, similar to the old system for the Russian Duma. Then the 9 national MEPs would be 3 Conservative, 2 UK Independence Party, 2 Labour, 1 Liberal Democrat and 1 Green.

Another option would be to use a different system for the lower-tier regional MEPs. 51 is 17 times 3, and European law requires each electoral area to return at least 3 MEPs. In the Republic of Ireland, the Single Transferable Vote is used to elect the Dáil Éireann, with each of the 43 constituencies returning between 3 and 5 Teachtaí Dála. Looking at the regions above, the 4 smallest could be single constituencies returning 3 to 5 MEPs, while the 5 largest could be divided into two constituencies. Or a set of 17 three-member constituencies could be drawn up. Or a natural patchwork of 3-to-5-member constituencies could be created roughly following the boundaries of collections of counties.

Sunday, 21 April 2013

By George - A Moveable Saint's Day

I remember back in 1999 chatting to a friend's dad, who worked for the Scouts. And he mentioned the problem they would have the following year for the St George's Day parade, as Easter Day would fall on St George's Day.

I corrected him and said that Easter would be on 23 April, and it can never be on St George's Day.

That might seem odd, but have a look at the Church of England's Common Worship which lists how feasts should be transferred.

With Easter being on a luni-solar calendar (and the timing of Lent, Pentecost and Trinity Sunday depending on it), the year not being made of a complete number of weeks, and other festivals (especially Christmas) being fixed, then there are going to be clashes.

Easter is a powerful thing in this respect as it sweeps clear anything else. Anything falling on a Sunday between the First Sunday of Lent (6 weeks before Easter) and Trinty Sunday (8 weeks after Easter) has to be transferred - normally to the Monday. But there is also a 15-day period, from Palm Sunday to the Second Sunday of Easter, where everything not related to Easter is moved.

Given how late Easter can be, it is clear this can include St George's Day.

Common Worship gives the rules as:

When St George's Day or St Mark's Day [25 April] falls between Palm Sunday and the Second Sunday of Easter inclusive, it is transferred to the Monday after the Second Sunday of Easter. If both fall in this period, St George's Day is transferred to the Monday and St Mark's Day to the Tuesday.

So, what do this mean in practice? Easter can fall between 22 March (which it next does in 2285) and 25 April (which it next does in 2038). If 23 April is a Sunday then we have Easter on 26 March, 2 April, 9 April, 16 April or 23 April.

During the period from 2000 to 4000, the most common day for Easter is 10 April (83 times) - although this doesn't happen next until 2039. How often do these Easters occur and what happens to St George's Day?:

  • 26 March - 58 times (next is 2062). 23 April is then the Fifth Sunday of Easter and St George's Day is transferred to Monday 24 April
  • 2 April - 51 times (next is 2051). 23 April is then the Fourth Sunday of Easter and St George's Day is transferred to Monday 24 April
  • 9 April - 74 times (next is 2023). 23 April is then the Third Sunday of Easter and St George's Day is transferred to Monday 24 April
  • 16 April - 76 times (next is 2017). 23 April is then the Second Sunday of Easter and St George's Day is transferred to Monday 24 April
  • 23 April - 22 times (next is 2079). St George's Day is transferred to Monday 1 May

St George's Day in May raises an interesting issue relating to the idea of a St George's Day Bank Holiday, which I will come to later.

The next set of dates is when 23 April is not a Sunday but is within a week of Easter. Hence Easter falling on or after 17 April:

  • 17 April - 63 times (next is 2022). St George's Day is transferred to Monday 25 April or Tueaday 26 April (curiously, the guidance in Common Worship is unclear as St Mark's Day does not fall within a week of Easter Day. Although from memory, in 2011, Easter Day was on 24 April and St Philip's & St James' Day [1 May - which was the Second Sunday of Easter] was moved to Wednesday 4 May not not to Monday 2 May - with St George's Day being 2 May and St Mark's Day 3 May - so I will assume that St George's Day is on 25 April and St Mark's Day is moved to 26 April)
  • 18 April - 61 times (next is 2049). St George's Day is transferred to Monday 26 April
  • 19 April - 71 times (next is 2071). St George's Day is transferred to Monday 27 April
  • 20 April - 78 times (next is 2014). St George's Day is transferred to Monday 28 April
  • 21 April - 73 times (next is 2019). St George's Day is transferred to Monday 29 April
  • 22 April - 42 times (next is 2057). St George's Day is transferred to Monday 30 April
  • 24 April - 36 times (next is 2095). St George's Day is transferred to Monday 2 May
  • 25 April - 21 times. St George's Day is transferred to Monday 3 May

So, in this period of 2,001 years, St George's Day is on:

  • 23 April - 1,275 times
  • 24 April - 259 times
  • 25 April - 63 times
  • 26 April - 61 times
  • 27 April - 71 times
  • 28 April - 78 times
  • 29 April - 73 times
  • 30 April - 42 times
  • 1 May - 22 times
  • 2 May - 36 times
  • 3 May - 21 times

So, just under 64% of the time, St George's Day falls on 23 April.

One thing to note is that 79 times it is in May. Not just any day in May, but the first Monday in May.

The Banking & Financial Dealings Act 1971 sets the current rules for Bank Holidays (the Trades Union Congress has produced a useful guide to the history of Bank Holidays). Some are set in legislation, but others can be declared by the Queen at a meeting of the Privy Council, e.g. for the Diamond Jubilee or the wedding of the Duke and Duchess of Cambridge.

The "May Day" Bank Holiday - the first Monday in May - falls into this category. This year it's 6 May, as declared at a Privy Council meeting last May. By convention it is the first Monday in May, and hence, some years, it coincides with St George's Day.

There are calls for a St George's Day Bank Holiday, with a campaigning group set up. In the 2010-12 session, there was a St George's Day & St David's Day Bill introduced by Nadhim Zahawi, the Conservative MP for Stratford-on-Avon, which was discussed in the House of Commons. With regards to St George's Day, the proposals were:

  • If 23 April was a working day (i.e. between Monday and Friday, but not Good Friday or Easter Monday) then that would be a Bank Holiday
  • If 23 April were a Saturday (but not the day before Easter Day) then Friday 22 April would be a Bank Holiday
  • If 23 April were a Sunday (but not Easter Day) then Monday 24 April would be a Bank Holiday
  • If 23 April were between Good Friday and Easter Monday then the "Secretary of State" (preumably the Culture, Media & Sport Secretary) would determine the date

There is one objection to this idea, which is that there can be a Bank Holiday overload. If Easter Day is 23 April or later, then we have two Bank Holiday weekends (a 4-day one and a 3-day one) in succession. Throwing St George into the mix increases this.

As St George's Day and May Day are traditional English festivals, I wonder if the solution is to have a single celebration, a Bank Holiday combining them. 27 April is the midpoint, and we have a tradition of Bank Holidays being on a Monday, so the Monday nearest to that could be it. And that would be the last Monday in April.

If that is Easter Monday (which would happen if Easter Day was on 23 April or later) then move the St George's Day & May Day Bank Holiday to the last Tuesday in April.

Sunday, 14 April 2013

The 2011 Scottish Election on Westminster Constituencies

One of my interests is analysing electoral statistics, and there is something I've been working on recently.

In May 2011 there was an election to the Scottish Parliament where the Scottish National Party won 53 constituencies, Labour 15, the Conservatives 3 and the Liberal Democrats 2.

Originally, the Scotland Act 1998 ensured that the constituencies for the House of Commons (Westminster) and the Scottish Parliament (Holyrood) were coterminous - with the exception of Orkney & Shetland which is one constituency at Westminster but two at Holyrood.

Hence the first two elections to Holyrood - May 1999 and May 2003 - were on the same constituencies as those used at Westminster for the May 1997 and June 2001 general elections.

From time to time, the Boundary Commission for Scotland produces recommendations for new and amended constituencies for Westminster - which under the Scotland Act 1998 automatically become those for Holyrood. The 5th review was presented to Alistair Darling, then the Transport Secretary and Scottish Secretary, in December 2004, in time for the May 2005 general election.

The Parliamentary Constituencies Act 1986 required that there be at least 71 Scottish constituencies in the House of Commons, but this was removed by the Scotland Act 1998.

While the review was being conducted, one issue was noticed. The number of Scottish MPs would be reduced from 72 to 59, and according to the Scotland Act 1998, the number of constituencies at Holyrood should therefore be reduced from 73 to 60. For Holyrood, Scotland is split into 8 regions, each returning 7 additional Members of the Scottish Parliament to make the result more proportional.

The Scotland Act 1998 stated that there should be 8 regions still, but the number of additional MSPs to constituency MSPs should be approximately 56:73, as per the origional division. Hence, there should be just 46 additional MSPs, with each region having 5 or 6.

This would have reduced Holyrood from 129 MSPs to 106, and so Westminster passed the Scottish Parliament (Constituencies) Act 2004 which "decoupled" the constituencies, so that Holyrood would remain with 73 constituencies which would be different to the Westminster ones.

This need not be the only option. Ideas I remember floating around at the time were that each Westminster constituency would return one male and one female MSP (making 118 constituency MSPs - and in the political climate of the time these would be predomninantly Labour) and then 11 additional MSPs. Or grouping Westminster constituencies together to elect Holyrood by Single Transferable Vote - a grouping of 2 constituencies would return 4 or 5 MSPs whilst a group of 3 constituencies would return 6 or 7.

Hence, the May 2007 Holyrood election took place using the constituencies used at the 1997 and 2001 Westminster elections - not the 2005.

By the time of the 2011 Holyrood election, the constituencies were pretty old - they were drawn up over 15 years earlier, and ripe for revision. Hence the Boundary Commission for Scotland drew up new constituencies.

There has been a political drift between Westminster and Holyrood. At the 1999 Holyrood election, only 3 constituencies were won by a different party to the 1997 Westminster one. These were:

  • Aberdeen South - held by Labour (who won it from the Conservatives) at Westminster, but won by the Liberal Democrats' Nicol Stephen at Holyrood. Stephen had won Kincardine & Deeside from the Conservatives at a by-election in November 1991 but lost it at the April 1992 general election. In June 2005 he became leader of the Liberal Democrat MSPs and hence Deputy First Minister - a position he held until the Labour/Liberal Democrat administration lost office at the 2007 election.

  • Falkirk West - held by Labour's Dennis Canavan at Westminster, and won by him - this time as an Independent - at Holyrood. Canavan had not been placed on Labour's approved list of Holyrood candidates, but decided to stand anyway, and was expelled from Labour in March 1999.

  • Inverness East, Nairn & Lochaber - held by Labour (who won it from the Liberal Democrats) at Westminster, but won by the Scottish National Party's Fergus Ewing - now the Scottish Minister for Energy, Enterprise & Tourism - at Holyrood.

One thing to notice about 1999 is that the regional vote is fairly close to the constitunecy one. The Scots were still finding their feet electorally, and I guess there was still a "Westminster mentality" - i.e. you vote for a party and whichever gets the most votes wins - rather than a full appreciation of how the Additional Members System works.

Move on to 2003 and things have changed. In addition to the three constituencies above, there are other constituencies which were represented by different parties at Westminster and Holyrood:

  • Aberdeen North - held by Labour at Westminster but gained by the Scottish National Party's Brian Adam at Holyrood
  • Ayr - held by Labour at Westminster but gained by the Conservatives' John Scott at Holyrood, who had gained the seat in a by-election in March 2000 following the resignation of Labour MSP Ian Welsh
  • Dundee East - held by Labour at Westminster but gained by the Scottish National Party's Shona Robison - now the Scottish Minister for Commonwealth Games & Sport - at Holyrood
  • Edinburgh Pentlands - held by Labour at Westminster (where former Foreign & Commonwealth Secretary Malcolm Rifkind failed to reclaim the seat he lost at the 1997 election) but gained by the Conservatives' David McLetchie - at the time the leader of the Conservative MSPs and a frequent taxi passnger - at Holyrood
  • Edinburgh South - held by Labour at Westminster but gained by the Liberal Democrats' Mike Pringle at Holyrood. At the time one of my colleagues was a Liberal Democrat activist, and we agreed that Labour would lose it. He believed the Scottish National Party would win, but my view was that if they can't win a seat with Margo MacDonald as their candidate, then it's unwinnable (8 years later I was proved wrong) and that the Liberal Democrats would gain it from third place
  • Glasgow Springburn - won by the Speaker from Labour at Westminster (in October 2000 its MP, Michael Martin, had been elected Speaker of the House of Commons), but held by Labour's sitting MSP, his son Paul, at Holyrood
  • Ochil - held by Labour at Westminster but won by the Scottish National Party's George Reid - who went on to become the Scottish Parliament's Presiding Officer - at Holyrood
  • Strathkelvin & Bearsden - held by Labour at Westminster but gained by Independent Jean Turner at Holyrood

The methodology is basically the Rallings-Thrasher method. Take the constituency results for the Scottish Parliament election and use tables of local election results (I used the ones from Electoral Calculus) to break these results down into ward results (or parts of a ward where a ward is split across two or three constituencies). The next step is to build up a Westminster result using the wards (and often parts of wards) in that constituency. Scottish local government wards can be quite large (over 20,000 electorate) and one assumption is that support for each party is evenly spread across a particular ward. Some of the hyper-marginal constituencies we get in Edinburgh and Glasgow could have gone the other way.

One quick impression is how across much of Scotland the classic two-party system is back. And those two parties are Labour and the Scottish National Party, with the Conservatives and Liberal Democrats being the rag-tag-and-bobtail candidates. This is the sort of situation where 40% of the vote leads to defeat not victory.

As an overview of the shift from the May 2010 general election:

  • The Scottish National Party gains 31 seats from Labour, 9 from the Liberal Democrats and 1 from the Conservatives - a gain of 41 seats, bringing their tally from 6 to 47
  • Labour loses 31 seats to the Scottish Natonal Party and 1 to the Conservatives - a loss of 32 seats, bringing their tally from 41 to 9
  • The Conservatives gain 1 seat from Labour, 1 from the Liberal Democrats but lose 1 to the Scottish National Party - a net gain of 1 seat, bringing their tally from 1 to 2
  • The Liberal Democrats lose 9 seats to the Scottish National Party and 1 to the Conservatives - a loss of 10 seats, bringing their tally from 11 to 1

With 47 out of 59 seats, the Scottish National Party would 80% of the seats. For comparison, Labour's best was 56 out of 72 at the 1997 election - which is 78%.

Another way of looking at it is that a party would need 425 English seats to be doing that well in England.

Labour's 9 seats would be its lowest number since the October 1931 general election.

For the Conservatives, third place and 2 seats would be their best result since the 1992 election(!). And being just 7 seats behind Labour is equal to the combined Scottish Unionist and National Liberal success at the October 1959 general election. The May 1955 general election, where the Scottish Unionists and the National Liberals managed over half the vote and a majority of Scottish seats between them is another era.

For the Liberal Democrats, their sole seat is their worst result since 1959, when they were also just reduced to Orkney & Shetland. The last time they came fourth in terms of the number of seats was at the October 1974 general election.

The seats won by the Scottish National Party would be:

The seats won by Labour would be:

The seats won by the Conservatives would be:

The seat won by the Liberal Democrats would be:

  • Orkney & Shetland - majority of 14.12% over Independent or majority of 23.92% over the Scottish National Party

At the 2011 election, there were Independent candidates coming second to the Liberal Democrats in both Orkney Islands and Shetland Islands whose combined voted was greater than the Scottish National Party's.

Saturday, 6 April 2013

Thoughts On Clara Oswin Oswald And The Great Intelligence

Last Saturday saw the return of Doctor Who in The Bells of Saint John, which saw Jenna-Louise Coleman make her debut as new companion Clara Oswald, who does not have a middle name.

Well, it was her third appearance. Last year she appeared as Oswin Oswald, junior entertainments officer on the starship Alaska in Asylum of the Daleks - later revealed to have been turned into a Dalek herself, and later on as Clara Oswin Oswald, Victorian barmaid and governess in The Snowmen, which saw the return of the Great Intelligence - also behind the plot last Saturday.

A few things were noticeable. Firstly, Oswin was a computer whizz, able to hack into the Daleks' path web. And the Doctor tells her that the Daleks converted her because she was a genius, and the Daleks need geniuses.

A genius, eh?

Is the Doctor saying the Daleks wanted her because she has a great intelligence?

And that the Daleks need Great Intelligences?

In last week's episode, Alexei notes that Clara is:

Very clever, but no computer skills.

And Miss Kizlet responds:

Upload her anyway, splice her a computer skills package.

A bit later Kizlet states what they do:

We're preserving living minds, in permanent form, in the data cloud. It's like immortality - only fatal.

Bear in mind, from what we know from The Abominable Snowmen and The Web of Fear, the Great Intelligence is non-corporeal.

After the Doctor manages to download Clara from the data cloud, he explains the situation:

There's something in the Wi-Fi. OK, this whole world is swimming in Wi-Fi. We're living in a Wi-Fi soup. Suppose something got inside it. Suppose there was something living in the Wi-Fi, harvesting human minds, extracting them. Imagine that. Human souls trapped like flies in the World Wide Web, stuck for ever, crying out for help. - which Clara compares to Twitter!

Clara keeps her impressive computer knowledge after being downloaded again. But what if she wasn't fully downloaded? Or what if she were copied in the data cloud? Is that what explains Coleman's two previous experiences?

Losing The Winter Sky

Thursday evening, at about 11, I had just got home from work so popped out to Tesco to buy some milk. And this was the sky (from Heavens Above):

The thing that sticks out is that the winter stars are getting a bit low in the west. Of course, dominating that area of the sky is the planet Jupiter - it entered Taurus on 14 May 2012, and will enter Gemini on 27 June 2013, remaining there until 7 July 2014 (see an earlier post for what Taurus and Gemini look like).

The slower-moving Saturn doesn't reach Taurus until 22 July 2029.

When I looked at the various definitions of the year, I also showed what Pisces looked like. You'll notice that along the bottom of the charts the x-axis is given in hours - these hours give the right ascension of any object in the sky.

When the First Point of Aries is due south, then the siderial time is 0 hours. You'll see to the left of the 0h is 1h - and when that line is due south, the siderial time is 1 hour.

This is siderial time - time by the stars. The First Point of Aries takes 23.9344699 hours to return to due south, and we call this the siderial day. So, one siderial hour is equivalent to 59 minutes 50 seconds.

The sky (ignoring solar system objects for the moment) takes 3 minutes 56 seconds short of a day to return to where it was. So, the sky you see at a certain point is what it will look like about half an hour earlier in a week's time and about 2 hours earlier in a month's time.

Hence the sky I saw was the same as midnight in early March (we were then on Universal Time [UT], now on Central European Time [CET]), 2 am in early February, 4 am in early January and 6 am in early December.

These evening stars are setting about 4 minutes earlier each night - but sunset is getting later, which increases the speed with which they disappear into the evening twilight.

And similarly, morning stars are rising about 4 minutes earlier each night - but sunrise is getting earlier.

We see a reverse effect from late summer to autumn.

The following charts (again from Heavens Above) show this. This is the sky at civil darkness (8.12pm CET) on 1 April:

Compare this with civil darkness (9.05pm CET) on 1 May:

Sirius, Orion and Taurus are now setting, while in the east we are startig to get the summer objects such as Vega and Hercules.

What about the morning? On 1 April, civil twilight is 6.08am CET:

The south-eastern sky is dominated by the "Summer Triangle" of Deneb in Cygnus, Vega in Lyra and Altair in Aquila. Low in the south are the zodiacal constellations of Scorpius and Sagittarius, frustratingly low - indeed parts of them are too far south to be seen even from the south coast of England. In June 1998 I was on an observing run using the Isaac Newton Telescope at the Observatorio del Roque de los Muchachos on La Palma, mainly to observe LO Pegasi. Now, on an observing run you spend most time inside, but one night I was unwell and couldn't use the telescope, but had a look out in the evening from my bedroom window and saw Scorpius in its entirety at a decent altitude and realised what we are missing in this country.

The other thing to note is how low in the east we are starting to get the autumn Square of Pegasus and its neighbour Andromeda (parts of which are always above the horizon).

I mentioned that you should expect the sky (with the exception of solar system objects) to be the same a couple of hours earlier each month. So, this sky is that at about:

  • 4am on 1 May (civil twilight 5.02am)
  • 2am on 1 June (civil twilight 4.13am)
  • midnight on 1 July (civil twilight 4.10am)
  • 10pm on 1 August (civil darkness 9.31pm)
  • 8pm on 1 September (sunset 7.51pm, civil darkness 8.25pm)
  • So, this sky is the sky about an hour before civil twilight in early May, and a couple of hours before civil twilight in early June.

    Indeed, if we look at the sky at civil twilight on 1 May, then we see not much has changed:

    And if we go on to the sky at civil twilight on 1 June:

    So, we've lost Leo, Virgo and Scorpius completely, and the autumn constellations are becoming a bit more prominent.

    So, what we see in the spring is that we lose the winter constellations quickly into the evening twilight, but sunrise getting earlier means we only slowly gain the late summer and autumn ones.

    There is a reverse effect in autumn. When I was able to actively observe (now have a flat and loads of streetlights), this was my favourite time of year. September especially - I recall one September when I had 10 clear evenings in a row.

    But there are two other factors that autumn has in its favour. The first is that the evening sky is stable. Firstly, look at the sky at civil darkness on 15 August (9.03pm CET):

    One thing to notice is this year we have Venus low in the west - it was at superior conjunction (behind the Sun) on 28 March and is now moving towards greatest eastern elongation (when it is furthest from the Sun in the sky) on 1 November.

    You might - correctly - say that 15 August isn't autumn, but this is the point in the year when the nights are starting to draw in, when you might be out in an evening and wish you'd put a jumper on, when you might have an indoor light on by the time you go to bed.

    Now move on to civil darkness on 15 September (7.53pm CET):

    Spot the difference. Well, not much really - although Venus and Saturn will be nearer each other (they will be at their closest on 20 September).

    And now civil darkness on 15 October (6.46pm CET):

    With regards to the stars, we have now lost Leo and Virgo - and indeed Saturn is now gone - and the autumn stars are a bit higher in the east. But not all that different to the civil darkness sky from two months previously.

    Note that Venus is near the bright red star Antares - they are at their closest on 16 October.

    And finally, civil darkness on 15 November (4.55pm UT):

    Compared to 3 months earlier, Leo, Virgo and Scorpius are gone. The bright orange star Arcturus is getting lower. In the east Perseus, Andromeda, Triangulum and the Square of Pegasus are higher up, and Aries and Pisces have appeared.

    This is the sky as it would be around midnight 3 months earlier - so roughly 3 hours after it got dark.

    If the evening sky is stable, the other attraction of autumn is that the morning sky is anything but! Start with the sky at civil twilight on 15 August (5.16am CET):

    By then, Jupiter will have moved into Gemini, and you might - after a night of observing the Perseid meteor shower - get your chance to wrap up the observing session by seeing Mercury.

    Starting to make their appearance are some of the winter constellations - Orion, Taurus, Gemini, while the Summer Triangle is getting low in the west.

    Now move on a month to civil twilight on 15 September (6.08am CET):

    And the sky has really moved on. Perseus, Auriga and Taurus are nearly at their best. The winter constellations - with the arrival of the bright stars Sirius and Procyon - are dominating the east and south east. Leo is making its arrival.

    And we also get Mars, down in Cancer. It is nearly at its faintest. if you look at Gemini, then there are the two brighter stars at the left end. The lower one is Pollux and the higher one Castor. Mars and Castor should be about the same brightness.

    Mars won't be at its best until 8 April 2014, when it will be about as bright as Sirius.

    If you look back at that mid-August civil twilight sky, then this will be the sky about 1/4 past 3 mid-September - roughly 3 hours before it gets light.

    We now move another month on to civil twilight on 15 October (6.56am CET):

    The winter constellations and Jupiter dominate the southern sky. The familiar autumn constellations like Pegasus and Pisces are starting to set. Out of the Summer Triangle, only Deneb is visible, low down in the north/north-west.

    Having set in the evening, Arcturus has now risen again and is low in the east. Leo is prominent in the east.

    That mid-September civil twilight sky is now the sky at about 1/4 past 4 (about 2 3/4 hours before it gets light) and the mid-August civil twilight sky is now the sky at about 1/4 past 1 (about 5 3/4 hours before it gets light).

    And finally, civil twilight on 15 November (6.45am UT):

    Even though it is still autumn, the main winter constellations are getting low. In the east, the sky is domiated with the spring constellations, with Vega - like Arcturus - have risen after setting earlier in the night.