John Grenham – Irish Roots

Poor Law Unions are the geographical areas used to collect records of births, marriages and deaths in Ireland. They are still today, astonishingly, based on the old Victorian welfare system. Of course we had to map them.

The Irish Poor Law Act of 1838 divided the island into 130 districts, each with a workhouse at its heart, usually situated in the largest local market town. A property tax was levied to pay for the operations of the workhouse, so the primary rationale for the area covered was that it should produce the necessary tax take. As a result many Unions ignored existing parish and county boundaries, unlike their English equivalents.

By 1852, the Famine had pushed the system to breaking point and new Unions were carved out of the existing districts, particularly in the south and west, where public hunger and distress had been at their most intense. For Irish genealogy, the direct relevance of the Poor Law begins in 1864, when the Victorian public health system based on the Unions took on state registration of births, marriages and deaths. A typically rigid hierarchy ran the whole thing. Local registrars collected registrations in subdivisions of the Union and passed them to the Superintendent Registrar in charge of the each Union, who made copies and passed them on in turn to the Registrar General in Dublin.

When centralised annual indexes were created by the Registrar General for all Irish births, deaths and marriages, the Poor Law Unions were the placenames used in the indexes: they are the locations that appear in the indexes online at IrishGenealogy and FamilySearch. Because most of them cover cross county and parish boundaries, they can be very deceptive.

Egregious examples: Edenderry Union includes parts of counties Kildare, Meath and Offaly, even though Edenderry town is 100% Offaly. New Ross, in Wexford, also covers bits of Carlow and Kilkenny. Ballyshannon takes in equal parts of counties Donegal, Leitrim and Fermanagh, Waterford Union includes most of south Kilkenny …

The point is that the system makes it all too easy to search for your ancestors in the wrong location. Hence the maps.

The townland lists used on to create the maps come from a marriage of the 1851 Townlands Index to the 1885 collection of pamphlets Townlands in Poor Law Unions, produced by the Registrar General to let local registrars know which areas they were responsible for. The pamphlets were collected and republished in book form in 1997 by the late, great George Handran, and republished on CD-ROM by Eneclann in 2006. There were significant changes and reoganisations of the local registrars’ districts over the years, with most of them detailed in George’s book. Dissolved unions (for example, Donaghmore) are identified on our maps, but all of the maps have to be taken as snapshots of an evolving system. Again, George supplies all the nitty-gritty.

The usefulness of having searchable place names from the the Superintendent Registrar’s Districts and the local registrars’ districts lies mainly in identifying the places given in the original registrations. Standard spellings were not used and the handwriting can be fiendish. All registrations include both PLU and local registrar’s district, narrowing the possibilities. An example:

Poor Law Unions are the geographical areas used to collect records of births, marriages and deaths in Ireland. They are still today, astonishingly, based on the old Victorian welfare system. Of course we had to map them.

The Irish Poor Law Act of 1838 divided the island into 130 districts, each with a workhouse at its heart, usually situated in the largest local market town. A property tax was levied to pay for the operations of the workhouse, so the primary rationale for the area covered was that it should produce the necessary tax take. As a result many Unions ignored existing parish and county boundaries, unlike their English equivalents.

By 1852, the Famine had pushed the system to breaking point and new Unions were carved out of the existing districts, particularly in the south and west, where public hunger and distress had been at their most intense.

For Irish genealogy, the direct relevance of the Poor Law begins in 1864, when the Victorian public health system based on the Unions took on state registration of births, marriages and deaths. A typically rigid hierarchy ran the whole thing. Local registrars collected registrations in subdivisions of the Union and passed them to the Superintendent Registrar in charge of the each Union, who made copies and passed them on in turn to the Registrar General in Dublin.

When centralised annual indexes were created by the Registrar General for all Irish births, deaths and marriages, the Poor Law Unions were the placenames used in the indexes: they are the locations that appear in the indexes online at IrishGenealogy and FamilySearch. Because most of them cover cross county and parish boundaries, they can be very deceptive.

Egregious examples: Edenderry Union includes parts of counties Kildare, Meath and Offaly, even though Edenderry town is 100% Offaly. New Ross, in Wexford, also covers bits of Carlow and Kilkenny. Ballyshannon takes in equal parts of counties Donegal, Leitrim and Fermanagh, Waterford Union includes most of south Kilkenny …

The point is that the system makes it all too easy to search for your ancestors in the wrong location. Hence the maps.

The townland lists used on to create the maps come from a marriage of the 1851 Townlands Index to the 1885 collection of pamphlets Townlands in Poor Law Unions, produced by the Registrar General to let local registrars know which areas they were responsible for.

The pamphlets were collected and republished in book form in 1997 by the late, great George Handran, and republished on CD-ROM by Eneclann in 2006. There were significant changes and reoganisations of the local registrars’ districts over the years, with most of them detailed in George’s book. Dissolved unions (for example, Donaghmore) are identified on our maps, but all of the maps have to be taken as snapshots of an evolving system. Again, George supplies all the nitty-gritty.

The usefulness of having searchable place names from the the Superintendent Registrar’s Districts and the local registrars’ districts lies mainly in identifying the places given in the original registrations. Standard spellings were not used and the handwriting can be fiendish. All registrations include both PLU and local registrar’s district, narrowing the possibilities.

An example:

The more we immerse ourselves in the online maps, the clearer it becomes that all maps are provisional. Two dimensions can never truly embody four. Contour lines and hill-shading might approximate the third dimension, but the fourth, time, is always out of reach. Every map is only a snapshot of a moving train.

What prompted such fortune-cookie musings is growing awareness of the disparities between the place-names listings on the site and the new maps that represent them. The listings are based on the 1851 Townlands Index, which reproduces the standardised place-names used for the 1851 census. These standard versions were created by the Ordnance Survey between 1831 and 1841 and were also used by Griffith for the Primary Valuation. The problem is that some of the Ordnance Survey parish boundaries in the Index had changed by the time Griffith came surveying.

The most egregious example is the parish of Ahoghill in Antrim. In the 1851 Index, this consisted of a huge area, almost 150 sq km, roughly 60 square miles. By the time Griffith’s surveyors arrived in the late 1850s, it had been divided in three, Ahoghill, now down to c. 50 sq km/20 sq m, Craigs (51 sq km) and Portglenone (44 sq km).

So, do we stick with the 1851 listing and break the connection to the Griffith’s records? Or mess with the 1851 to make it fit Griffith? The aim of the site is to provide tools and shortcuts for research, not to reproduce prefect facsimiles, so there can only be one choice. Messing and Griffith’s it is.

The Antrim map is already rejigged to represent the 1861-62 Griffith’s survey. Now we just have to pick apart all the record listings for Ahoghill and reknit them into the three smaller parishes. Argh.

Because Griffith’s was published relatively early (1847-1854) for southern counties, with Ulster counties appearing in the 1860s, the areas most affected by the discrepancies are in Derry, Fermanagh, Antrim and Down. Parishes on the Ards Peninsula in Down appear to have been shuffled crazily in the 1860s, with the mapped areas bearing very peculiar relationships to the listings. There are many, many townlands mapped into one parish, but recorded in another in the Valuation.

We’ll correct the worst offenders over time, but in many cases the effort is pointless. It’s also not clear that some of the parishes scattered in multiple parts on the maps are entirely accurate, at least for Griffith’s. So sup with a longish spoon and gird your loins with that trusty genealogical scepticism.

(A quieter caveat is that we’ve finally implemented the soft paywall for the maps. In the first flush of enthusiasm we forgot about it.

But even without paying, almost everything is free via zooming on the county pages (e.g https://www.johngrenham.com/browse/county_civil.php?county=Roscommon), with the paywall counter starting only if you click through to a parish page. Shh.)

 

After putting up the all-singing, all-dancing, click-and-zoom civil parish maps in July, it occurred to us that the townlands available from Open Street Map (townlands.ie) would fit nicely into them. If we just could get the data down to manageable file-sizes. So we squeezed and squeezed and … Whoopee. All-singing, all-dancing, and now break out the top hat and cane.

Then, after repeatedly hunting inside various parish maps for a townland, we thought it would be nice if you could just click on the townland name to drop a marker showing exactly where it was. So we did that. Yeehaww. All-singing, all-dancing, add white gloves to the top hat and cane.

 

 

But that only worked if the townland names on the maps matched those in our own listing. And in about 10% of cases, more than 6,000 names, they didn’t. As my Auntie Breda used to say, it would give you a pain in your Laladooshia. But there was no choice. At the end of July, we began the process of standardising the names, one by irritating one. There were plenty of typos in the OSM lists and (mumbles) one or two in our own. Many came down to commas and brackets and hyphens. The standard we aimed for was the Griffith’s spelling, so no Irish versions are included, and nothing later than the 1860s. This is where the differences between OSM and ourselves came into play: They aim for contemporary geographical accuracy; we just want to make it easier and more intuitive to get at the records.

Anyway, that process ended last week, with about 98% of place names now plonking down markers when you click them.

Then the fact that the markers were unclickable began to annoy me. Don’t just sit there, you lazy little marker, do something. So I made a popup from the marker to include links to records covering that townland: Griffith’s, civil BMD records, 1901, 1911. Wahey, the circus is in town.

What this means is that for 98% of townlands in rural Ireland, you can now click through to the universally relevant records from the 1850s to the 1920s, while seeing the location of each townland in relation to its neighbours. A historical snooper’s dream: in other words, mine. Enjoy.

(In the course to the eyeball-to-eyeball engagement with the townlands, I came across some doozies. Here, for your delectation, are a few of my favourites:

• • • America and Liberty, just across the fields from my ancestral homeplace, probably testament to local eighteenth-century support for the American Revolution.
    • The geographically confused: Gibralter, Antigua, Siberia, Lilliput and the hard-to-find Castlelost.
    • Rows that got out of hand: Bastardstown, Slutsend, Stranglewilly, Controversyland
    • Good, plain Presbyterians: John Gillins, True, Christendom)

Irish civil parishes are strange beasts. For good or ill, knowing about them is  essential for Irish research.

They came into existence, as simple parishes, after the twelfth century attempt to tame the exotic Irish church and bring it into line with Roman norms. As well as having scandalous marriage laws, outrageous hairstyles, deadly book illustrators and a peculiar way of calculating the date of Easter, the early Irish church disregarded the Imperial Roman territorial divisions of parishes and dioceses. We preferred vast monastic holdings that resembled secular kingdoms in power and wealth, and were passed on, like kingdoms, from father(-Abbot) to son(-Abbot).

We eventually had clerical celibacy, parishes and manners put on us by the Anglo-Normans.

Until the sixteenth century, however,  parishes were only in widespread use in the English-controlled areas of the East. The Reformation in Tudor England transformed the Church there into an arm of the state, and then they conquered the entire island. As a result, a parish structure was finally extended country-wide and became the bailiwick of the Anglican Church in Ireland, the Church of Ireland. Because church and state were so intertwined, these parishes then also became the geographic basis of public administration, hence “civil” parishes. For two centuries from roughly 1700 to 1900, tax records, militia levies, rentals and censuses were all based on the civil parish.

Why bring this up now? Because we’ve just implemented a major upgrade to the civil parish maps on this site. Maps were always important, because you need to know where the blasted parishes are, as well as what townlands they contain and which other parishes adjoin them. But the old maps were a little … underwhelming. With the new ones, you can zoom in and out, see where the constituent townlands are and best of all enjoy the whole thing in glorious Candy-Crush colour.

The colour began as an afterthought, but turned out to be a revelation. Many parishes have gained or lost townlands over the centuries and now consist of multiple separate areas. Colour makes the links between the various parts jump out. Have a look at Drumcolumb parish in Sligo, scattered in four distinct parts. Or my favourite, Shrule in Mayo, which incorporates a townland from the separate parish of Kilmainebeg, with that townland then incorporating a separate townland from Shrule. M.C Escher, eat your heart out.

All of this was only possible because of the wonderful work of OpenStreetMap, in particular the mappers behind www.townlands.ie, who made all their data open-source. For some extraordinary work by one of the people behind that site, have a look at Brian Hollinshead’s astonishing www.dublinhistoricmaps.ie. It is Map-Nut Nirvana. Not only does it include every single map of Dublin ever published, all overlaid on contemporary street plans, it also includes some extremely useful guides, for example to local registrars’ districts. As well as some deeply odd ones: Dublin Compressed Gas Filling Stations?

Once you begin to get a taste for this stuff …

I never studied history academically. In fact, my last formal schooling in the subject ended at the age of fifteen. So, like all autodidacts, I have a certain twitchiness about the lacunae in my knowledge (I kept up the Latin). I’ve just finished Dáibhí Ó Cróinín’s Early Medieval Ireland 400-1200 (2^nd ed. Longman, 2016), in an attempt to fill in some of the gaps. (Full disclosure: Dáibhí was a school friend of mine more than fifty years ago). The book isn’t always an easy read, though there some nice flashes of impish humour.

What really shines is the treatment of sources. The second edition includes the best short account I’ve read of the potential of early Irish genealogies. These things are vast. Nothing on their scale exists anywhere else in the world, and they have puzzled and frustrated historians and genealogists for generations. The earliest surviving manuscripts date to the 1100s, but the historical individuals they cover go back to at least the fifth century, 600 years earlier, and of course the makey-uppy individuals go back to the Garden of Eden. They cover tens of thousands of families, some in biblical lists of “who-begats”, and others in great spreading collateral branches. But they include not a single date and are distorted by centuries of revision and rewriting, as one ruling family usurped another and needed to show their ancient legitimacy. So using them as historical or genealogical sources is almost impossible. Half may be true: but which half?

What the book does is show how it is (sometimes) possible to cross-reference a genealogy with some of the Irish annals, for example to identify a year of death and thus an entire extended family. This is not easy, to put it mildly, requiring fluency in Old Irish and Latin, familiarity with continental medieval manuscripts, and the patience of a saint, a constellation of qualities hard to find in one person. Step forward Dáibhí.

Miraculously, the wonderful section on medieval sources is part of the free sample download on Kindle. Sorry Dáibhí.

A more general point I took from the book is the extent to which Gaelic Irish surnames encode the story of medieval Ireland. Yes, all of those names are patronymic, Ó, ‘grandson of’, and M[a]c, ‘son of’, in a way that shows just how important descent and genealogy was in Ireland when hereditary surnames began in the 10^th and 11^th centuries. But the stems of the surnames, the personal names of the individuals from whom descent is being claimed, embody a world that existed long before there were surnames.

There was a lot of fighting in that world. Warrior-derived surnames abound: Cath means ‘battle’, with the personal name Cathán meaning ‘battler’. Ó Catháin, ‘grandson of the battler’, arose in at many different areas and was anglicized as ‘Keen’ (Co. Down), as Kane and Keane (Cos Derry and Galway), as Kane in Mayo … That most Irish of all names, Murphy, in modern Irish Ó Murchú, comes from the personal name Murchadh, Mara, ‘sea’ and Cú, ‘hound’, so ‘sea-fighter’. There must have been a lot of them: the surname arose in at least seven different areas. The many other warrior surnames include Coakley, Conville, Donoghue, Dooley, Duncan , Falahee, Ganley, Glancy, Hanrahan, Hargy, Horohoe, Hourihane, Kemp, Kimber, Lambe, McCamley, McCann, Clancy, McDonagh, McEvilly, McKinley, McMurrough, Loane, Looney, Shanley, Staunton , Moraghan, Murchan, – all based on fighting.

We were also quite pious between the fifth and ninth centuries, though our attitudes to marriage and inheritance scandalised celibate British and Roman fellow clerics (Ó Cléirigh, grandson of Cleary, from Cléireach, a cleric who shouldn’t be having grandchildren). Surnames actually including the word “Monk” (manach) include the anglicised Moynes, Monaher, Monaghan, Mannix, Manahan, Managh, McEvaney, McCavana … Plenty of offspring there. Much more numerous are the Mac Giolla (‘son of the devotee of ‘) surnames: Gilboy, Gilbride, Gilchrist, Gildea, Gilfedder, Gilfoyle, Gilhooley, Gillane, Gillard, Gilleese, Gillespie, Gilligan, Gilmartin, Gilmore, Gilmurray, Gilsenan and many more. My favourite are the Mul- surnames. The Irish maol literally means ‘bald’, and was used to refer (disrespectfully?) to churchmen because of the distinctive tonsures they sported. Hence Mulcahy, Mulderrig, Muldoon, Mulhall, Mulholland, Mullally, Mullane, Mullarkey, Mulleady, Mullen, Mulligan, Mulready, Mulrooney, Mulvey, Mulvihill …  If you know a middle-aged man with one of these names who’s going bald, cheer him up by telling him about the Irish tonsure.

The Vikings first came to Ireland at the end of the 8^th century and left plenty of progeny bearing personal names that went on to become the roots of hereditary surnames: Doyle (Dubh Gall, dark foreigner, Danish apparently), Groarke (Hrothrekr, Norse), Higgins (Mag Uiginn, son of the viking), Loughlin (Ó Lochlainn, grandson of the viking), McManus (Mac Maghnuis, from Mánus, a Norse forename, Reynolds (Mac Raghnaill again Norse), Sugrue (Sigfrid), Beirne (Bjorn), Broderick (Brudar).

I could go on. In fact I have. If you want more about surnames coming into English, there’s a series of new rants on YouTube. Dáibhí’s book is on Amazon.  Most of the medieval annals are online at the University College Cork CELT website. Some of the genealogy manuscripts are at University College Dublin’s Irish Script on Screen site.

IrishGenealogy’s civil records section is the greatest thing since sliced bread. Almost single-handedly, it has revolutionised Irish family history research.  By making almost 16 million births, marriages and deaths open and free, it also enabled whole fields of research – into infant mortality, townland history, occupational history – that had been closed books. So all hail IrishGenealogy.

Except (there had to be an “except”) … over the years, I’ve regularly noticed peculiarities in the site’s search results: the same search parameters sometimes bring back different results or the results include many more records than expected.

So, in the same spirit of enquiry that led me to dismantle Santa’s 1959 gift of a wonderful toy Winchester, I went down the rabbit-hole that is the IG search interface. Be warned. This gets geeky.

First the simple search: Enter a first name and a last name here and the site will search its forename records (and all forename variants) for that first name and search its surname records (and all surname variants) for the last name. But there is no way here to turn off the variant search. So entering “Jeremiah”, always also finds Darby and Jerry and Gerard and Gerald, entering Smith always also finds Smyth, Smythe, Smithe ….

In addition, for births after 1899, where the record includes the mother’s maiden surname, that surname and its variants are also searched.  So a search for any “Sadler” born in 1900 also returns all births where the mother’s name is Sadler or Sadlier or Sadleir … There is no way to confine the search to the child’s surname only.

The innocuous-sounding “More Search Options” is even more complex, especially when searching on two names. The crucial distinction to keep in mind is between the names entered at the top of the page and any other names you enter for mother’s maiden name or spouse.

First, like the simple search, the main forename and surname boxes search only forenames and surnames. And again, for births the mother’s maiden name is always searched. Here you can turn off variants by ticking the “exact match” box, but this won’t turn off the mother’s maiden name search. So we have actually no way on the site to search just for a child’s birth surname.

The most important quirk in the “More” area is that variants are only ever applied to the main surname, not any other names you enter. So searching for a Walsh with mother’s maiden name Burke (after 1899) will find all Walsh/Welsh/Walch children with mother’s name spelt exactly “Burke”. No Bourk or Bourke or De Burca mothers.

This can be especially tricky with marriage searches. Again, variants are only applied to the two main names, not anything you enter in the “2nd Party” box. Search for a Derby Moriarty who married a McElligot and you’ll get all variants of Derby and all variants of Moriarty but only the exact spelling of McElligot. And I count at least 13 variant spellings of McElligot in the marriage records.

Because IG marriages cover more than a century, and because marriage records are the most useful of the BMD trifecta, the moral is clear: always reverse names and do a second marriage search, whether or not you think you’ve found what you want.

Finally, remember that wild cards are always usable in all the name boxes in the “More” area.

Enough? I could go on … if you really want more, here’s a YouTube with lots of examples.

And I never got the wonderful Winchester to work again. Another moral there.

We’ve just done a big upgrade to the civil births maps, adding the 8 years 1914 to 1921, to match the civil deaths and marriages already covering those years. So three cheers, claps on the back, pints of shamrock all round. Except …

I thought the hair-raising LDS death and marriage transcriptions were the limit. No they weren’t.
Jhonston 1920
Johnnston 1915
Johnsron 1916
Johnstin 1914
Johsnton 1920
Or try:
Mujphy 1917
Murphu 1917
Murpjy 1921
Murpny 1915
Murpyy 1916

A number of demons are in play. First, from 1914 to 1918 many of the civil servants running the registration system were away at war. Those who took over seem to have had poor literacy skills, to put it kindly. Then from 1918 to 1921, came the War of Independence, with continual sabotage of all the existing (British) systems of administration, including the civil registration system. So there are plenty of good reasons why your ancestors’ birth, marriage or death might not be there.

However, in the case of birth records from about 1911, there also seems to have been a bit of a breakdown in the transcription system used by IrishGenealogy. MAH MT KVAX born in Belfast in 1920? I don’t think so. If you look at the original, the child was May McKeag.

The only way a transcriber could have produced this is by trying to type with their head instead of their fingers. Banging their forehead off the keyboard, in other words.

Whoever that poor suffering transcriber was, they left a trail of anguish: VUFARU MT TONNVLL; TACYVRZNV MAYONH ;   BAMDVL ZRFZN.

The problems seem to be specific to IrishGenealogy births from about 1911 to 1921, so if you’re not finding a birth that should be there in that period, sweat other sources – FamilySearch, Rootsireland, the Northern Ireland GRO, even the printed indexes in the Dublin GRO search room.

Bamdvl Zrfzn, long-lost scion of the Irwins, I feel your pain.

 

 

 

In Irish research, birth, marriage and death records are most definitely not created equal. State death records capture only incidental family information, births give just a single generation, but a marriage record supplies both fathers’ names and occupations, the couple’s ages, addresses and occupations, their witnesses, the clergyman’s name, the church …  Loads and loads of loverly threads for us all to follow.

Marriages are also the earliest, starting in 1845, and the easiest to find, with two big bull’s-eyes, bride and groom, both appearing in the indexes, begging to be cross-referenced.

So happy days.  We’ve now got the Latter Day Saints transcripts of Irish General Register Office marriage indexes 1845-1922 mapped and available on the site. Check out Hession, for example .

In the great gappy jigsaw puzzle that is Irish research, there are only four universally useful record-sets:

• Griffith’s,
• Parish registers,
• 1901 and 1911 censuses
• GRO birth, marriage and death records.

These new marriage maps are our final piece of that puzzle. It’s only taken us ten years.

Because the records are a bit different, we’ve treated them a bit differently. You can search using a forename, and the double surname search offers an option to check for actual marriages between the two surnames on IrishGenealogy. Calloo, callay, oh frabjous day.

If you detect a certain hesitancy in the enthusiasm, it’s because time spent cloistered with these records has once more provided a close-up of their imperfections. Again, numerous records are skipped, duplicated and mangled. Whoever was responsible for the 1890s once again used their spreadsheet fill-down function as a shortcut for duplicate records. Except that many many are not duplicate, and so are just plain wrong. And of course the same problems here and on FamilySearch are also in the licensed copies on Ancestry and FindMyPast. Sup with a very long spoon.

Before getting involved in mapping their marriage (and death) indexes, if you’d asked me what I thought of the FamilySearch validation process, I’d have said it was pretty good. Years back, I signed up online to be a transcriber for them (just being nosy) and it seemed like a serious business: double transcripts automatically checked against each other, with conflicts resolved by a third party. Mar dhea.

And what do I think of the FamilySearch validation process now? It would have been a good idea.

 

 

Grief is one of the drivers of genealogy, whether we acknowledge it or not, and a reason why most of us are middle-aged or older: only after losses brought by age do you feel the need to slow the decay involved in forgetting. So grief can be put to use. It can even be beautiful.

What brought an odd thought like this this to mind was listening to a folk-song, “The Lost Words Blessing”, part of a musical version of the children’s book The Lost Words.  A 2017 collaboration between the nature writer Robert Macfarlane and the painter Jackie Morris, the book was a response to the exclusion of twenty names for everyday nature from the Oxford Junior Dictionary, due to their underuse by contemporary children. The lost words included such ordinary things as Acorn, Wren, Hare, Otter, Lark … Their omission from the Dictionary was simply recognition that nature  has receded farther and farther from the lives of children, a tiny poignant symptom of the vast extinctions happening around us in the natural world as we consume more and more of it.

The paintings and poems in the book are extraordinary, and I have no doubt they succeed in their aim of intriguing children into love of these wonderful ordinary creatures who are leaving us forever. But the book is necessarily an elegy for the natural world it celebrates, with some of its beauty coming from that elegiac tinge.

In the song, that atmosphere becomes almost unbearably intense. The words bless a child entering into the world and pray for the child to recognize and take on the natural qualities of the heron and the kingfisher and the otter, even as the animals themselves are ceasing to exist. They will somehow survive in that way, as an afterlife, a glint of light in a starling’s eye reflected out into the universe “past dying stars exploding” – “Like the little aviator, sing your heart to all dark matter”.

It is a kind of survival, but only in the sense that our ancestors survive in a family tree. It is making beautiful use of grief.

As you might gather, I found the song very moving. That might have been due to listening to it on Hogmanay with a glass of whiskey in my hand and hearing a Scottish singer wonderfully rhyme Otter with Water. See what you think yourself. Song here, words here.

Happy New Year.